Public report of hankcs/HanLP, published 25 October 2026. It has the same score, grades and number of
findings as the full report. The details of security findings (where they are in the code, which known
vulnerabilities and which passwords or keys were found) are left out, so nobody can use them against the project.
Public
Executive reportMeasured under the Code Assurance Index, an open standard for code quality, using its published scoring rules (rubric version 2026.10.5). Watchdog signed this survey and filed it in the standard's public register, where anyone can look it up by its survey ID. Check this survey →
Signed and filed in the standard's public register, 25 October 2026, 08:17 UTC · Survey ID cd_be827990fb0542338c3fa7130a7ddfaf
hankcs/HanLP
Measured 25 October 2026, 08:15 UTC, at code version 281e8ef
No regression since the last survey
A regression is a drop of 0.5 points or more in the score, or of 2.5 or more out of 10 in a dimension, since the last survey.
CAI scoreAdequate
CriticalWeakAdequateStrongExemplary
No change since the last survey
The score is built from these lenses. Each lens grades one area of the code, such as how easy it is to change or how well it is protected. The weakest lens counts most, so one weak area pulls the score down instead of being averaged away.
A medium codebase: 50,355 lines of product code in 5 projects, tests not counted · Estimated rebuild effort: about 0.3 person-years · Weakest lens: Code Health, 46
Findings by how serious they are
14 critical389 serious77 minor
11 dimensions are left out of the score. For each one, Watchdog either couldn't measure it in this code, or found no evidence of the safeguard it looks for, such as a tested backup restore. They count neither for nor against the score, and Limitations lists each one with its reason.
This survey was produced by
Watchdog
Producer
Canine Development
Analyzer
Watchdog engine 1.0.0
Measured
25 October 2026, 08:15 UTC
A measurement, not a certificate. The Code Assurance Index records a number
together with the evidence it was worked out from, so anyone can recalculate it, but it doesn't certify or guarantee
the code. The standard was written by Canine Development, which also builds Watchdog. It is open to any tool that
implements it, and today Watchdog is the only one that does.
The score is deterministic, which means the same code always gets the same score. Watchdog works it out with tools that follow published rules, and an AI model never decides it, so anyone who applies the same rules to the same code gets the same number. How it is done is explained under Methodology ▸
57of 57 dimensions measured by tools.
467of 480 findings point to an exact file and line, so you can open each one in the code. The other 13 concern the repository as a whole, such as missing documentation.
57dimensions across 6 lenses, read from 50,355 lines of product code in 5 projects.
hankcs/HanLP falls short of our standard for production, held back most by its Code Health lens, the quality of the code itself. It scores 53 out of 100 and is rated Adequate, and all 14 of its critical findings are security problems.
Code Health, at 46, is the weakest lens and counts most in the score. The best first step is adding automated security scanning, worth about 4 points. Declaring which values may be missing would lift Code Health itself, worth about 3 points. Raising Code Health to 70 would lift the score to about 58.
Six places in the code run text as code, which an attacker could use. Four pipeline tools are referenced by a changeable name, so the build pipeline can be tampered with, and two pipeline steps install packages that can change without warning. Replacing text run as code with direct calls, pinning each tool to an exact version and pinning what the pipeline installs would close these risks. Updating the vulnerable packages would fix the package with a known security hole, and the password or key in the code's history needs replacing, with the old one switched off.
The product code is 50,355 lines, about €49,000 to rebuild.
A lens is capped at Adequate while one of its dimensions scores below 4 out of 10, however well the rest of the lens does. That way a serious problem in one place always shows in the lens's rating.
How the score is built: what each lens counts forThe lenses don't count equally. The weakest lens counts most, the next weakest a little less, and so on, so a weak area pulls the score down instead of being averaged away. In the bar below, each part's width shows how much that lens counts and its colour shows its rating. Improving the widest part moves the score most, and as that lens improves, it counts for less, because another lens becomes the weakest.
If Code Health rose from 46 to 70, the lowest score rated Strong, the score would rise from 53 to about 58.
Code composition: what the lines of code are
Product code 99%Tests 1%
Rebuild cost: about €49,000 · Estimate
Cost to rebuild
€16,000 to €82,000, for 273 to 866 hours of work, about 0.3 person-years for one engineer
Difficulty of the problem
Standard, judged from the kind of software and how it is built. Code that solves a harder problem, such as payments or medical bookings, costs more per line to write.
Quality factor
×0.7 at a score of 53. Better code takes more work to write, so the factor is ×1.0 at a score of 70, lower below it and higher above it.
Size
Medium: 50,355 lines of product code
How the estimate is made: Watchdog prices the hours a team would need to write this code again, at €60 to €95 an hour. Declarations and boilerplate are priced at a fast rate and logic at a slower one, and the logic is then multiplied by the difficulty of the problem and by the quality factor. The estimate can be about 30% off either way, and it moves most with the hourly rate and the difficulty. For this code, every line is priced as straight logic, which keeps the estimate on the low side.
Top priorities
The three actions that raise the score most for the work they take. The full ranked list is in the Roadmap.
1
Add a SAST step to CI running what this repository's stack ships: bandit, `semgrep --config=p/python`, or CodeQL's python pack — so a security regression fails the build instead of landing.
Declare the default where Java declares it: `@NullMarked` (org.jspecify.annotations) in each package's `package-info.java`, or once in `module-info.java`, and run NullAway (Error Prone) or the Checker Framework so the declaration is enforced.
Then annotate the real exceptions `@Nullable` and resolve the suppressions rather than adding to them: a `@SuppressWarnings("NullAway")` or `castToNonNull(x)` on a value the checker could show to be null is a `NullPointerException` waiting for the path that makes it true — where the value genuinely cannot be null, say something the checker can see instead (a null check, `Objects.requireNonNull`, a non-null field initialised in the constructor).
Each conclusion below combines several dimensions into one point that no single dimension shows, and lists the measurements it rests on.
Where the investment is most exposed · Value at risk
This code would cost about €49,000 to build again, about 0.3 person-years of work. Its weakest lens is Code Health, at 46 (Weak). Code Health is about how easy the code is to read and change safely, and code that is hard to read is slow and risky to change, so every future change costs more than it should. Because the weakest lens counts most in the score, work on Code Health also moves the score most.
Based on: the rebuild estimate · the Code Health grade · how much each lens counts
What to do: Start with the Code Health actions in Top priorities.
If you do one thing first · First action
Add a SAST step to CI running what this repository's stack ships: bandit, `semgrep --config=p/python`, or CodeQL's python pack — so a security regression fails the build instead of landing. Of all the actions in this report, it raises the score most for the work it takes: about 4 points for medium effort. The next survey will show the change in the score.
Based on: the ranked actions in Top priorities, by score gain and effort
What hard-to-read code costs each year · Cost
Four dimensions measure how easy the code is to work in: two measure how tangled its logic is, one how much of it is copied, and one how well each part sticks to a single job. In this codebase they average 6.7 out of 10. Code at that level is slower to change than clean code, and Watchdog estimates the extra work at 3% to 7% on every change. The history of the code shows how much it changes: about 428 lines over the last 90 days, which is about 1,736 lines a year. At that pace, the extra work comes to an estimated 0.1 to 0.8 engineer-days a year, and it grows as more code changes.
How this is estimated: The number of changed lines is measured from the history of the code. The rest rests on two assumptions of Watchdog's own: code scoring 8 out of 10 or more adds no extra work, and each point below 8 adds 2.5% to 5.6% to every change; and an engineer changes 150 to 400 lines in a day. The result is a range for planning, not a measurement.
Based on: Cyclomatic Complexity · Cognitive Complexity · Code Duplication · Cohesion (LCOM4) · the code's change history
What to do: Start with the files listed under Churn × Complexity Hotspots, which change often and are hard to read, because that is where the extra work is paid most often.
Architecture: which parts of the code depend on which
The code is divided into modules, groups of code that each handle one part of the product, and a module depends on another when it uses code from it. The table lists the modules both down the side and across the top, in an order where each module should only use modules listed above it. A number in a cell means the module on that row uses the module in that column, and it counts the pairs of building blocks (types) that connect them. The one thing to look for is a mark above the diagonal line. It means two or more modules use each other in a loop (a dependency cycle), so none of them can be changed, tested or released on its own.
240 modules, with 263 dependencies between them. Every module uses only modules listed above it, so there are no loops.
The table shows the 40 modules with the most connections; 200 more are left out to keep it readable.
Which modules use which. The module on each row uses the module in each column, and the number counts the pairs of types that connect them. A cell above the diagonal is part of a loop.
common.keras_component uses common.component, so a change to common.component can break common.keras_component, while a change to common.keras_component can't break common.component.
Position
Below the diagonal: common.keras_component uses a module listed above it, which is the healthy direction.
common.keras_component uses common.transform_tf, so a change to common.transform_tf can break common.keras_component, while a change to common.keras_component can't break common.transform_tf.
Position
Below the diagonal: common.keras_component uses a module listed above it, which is the healthy direction.
common.torch_component uses common.component, so a change to common.component can break common.torch_component, while a change to common.torch_component can't break common.component.
Position
Below the diagonal: common.torch_component uses a module listed above it, which is the healthy direction.
common.transform uses common.vocab, so a change to common.vocab can break common.transform, while a change to common.transform can't break common.vocab.
Position
Below the diagonal: common.transform uses a module listed above it, which is the healthy direction.
layers.embeddings.contextual_word_embedding uses layers.embeddings.embedding, so a change to layers.embeddings.embedding can break layers.embeddings.contextual_word_embedding, while a change to layers.embeddings.contextual_word_embedding can't break layers.embeddings.embedding.
Position
Below the diagonal: layers.embeddings.contextual_word_embedding uses a module listed above it, which is the healthy direction.
layers.embeddings.contextual_word_embedding uses transform.transformer_tokenizer, so a change to transform.transformer_tokenizer can break layers.embeddings.contextual_word_embedding, while a change to layers.embeddings.contextual_word_embedding can't break transform.transformer_tokenizer.
Position
Below the diagonal: layers.embeddings.contextual_word_embedding uses a module listed above it, which is the healthy direction.
layers.scalar_mix uses common.structure, so a change to common.structure can break layers.scalar_mix, while a change to layers.scalar_mix can't break common.structure.
Position
Below the diagonal: layers.scalar_mix uses a module listed above it, which is the healthy direction.
transform.tsv_tf uses common.transform_tf, so a change to common.transform_tf can break transform.tsv_tf, while a change to transform.tsv_tf can't break common.transform_tf.
Position
Below the diagonal: transform.tsv_tf uses a module listed above it, which is the healthy direction.
common.dataset uses common.transform, so a change to common.transform can break common.dataset, while a change to common.dataset can't break common.transform.
Position
Below the diagonal: common.dataset uses a module listed above it, which is the healthy direction.
components.ner.biaffine_ner.biaffine_ner uses common.structure, so a change to common.structure can break components.ner.biaffine_ner.biaffine_ner, while a change to components.ner.biaffine_ner.biaffine_ner can't break common.structure.
Position
Below the diagonal: components.ner.biaffine_ner.biaffine_ner uses a module listed above it, which is the healthy direction.
components.ner.biaffine_ner.biaffine_ner uses layers.embeddings.embedding, so a change to layers.embeddings.embedding can break components.ner.biaffine_ner.biaffine_ner, while a change to components.ner.biaffine_ner.biaffine_ner can't break layers.embeddings.embedding.
Position
Below the diagonal: components.ner.biaffine_ner.biaffine_ner uses a module listed above it, which is the healthy direction.
components.ner.biaffine_ner.biaffine_ner uses common.torch_component, so a change to common.torch_component can break components.ner.biaffine_ner.biaffine_ner, while a change to components.ner.biaffine_ner.biaffine_ner can't break common.torch_component.
Position
Below the diagonal: components.ner.biaffine_ner.biaffine_ner uses a module listed above it, which is the healthy direction.
components.parsers.ud.ud_parser uses common.structure, so a change to common.structure can break components.parsers.ud.ud_parser, while a change to components.parsers.ud.ud_parser can't break common.structure.
Position
Below the diagonal: components.parsers.ud.ud_parser uses a module listed above it, which is the healthy direction.
components.parsers.ud.ud_parser uses common.torch_component, so a change to common.torch_component can break components.parsers.ud.ud_parser, while a change to components.parsers.ud.ud_parser can't break common.torch_component.
Position
Below the diagonal: components.parsers.ud.ud_parser uses a module listed above it, which is the healthy direction.
components.parsers.ud.ud_parser uses layers.embeddings.contextual_word_embedding, so a change to layers.embeddings.contextual_word_embedding can break components.parsers.ud.ud_parser, while a change to components.parsers.ud.ud_parser can't break layers.embeddings.contextual_word_embedding.
Position
Below the diagonal: components.parsers.ud.ud_parser uses a module listed above it, which is the healthy direction.
components.taggers.tagger_tf uses common.keras_component, so a change to common.keras_component can break components.taggers.tagger_tf, while a change to components.taggers.tagger_tf can't break common.keras_component.
Position
Below the diagonal: components.taggers.tagger_tf uses a module listed above it, which is the healthy direction.
components.taggers.transformers.transformer_transform_tf uses common.transform_tf, so a change to common.transform_tf can break components.taggers.transformers.transformer_transform_tf, while a change to components.taggers.transformers.transformer_transform_tf can't break common.transform_tf.
Position
Below the diagonal: components.taggers.transformers.transformer_transform_tf uses a module listed above it, which is the healthy direction.
components.taggers.transformers.transformer_transform_tf uses transform.tsv_tf, so a change to transform.tsv_tf can break components.taggers.transformers.transformer_transform_tf, while a change to components.taggers.transformers.transformer_transform_tf can't break transform.tsv_tf.
Position
Below the diagonal: components.taggers.transformers.transformer_transform_tf uses a module listed above it, which is the healthy direction.
components.amr.seq2seq.seq2seq_amr_parser uses common.structure, so a change to common.structure can break components.amr.seq2seq.seq2seq_amr_parser, while a change to components.amr.seq2seq.seq2seq_amr_parser can't break common.structure.
Position
Below the diagonal: components.amr.seq2seq.seq2seq_amr_parser uses a module listed above it, which is the healthy direction.
components.amr.seq2seq.seq2seq_amr_parser uses common.torch_component, so a change to common.torch_component can break components.amr.seq2seq.seq2seq_amr_parser, while a change to components.amr.seq2seq.seq2seq_amr_parser can't break common.torch_component.
Position
Below the diagonal: components.amr.seq2seq.seq2seq_amr_parser uses a module listed above it, which is the healthy direction.
components.amr.seq2seq.seq2seq_amr_parser uses common.dataset, so a change to common.dataset can break components.amr.seq2seq.seq2seq_amr_parser, while a change to components.amr.seq2seq.seq2seq_amr_parser can't break common.dataset.
Position
Below the diagonal: components.amr.seq2seq.seq2seq_amr_parser uses a module listed above it, which is the healthy direction.
components.mtl.tasks uses common.structure, so a change to common.structure can break components.mtl.tasks, while a change to components.mtl.tasks can't break common.structure.
Position
Below the diagonal: components.mtl.tasks uses a module listed above it, which is the healthy direction.
components.mtl.tasks uses transform.transformer_tokenizer, so a change to transform.transformer_tokenizer can break components.mtl.tasks, while a change to components.mtl.tasks can't break transform.transformer_tokenizer.
Position
Below the diagonal: components.mtl.tasks uses a module listed above it, which is the healthy direction.
components.mtl.tasks uses common.torch_component, so a change to common.torch_component can break components.mtl.tasks, while a change to components.mtl.tasks can't break common.torch_component.
Position
Below the diagonal: components.mtl.tasks uses a module listed above it, which is the healthy direction.
components.mtl.tasks uses layers.scalar_mix, so a change to layers.scalar_mix can break components.mtl.tasks, while a change to components.mtl.tasks can't break layers.scalar_mix.
Position
Below the diagonal: components.mtl.tasks uses a module listed above it, which is the healthy direction.
components.mtl.tasks uses common.dataset, so a change to common.dataset can break components.mtl.tasks, while a change to components.mtl.tasks can't break common.dataset.
Position
Below the diagonal: components.mtl.tasks uses a module listed above it, which is the healthy direction.
components.parsers.biaffine.biaffine_dep uses common.structure, so a change to common.structure can break components.parsers.biaffine.biaffine_dep, while a change to components.parsers.biaffine.biaffine_dep can't break common.structure.
Position
Below the diagonal: components.parsers.biaffine.biaffine_dep uses a module listed above it, which is the healthy direction.
components.parsers.biaffine.biaffine_dep uses common.torch_component, so a change to common.torch_component can break components.parsers.biaffine.biaffine_dep, while a change to components.parsers.biaffine.biaffine_dep can't break common.torch_component.
Position
Below the diagonal: components.parsers.biaffine.biaffine_dep uses a module listed above it, which is the healthy direction.
components.parsers.biaffine.biaffine_dep uses common.dataset, so a change to common.dataset can break components.parsers.biaffine.biaffine_dep, while a change to components.parsers.biaffine.biaffine_dep can't break common.dataset.
Position
Below the diagonal: components.parsers.biaffine.biaffine_dep uses a module listed above it, which is the healthy direction.
components.parsers.constituency.crf_constituency_parser uses common.structure, so a change to common.structure can break components.parsers.constituency.crf_constituency_parser, while a change to components.parsers.constituency.crf_constituency_parser can't break common.structure.
Position
Below the diagonal: components.parsers.constituency.crf_constituency_parser uses a module listed above it, which is the healthy direction.
components.parsers.constituency.crf_constituency_parser uses common.torch_component, so a change to common.torch_component can break components.parsers.constituency.crf_constituency_parser, while a change to components.parsers.constituency.crf_constituency_parser can't break common.torch_component.
Position
Below the diagonal: components.parsers.constituency.crf_constituency_parser uses a module listed above it, which is the healthy direction.
components.parsers.constituency.crf_constituency_parser uses common.dataset, so a change to common.dataset can break components.parsers.constituency.crf_constituency_parser, while a change to components.parsers.constituency.crf_constituency_parser can't break common.dataset.
Position
Below the diagonal: components.parsers.constituency.crf_constituency_parser uses a module listed above it, which is the healthy direction.
components.srl.span_bio.span_bio uses common.structure, so a change to common.structure can break components.srl.span_bio.span_bio, while a change to components.srl.span_bio.span_bio can't break common.structure.
Position
Below the diagonal: components.srl.span_bio.span_bio uses a module listed above it, which is the healthy direction.
components.srl.span_bio.span_bio uses layers.embeddings.embedding, so a change to layers.embeddings.embedding can break components.srl.span_bio.span_bio, while a change to components.srl.span_bio.span_bio can't break layers.embeddings.embedding.
Position
Below the diagonal: components.srl.span_bio.span_bio uses a module listed above it, which is the healthy direction.
components.srl.span_bio.span_bio uses common.torch_component, so a change to common.torch_component can break components.srl.span_bio.span_bio, while a change to components.srl.span_bio.span_bio can't break common.torch_component.
Position
Below the diagonal: components.srl.span_bio.span_bio uses a module listed above it, which is the healthy direction.
components.srl.span_bio.span_bio uses common.dataset, so a change to common.dataset can break components.srl.span_bio.span_bio, while a change to components.srl.span_bio.span_bio can't break common.dataset.
Position
Below the diagonal: components.srl.span_bio.span_bio uses a module listed above it, which is the healthy direction.
components.taggers.rnn_tagger_tf uses common.transform_tf, so a change to common.transform_tf can break components.taggers.rnn_tagger_tf, while a change to components.taggers.rnn_tagger_tf can't break common.transform_tf.
Position
Below the diagonal: components.taggers.rnn_tagger_tf uses a module listed above it, which is the healthy direction.
components.taggers.rnn_tagger_tf uses common.vocab_tf, so a change to common.vocab_tf can break components.taggers.rnn_tagger_tf, while a change to components.taggers.rnn_tagger_tf can't break common.vocab_tf.
Position
Below the diagonal: components.taggers.rnn_tagger_tf uses a module listed above it, which is the healthy direction.
components.taggers.rnn_tagger_tf uses components.taggers.tagger_tf, so a change to components.taggers.tagger_tf can break components.taggers.rnn_tagger_tf, while a change to components.taggers.rnn_tagger_tf can't break components.taggers.tagger_tf.
Position
Below the diagonal: components.taggers.rnn_tagger_tf uses a module listed above it, which is the healthy direction.
components.taggers.transformers.transformer_tagger uses common.structure, so a change to common.structure can break components.taggers.transformers.transformer_tagger, while a change to components.taggers.transformers.transformer_tagger can't break common.structure.
Position
Below the diagonal: components.taggers.transformers.transformer_tagger uses a module listed above it, which is the healthy direction.
components.taggers.transformers.transformer_tagger uses layers.embeddings.embedding, so a change to layers.embeddings.embedding can break components.taggers.transformers.transformer_tagger, while a change to components.taggers.transformers.transformer_tagger can't break layers.embeddings.embedding.
Position
Below the diagonal: components.taggers.transformers.transformer_tagger uses a module listed above it, which is the healthy direction.
components.taggers.transformers.transformer_tagger uses transform.transformer_tokenizer, so a change to transform.transformer_tokenizer can break components.taggers.transformers.transformer_tagger, while a change to components.taggers.transformers.transformer_tagger can't break transform.transformer_tokenizer.
Position
Below the diagonal: components.taggers.transformers.transformer_tagger uses a module listed above it, which is the healthy direction.
components.taggers.transformers.transformer_tagger uses common.dataset, so a change to common.dataset can break components.taggers.transformers.transformer_tagger, while a change to components.taggers.transformers.transformer_tagger can't break common.dataset.
Position
Below the diagonal: components.taggers.transformers.transformer_tagger uses a module listed above it, which is the healthy direction.
components.taggers.transformers.transformer_tagger_tf uses common.transform_tf, so a change to common.transform_tf can break components.taggers.transformers.transformer_tagger_tf, while a change to components.taggers.transformers.transformer_tagger_tf can't break common.transform_tf.
Position
Below the diagonal: components.taggers.transformers.transformer_tagger_tf uses a module listed above it, which is the healthy direction.
components.taggers.transformers.transformer_tagger_tf uses components.taggers.tagger_tf, so a change to components.taggers.tagger_tf can break components.taggers.transformers.transformer_tagger_tf, while a change to components.taggers.transformers.transformer_tagger_tf can't break components.taggers.tagger_tf.
Position
Below the diagonal: components.taggers.transformers.transformer_tagger_tf uses a module listed above it, which is the healthy direction.
1 pair of types, where a type in components.taggers.transformers.transformer_tagger_tf uses a type in components.taggers.transformers.transformer_transform_tf.
components.taggers.transformers.transformer_tagger_tf uses components.taggers.transformers.transformer_transform_tf, so a change to components.taggers.transformers.transformer_transform_tf can break components.taggers.transformers.transformer_tagger_tf, while a change to components.taggers.transformers.transformer_tagger_tf can't break components.taggers.transformers.transformer_transform_tf.
Position
Below the diagonal: components.taggers.transformers.transformer_tagger_tf uses a module listed above it, which is the healthy direction.
layers.embeddings.word2vec uses layers.embeddings.embedding, so a change to layers.embeddings.embedding can break layers.embeddings.word2vec, while a change to layers.embeddings.word2vec can't break layers.embeddings.embedding.
Position
Below the diagonal: layers.embeddings.word2vec uses a module listed above it, which is the healthy direction.
layers.embeddings.word2vec uses common.torch_component, so a change to common.torch_component can break layers.embeddings.word2vec, while a change to layers.embeddings.word2vec can't break common.torch_component.
Position
Below the diagonal: layers.embeddings.word2vec uses a module listed above it, which is the healthy direction.
layers.embeddings.word2vec uses common.transform, so a change to common.transform can break layers.embeddings.word2vec, while a change to layers.embeddings.word2vec can't break common.transform.
Position
Below the diagonal: layers.embeddings.word2vec uses a module listed above it, which is the healthy direction.
layers.embeddings.word2vec uses common.dataset, so a change to common.dataset can break layers.embeddings.word2vec, while a change to layers.embeddings.word2vec can't break common.dataset.
Position
Below the diagonal: layers.embeddings.word2vec uses a module listed above it, which is the healthy direction.
components.mtl.multi_task_learning uses common.structure, so a change to common.structure can break components.mtl.multi_task_learning, while a change to components.mtl.multi_task_learning can't break common.structure.
Position
Below the diagonal: components.mtl.multi_task_learning uses a module listed above it, which is the healthy direction.
components.mtl.multi_task_learning uses layers.embeddings.embedding, so a change to layers.embeddings.embedding can break components.mtl.multi_task_learning, while a change to components.mtl.multi_task_learning can't break layers.embeddings.embedding.
Position
Below the diagonal: components.mtl.multi_task_learning uses a module listed above it, which is the healthy direction.
components.mtl.multi_task_learning uses common.torch_component, so a change to common.torch_component can break components.mtl.multi_task_learning, while a change to components.mtl.multi_task_learning can't break common.torch_component.
Position
Below the diagonal: components.mtl.multi_task_learning uses a module listed above it, which is the healthy direction.
components.mtl.multi_task_learning uses components.mtl.tasks, so a change to components.mtl.tasks can break components.mtl.multi_task_learning, while a change to components.mtl.multi_task_learning can't break components.mtl.tasks.
Position
Below the diagonal: components.mtl.multi_task_learning uses a module listed above it, which is the healthy direction.
components.mtl.tasks.constituency uses layers.scalar_mix, so a change to layers.scalar_mix can break components.mtl.tasks.constituency, while a change to components.mtl.tasks.constituency can't break layers.scalar_mix.
Position
Below the diagonal: components.mtl.tasks.constituency uses a module listed above it, which is the healthy direction.
components.mtl.tasks.constituency uses common.dataset, so a change to common.dataset can break components.mtl.tasks.constituency, while a change to components.mtl.tasks.constituency can't break common.dataset.
Position
Below the diagonal: components.mtl.tasks.constituency uses a module listed above it, which is the healthy direction.
components.mtl.tasks.constituency uses components.mtl.tasks, so a change to components.mtl.tasks can break components.mtl.tasks.constituency, while a change to components.mtl.tasks.constituency can't break components.mtl.tasks.
Position
Below the diagonal: components.mtl.tasks.constituency uses a module listed above it, which is the healthy direction.
components.mtl.tasks.constituency uses components.parsers.constituency.crf_constituency_parser, so a change to components.parsers.constituency.crf_constituency_parser can break components.mtl.tasks.constituency, while a change to components.mtl.tasks.constituency can't break components.parsers.constituency.crf_constituency_parser.
Position
Below the diagonal: components.mtl.tasks.constituency uses a module listed above it, which is the healthy direction.
components.mtl.tasks.dep uses common.transform, so a change to common.transform can break components.mtl.tasks.dep, while a change to components.mtl.tasks.dep can't break common.transform.
Position
Below the diagonal: components.mtl.tasks.dep uses a module listed above it, which is the healthy direction.
components.mtl.tasks.dep uses layers.scalar_mix, so a change to layers.scalar_mix can break components.mtl.tasks.dep, while a change to components.mtl.tasks.dep can't break layers.scalar_mix.
Position
Below the diagonal: components.mtl.tasks.dep uses a module listed above it, which is the healthy direction.
components.mtl.tasks.dep uses common.dataset, so a change to common.dataset can break components.mtl.tasks.dep, while a change to components.mtl.tasks.dep can't break common.dataset.
Position
Below the diagonal: components.mtl.tasks.dep uses a module listed above it, which is the healthy direction.
components.mtl.tasks.dep uses components.mtl.tasks, so a change to components.mtl.tasks can break components.mtl.tasks.dep, while a change to components.mtl.tasks.dep can't break components.mtl.tasks.
Position
Below the diagonal: components.mtl.tasks.dep uses a module listed above it, which is the healthy direction.
components.mtl.tasks.dep uses components.parsers.biaffine.biaffine_dep, so a change to components.parsers.biaffine.biaffine_dep can break components.mtl.tasks.dep, while a change to components.mtl.tasks.dep can't break components.parsers.biaffine.biaffine_dep.
Position
Below the diagonal: components.mtl.tasks.dep uses a module listed above it, which is the healthy direction.
components.mtl.tasks.dep_2nd uses layers.scalar_mix, so a change to layers.scalar_mix can break components.mtl.tasks.dep_2nd, while a change to components.mtl.tasks.dep_2nd can't break layers.scalar_mix.
Position
Below the diagonal: components.mtl.tasks.dep_2nd uses a module listed above it, which is the healthy direction.
components.mtl.tasks.dep_2nd uses common.dataset, so a change to common.dataset can break components.mtl.tasks.dep_2nd, while a change to components.mtl.tasks.dep_2nd can't break common.dataset.
Position
Below the diagonal: components.mtl.tasks.dep_2nd uses a module listed above it, which is the healthy direction.
components.mtl.tasks.dep_2nd uses components.mtl.tasks, so a change to components.mtl.tasks can break components.mtl.tasks.dep_2nd, while a change to components.mtl.tasks.dep_2nd can't break components.mtl.tasks.
Position
Below the diagonal: components.mtl.tasks.dep_2nd uses a module listed above it, which is the healthy direction.
components.mtl.tasks.lem uses layers.scalar_mix, so a change to layers.scalar_mix can break components.mtl.tasks.lem, while a change to components.mtl.tasks.lem can't break layers.scalar_mix.
Position
Below the diagonal: components.mtl.tasks.lem uses a module listed above it, which is the healthy direction.
components.mtl.tasks.lem uses common.dataset, so a change to common.dataset can break components.mtl.tasks.lem, while a change to components.mtl.tasks.lem can't break common.dataset.
Position
Below the diagonal: components.mtl.tasks.lem uses a module listed above it, which is the healthy direction.
components.mtl.tasks.lem uses components.mtl.tasks, so a change to components.mtl.tasks can break components.mtl.tasks.lem, while a change to components.mtl.tasks.lem can't break components.mtl.tasks.
Position
Below the diagonal: components.mtl.tasks.lem uses a module listed above it, which is the healthy direction.
components.mtl.tasks.ner.biaffine_ner uses common.transform, so a change to common.transform can break components.mtl.tasks.ner.biaffine_ner, while a change to components.mtl.tasks.ner.biaffine_ner can't break common.transform.
Position
Below the diagonal: components.mtl.tasks.ner.biaffine_ner uses a module listed above it, which is the healthy direction.
components.mtl.tasks.ner.biaffine_ner uses layers.scalar_mix, so a change to layers.scalar_mix can break components.mtl.tasks.ner.biaffine_ner, while a change to components.mtl.tasks.ner.biaffine_ner can't break layers.scalar_mix.
Position
Below the diagonal: components.mtl.tasks.ner.biaffine_ner uses a module listed above it, which is the healthy direction.
components.mtl.tasks.ner.biaffine_ner uses common.dataset, so a change to common.dataset can break components.mtl.tasks.ner.biaffine_ner, while a change to components.mtl.tasks.ner.biaffine_ner can't break common.dataset.
Position
Below the diagonal: components.mtl.tasks.ner.biaffine_ner uses a module listed above it, which is the healthy direction.
components.mtl.tasks.ner.biaffine_ner uses components.ner.biaffine_ner.biaffine_ner, so a change to components.ner.biaffine_ner.biaffine_ner can break components.mtl.tasks.ner.biaffine_ner, while a change to components.mtl.tasks.ner.biaffine_ner can't break components.ner.biaffine_ner.biaffine_ner.
Position
Below the diagonal: components.mtl.tasks.ner.biaffine_ner uses a module listed above it, which is the healthy direction.
components.mtl.tasks.ner.biaffine_ner uses components.mtl.tasks, so a change to components.mtl.tasks can break components.mtl.tasks.ner.biaffine_ner, while a change to components.mtl.tasks.ner.biaffine_ner can't break components.mtl.tasks.
Position
Below the diagonal: components.mtl.tasks.ner.biaffine_ner uses a module listed above it, which is the healthy direction.
components.mtl.tasks.pos uses layers.scalar_mix, so a change to layers.scalar_mix can break components.mtl.tasks.pos, while a change to components.mtl.tasks.pos can't break layers.scalar_mix.
Position
Below the diagonal: components.mtl.tasks.pos uses a module listed above it, which is the healthy direction.
components.mtl.tasks.pos uses common.dataset, so a change to common.dataset can break components.mtl.tasks.pos, while a change to components.mtl.tasks.pos can't break common.dataset.
Position
Below the diagonal: components.mtl.tasks.pos uses a module listed above it, which is the healthy direction.
components.mtl.tasks.pos uses components.mtl.tasks, so a change to components.mtl.tasks can break components.mtl.tasks.pos, while a change to components.mtl.tasks.pos can't break components.mtl.tasks.
Position
Below the diagonal: components.mtl.tasks.pos uses a module listed above it, which is the healthy direction.
components.mtl.tasks.pos uses components.taggers.transformers.transformer_tagger, so a change to components.taggers.transformers.transformer_tagger can break components.mtl.tasks.pos, while a change to components.mtl.tasks.pos can't break components.taggers.transformers.transformer_tagger.
Position
Below the diagonal: components.mtl.tasks.pos uses a module listed above it, which is the healthy direction.
components.mtl.tasks.sdp uses common.transform, so a change to common.transform can break components.mtl.tasks.sdp, while a change to components.mtl.tasks.sdp can't break common.transform.
Position
Below the diagonal: components.mtl.tasks.sdp uses a module listed above it, which is the healthy direction.
components.mtl.tasks.sdp uses layers.scalar_mix, so a change to layers.scalar_mix can break components.mtl.tasks.sdp, while a change to components.mtl.tasks.sdp can't break layers.scalar_mix.
Position
Below the diagonal: components.mtl.tasks.sdp uses a module listed above it, which is the healthy direction.
components.mtl.tasks.sdp uses common.dataset, so a change to common.dataset can break components.mtl.tasks.sdp, while a change to components.mtl.tasks.sdp can't break common.dataset.
Position
Below the diagonal: components.mtl.tasks.sdp uses a module listed above it, which is the healthy direction.
components.mtl.tasks.sdp uses components.mtl.tasks, so a change to components.mtl.tasks can break components.mtl.tasks.sdp, while a change to components.mtl.tasks.sdp can't break components.mtl.tasks.
Position
Below the diagonal: components.mtl.tasks.sdp uses a module listed above it, which is the healthy direction.
components.mtl.tasks.tok.reg_tok uses common.transform, so a change to common.transform can break components.mtl.tasks.tok.reg_tok, while a change to components.mtl.tasks.tok.reg_tok can't break common.transform.
Position
Below the diagonal: components.mtl.tasks.tok.reg_tok uses a module listed above it, which is the healthy direction.
components.mtl.tasks.tok.reg_tok uses layers.scalar_mix, so a change to layers.scalar_mix can break components.mtl.tasks.tok.reg_tok, while a change to components.mtl.tasks.tok.reg_tok can't break layers.scalar_mix.
Position
Below the diagonal: components.mtl.tasks.tok.reg_tok uses a module listed above it, which is the healthy direction.
components.mtl.tasks.tok.reg_tok uses common.dataset, so a change to common.dataset can break components.mtl.tasks.tok.reg_tok, while a change to components.mtl.tasks.tok.reg_tok can't break common.dataset.
Position
Below the diagonal: components.mtl.tasks.tok.reg_tok uses a module listed above it, which is the healthy direction.
components.mtl.tasks.tok.reg_tok uses components.mtl.tasks, so a change to components.mtl.tasks can break components.mtl.tasks.tok.reg_tok, while a change to components.mtl.tasks.tok.reg_tok can't break components.mtl.tasks.
Position
Below the diagonal: components.mtl.tasks.tok.reg_tok uses a module listed above it, which is the healthy direction.
components.mtl.tasks.tok.tag_tok uses transform.transformer_tokenizer, so a change to transform.transformer_tokenizer can break components.mtl.tasks.tok.tag_tok, while a change to components.mtl.tasks.tok.tag_tok can't break transform.transformer_tokenizer.
Position
Below the diagonal: components.mtl.tasks.tok.tag_tok uses a module listed above it, which is the healthy direction.
components.mtl.tasks.tok.tag_tok uses common.transform, so a change to common.transform can break components.mtl.tasks.tok.tag_tok, while a change to components.mtl.tasks.tok.tag_tok can't break common.transform.
Position
Below the diagonal: components.mtl.tasks.tok.tag_tok uses a module listed above it, which is the healthy direction.
components.mtl.tasks.tok.tag_tok uses layers.scalar_mix, so a change to layers.scalar_mix can break components.mtl.tasks.tok.tag_tok, while a change to components.mtl.tasks.tok.tag_tok can't break layers.scalar_mix.
Position
Below the diagonal: components.mtl.tasks.tok.tag_tok uses a module listed above it, which is the healthy direction.
components.mtl.tasks.tok.tag_tok uses common.dataset, so a change to common.dataset can break components.mtl.tasks.tok.tag_tok, while a change to components.mtl.tasks.tok.tag_tok can't break common.dataset.
Position
Below the diagonal: components.mtl.tasks.tok.tag_tok uses a module listed above it, which is the healthy direction.
components.mtl.tasks.tok.tag_tok uses components.mtl.tasks, so a change to components.mtl.tasks can break components.mtl.tasks.tok.tag_tok, while a change to components.mtl.tasks.tok.tag_tok can't break components.mtl.tasks.
Position
Below the diagonal: components.mtl.tasks.tok.tag_tok uses a module listed above it, which is the healthy direction.
components.ner.ner_tf uses common.transform_tf, so a change to common.transform_tf can break components.ner.ner_tf, while a change to components.ner.ner_tf can't break common.transform_tf.
Position
Below the diagonal: components.ner.ner_tf uses a module listed above it, which is the healthy direction.
components.ner.ner_tf uses common.keras_component, so a change to common.keras_component can break components.ner.ner_tf, while a change to components.ner.ner_tf can't break common.keras_component.
Position
Below the diagonal: components.ner.ner_tf uses a module listed above it, which is the healthy direction.
components.ner.ner_tf uses components.taggers.transformers.transformer_transform_tf, so a change to components.taggers.transformers.transformer_transform_tf can break components.ner.ner_tf, while a change to components.ner.ner_tf can't break components.taggers.transformers.transformer_transform_tf.
Position
Below the diagonal: components.ner.ner_tf uses a module listed above it, which is the healthy direction.
components.ner.ner_tf uses components.taggers.rnn_tagger_tf, so a change to components.taggers.rnn_tagger_tf can break components.ner.ner_tf, while a change to components.ner.ner_tf can't break components.taggers.rnn_tagger_tf.
Position
Below the diagonal: components.ner.ner_tf uses a module listed above it, which is the healthy direction.
components.ner.ner_tf uses components.taggers.transformers.transformer_tagger_tf, so a change to components.taggers.transformers.transformer_tagger_tf can break components.ner.ner_tf, while a change to components.ner.ner_tf can't break components.taggers.transformers.transformer_tagger_tf.
Position
Below the diagonal: components.ner.ner_tf uses a module listed above it, which is the healthy direction.
components.tokenizers.tok_tf uses common.keras_component, so a change to common.keras_component can break components.tokenizers.tok_tf, while a change to components.tokenizers.tok_tf can't break common.keras_component.
Position
Below the diagonal: components.tokenizers.tok_tf uses a module listed above it, which is the healthy direction.
components.tokenizers.tok_tf uses transform.tsv_tf, so a change to transform.tsv_tf can break components.tokenizers.tok_tf, while a change to components.tokenizers.tok_tf can't break transform.tsv_tf.
Position
Below the diagonal: components.tokenizers.tok_tf uses a module listed above it, which is the healthy direction.
components.tokenizers.tok_tf uses components.taggers.transformers.transformer_transform_tf, so a change to components.taggers.transformers.transformer_transform_tf can break components.tokenizers.tok_tf, while a change to components.tokenizers.tok_tf can't break components.taggers.transformers.transformer_transform_tf.
Position
Below the diagonal: components.tokenizers.tok_tf uses a module listed above it, which is the healthy direction.
components.tokenizers.tok_tf uses components.taggers.rnn_tagger_tf, so a change to components.taggers.rnn_tagger_tf can break components.tokenizers.tok_tf, while a change to components.tokenizers.tok_tf can't break components.taggers.rnn_tagger_tf.
Position
Below the diagonal: components.tokenizers.tok_tf uses a module listed above it, which is the healthy direction.
components.tokenizers.tok_tf uses components.taggers.transformers.transformer_tagger_tf, so a change to components.taggers.transformers.transformer_tagger_tf can break components.tokenizers.tok_tf, while a change to components.tokenizers.tok_tf can't break components.taggers.transformers.transformer_tagger_tf.
Position
Below the diagonal: components.tokenizers.tok_tf uses a module listed above it, which is the healthy direction.
At a glance — Code Health · 46% · Weak · gated by D2, X4, X5 ·
Findings mapped to OWASP categories; the specific CVEs/secrets are in the Security dimension cards below and findings.md (redacted only on the public version of this report).
OWASP category
Findings
Severity
A05:2025 — Injection
18
High / Critical
A03:2025 — Software Supply Chain Failures
5
High / Critical
A07:2025 — Authentication Failures
2
High / Critical
Roadmap
First, integrate static security analysis into the CI pipeline to prevent security regressions from reaching production. Second, enforce nullable reference types across Java packages using NullAway or the Checker Framework, resolving suppressions by addressing actual null risks rather than hiding them. Third, replace interpolated log messages with structured logging to improve observability. Fourth, document significant design decisions in a dedicated ADR folder to capture context and consequences. Finally, eliminate swallowed exceptions by ensuring all caught errors are properly handled or logged.
Ranked by impact ÷ effort. "Helps" is the estimated gain on the 0–100 health score.
Do this
Helps
Effort
Dimension
Add a SAST step to CI running what this repository's stack ships: bandit, `semgrep --config=p/python`, or CodeQL's python pack — so a security regression fails the build instead of landing.
Declare the default where Java declares it: `@NullMarked` (org.jspecify.annotations) in each package's `package-info.java`, or once in `module-info.java`, and run NullAway (Error Prone) or the Checker Framework so the declaration is enforced. Then annotate the real exceptions `@Nullable` and resolve the suppressions rather than adding to them: a `@SuppressWarnings("NullAway")` or `castToNonNull(x)` on a value the checker could show to be null is a `NullPointerException` waiting for the path that makes it true — where the value genuinely cannot be null, say something the checker can see instead (a null check, `Objects.requireNonNull`, a non-null field initialised in the constructor).
Record your significant design decisions. Write one short document per decision, giving its date, the situation, the decision and its consequences, and keep them together with your other design documents. A docs/adr folder with each file named by number and title is the easiest place to find them.
Nothing pauses a release for a human: publish as a draft release (or gate the release job on a protected tag/manual dispatch) so a bad build can be stopped before users can download it.
Every finding carries one of four grades. Three say how serious it is. The fourth says this
survey could not settle it — and it is a grade, not a gap.
Critical — 14
A definite problem that already costs you something and drags the score down: a
missing authorisation check, a dependency with a known exploit, a build that does not reproduce. Failure here
tends to cause failures elsewhere.
Serious — 389
Likely wrong, but not failing yet. It degrades
the codebase over a longer horizon and can cause failures elsewhere — not urgent this week, not something to
carry for two years either.
Minor — 77
Recorded, with no effect on how the codebase functions.
Present so the survey is complete, not because it needs doing.
Could not be resolved — 11
Something this survey could not settle
from the outside, and which could be critical or serious. Either a control was required and no
positive evidence of it exists in the repository — a backup job that nothing shows was ever restored from proves
nothing about restores — or our own analysis could not run over that part of the tree. This is not a clean
result. These are excluded from the score rather than awarded a pass, so the number on the cover neither
rewards nor penalises them: if you act on this survey without resolving them, you carry that risk yourself. Each
one is named under Limitations.
Methodology & how to trust this report
Watchdog is a deep, periodic assessment — run each sprint, monthly, or quarterly, taking the time to go wider and deeper than a quick check and surfacing in one coherent report what you'd otherwise piece together from a dozen separate tools. It scores deterministically: the same commit yields the same score, every run. 57 of 57 evaluated dimensions are computed purely by tools and static analysis (confidence 1.0); none rely on LLM judgement. Overall confidence is 0.9 — the weighted average across measured dimensions; it falls as more of the score leans on LLM-assisted judgement and rises when it's fully tool-backed.
Every figure here is one of three kinds, and we label which: ✓ Measured — a deterministic fact (LoC, complexity, coverage); ~ Modeled — an estimate from a stated model (cost, effort, value-at-risk), always a range with its assumptions, never a precise fact; ◐ Advisory — an LLM prose judgement. We never present a modelled estimate as if it were measured. Perfect or absent scores carry their provenance too (ADR-0011): ✓ Tool-verified means the property itself was measured across the surface; ○ Nothing flagged means the probes came back clean — a claim bounded by what a repository can show; ⊘ Not evidenced means a working control (a tested restore, an automated rollback) showed no positive evidence — absence of evidence is not evidence of a control, so it's excluded from the score rather than awarded a spurious 10; ◐ Sampled · advisory marks an LLM verdict over a bounded sample — advisory, never a deterministic measurement.
What we checked — 57 dimensions across the health lenses
Each chip is a dimension scored from real signals across architecture, testing, dependencies, security & compliance, documentation, git-history and code quality — in one coherent pass. A surface report typically covers a handful.
How to trust any code-health report — three questions
Can you open the finding? Real findings cite a repo-relative file and line you can open at the cited line — never an absolute scratch path. Here, 467 of 480 do; the remainder are repo-wide signals — a dimension-level measurement, not a single line. (Every path in this report is repo-relative by construction: paths are normalized at the producer and the report is rejected if any rooted path leaks through.)
Is there a tool behind the number? Every score below names the method that produced it — Roslyn, git, a scanner, or (for a handful of documentation/naming dimensions) an LLM labelled sampled · advisory — not a narrative.
Does re-running give the same result? Run it again on the same commit and the score — and this report, byte for byte — is identical. A report whose numbers move between runs is describing the run, not the code.
This report answers yes to all three. That's the bar to hold any assessment to.
Tools & methods
The actual versions used this run (captured at analysis time) — re-run on the same commit for the identical score.
Method
Backs
Version
Evaluator
Roslyn static analysis
Complexity, cohesion, coupling, dead code, API surface, layering
What ran differently this time — a tool absent, degraded, or that fell back to an estimate. Named openly, not folded silently into the scores. A degraded run also records its exact cause in diagnostics.md.
D8 Code Coverage — not measured this run — Watchdog could not measure this here. That is a gap on our side — a collector, parser or image we have not built yet — and it is neither a defect in this repository nor evidence that the check would have passed. Coverage NOT MEASURED: test source is present (.py) but the built-in coverage collector has no runner for this repository's ecosystem — so this suite was never executed by it. Not scored — this is a gap in the analyzer's language coverage, not a defect in the repo. To have real coverage read, produce a coverage report in a standard format (`coverage run -m pytest` then `coverage xml`) and commit it — a hosted scan measures a clone of the repository, so a report that exists only in a working tree, a CI runner's or your own, never reaches it; the artefact is commonly gitignored, so `git add -f` that one file (or un-ignore its path) and commit it alongside the code it measures. You can widen what we reach: optional: produce a coverage report in a standard format (`coverage run -m pytest` then `coverage xml`) and commit it — a hosted scan measures a clone of the repository, so a report that exists only in a working tree, a CI runner's or your own, never reaches it; the artefact is commonly gitignored, so `git add -f` that one file (or un-ignore its path) and commit it alongside the code it measures — then the real number is read on the next scan.
D12 Dependency Hygiene — not measured this run — Watchdog could not measure this here. That is a gap on our side — a collector, parser or image we have not built yet — and it is neither a defect in this repository nor evidence that the check would have passed. Not scored — 1 Maven/Gradle manifest(s) were read, stating a version for 2 dependency(ies) and leaving 0 to a parent or an imported BOM, but no build resolved the dependency graph, so the outdated signal was not established. NOT a finding that these dependencies are current or healthy.
D16 Bus Factor — not measured this run — This is a true statement about the repository that carries nothing for its owner to act on, so it is reported here rather than as a defect in their code. Single-maintainer repository — bus factor is not applicable (47 contributor(s) across 1812 commit(s) sampled, automation and bot accounts excluded). One of them holds 94% of the history; the other 46 hold 0.1% each on average, below the 5% at which there is somebody to hand the work to. That is a single maintainer with drive-by contributors, not a team whose knowledge has concentrated — so the bus factor is not applicable and there is nothing here for the owner to act on.
D19 Documentation Quality — measured, with a gap in what it reached — Watchdog measured this, but not all of it. What it did not reach is a gap on our side — a collector, parser or image we have not built yet — so the numbers on that dimension cover less than the repository, and the part left out is not evidence that it would have passed. Each ecosystem's manifests (what makes a folder runnable or a published package) need their own proof before D19 judges its units. Until then those folders' READMEs are judged only as documents (boilerplate, structure), never counted as complete.
D21 Naming Consistency — not measured this run — Watchdog could not measure this here. That is a gap on our side — a collector, parser or image we have not built yet — and it is neither a defect in this repository nor evidence that the check would have passed. D21's deterministic naming rules are gated on C# first (docs/plans/deterministic-engine.md, P5): each other language is switched on only after its own concept suite and A/B gate pass. Until then its declarations carry no naming verdict this run — this says nothing about whether its names are consistent.
D22 Internal API Consistency — not measured this run — Watchdog could not measure this here. That is a gap on our side — a collector, parser or image we have not built yet — and it is neither a defect in this repository nor evidence that the check would have passed. D22 identifies the intentionally-exposed surface from `IsPackable` and `.Contracts` project names, MSBuild conventions read off the loaded project set. This target exposed no such projects, so the probe never ran; this says nothing about whether the repository has a public API. This repository commits no C#/VB source at all, so there was never an MSBuild project set to read these conventions off. That is OUR side and it is an ADMISSION gap, not an environment fault: it declares a published package (REDACTED), but no published-package marker D22 reads admitted any project here. D22 DOES collect the public API of every language in this tree, and admits a project only when its manifest makes that ecosystem's publication act in a form the marker recognises — so what is missing is not a collector but an admission rule for the publication act this repository makes, if it makes one. The remedy is a better discriminator for that marker — not a collector, and no change to the scan image can close it. The marker's rules are measured to refuse applications, so one refused repository is the rule working as calibrated; several are the evidence that it is too strict.
D24 Comment Value — not measured this run — Watchdog could not measure this here. That is a gap on our side — a collector, parser or image we have not built yet — and it is neither a defect in this repository nor evidence that the check would have passed. This check's reader does not cover the language this repository's product is written in, so it had nothing of the product to read. That is a gap in this analyzer's language reach — not a finding about this repository.
D27 Navigability — not measured this run — Watchdog could not measure this here. That is a gap on our side — a collector, parser or image we have not built yet — and it is neither a defect in this repository nor evidence that the check would have passed. Navigability is taken over call sites whose callee resolved. For C# that population includes calls into referenced assemblies; for this language only calls declared inside the repository count, so the abstention threshold reads the repository's use of its standard library as a resolution failure. That is a limit of this analyzer, not a build problem in this repository.
C1 Data Protection — not measured this run — Watchdog could not measure this here. That is a gap on our side — a collector, parser or image we have not built yet — and it is neither a defect in this repository nor evidence that the check would have passed. These personal data controls are read from declarative annotations, request middleware, entity/column names and guard methods in a C# source model, and none was loaded on this run, so there was nothing to gather. That is a gap in this analyzer's language reach — not a finding that the repository lacks personal data controls.
GD1 Unfinished & placeholder code — not measured this run — Watchdog could not measure this here. That is a gap on our side — a collector, parser or image we have not built yet — and it is neither a defect in this repository nor evidence that the check would have passed. This check reads C# syntax, F#, JavaScript/TypeScript, Dart, Ruby and Scala source only, and no C#, F#, JavaScript/TypeScript, Dart, Ruby or Scala was found, and this repository's Java, Python is not read yet, so it had nothing of this repository's product to read. That is a gap in this analyzer's language reach — not a finding that the repository is free of what this check looks for.
IC1 Incompleteness & stubs — not measured this run — Watchdog could not measure this here. That is a gap on our side — a collector, parser or image we have not built yet — and it is neither a defect in this repository nor evidence that the check would have passed. This check reads C# syntax, JavaScript/TypeScript, Java, Dart, Ruby and Scala source only, and this repository's Python is most of its product and is not read yet, so it had nothing of this repository's product to read. That is a gap in this analyzer's language reach — not a finding that the repository is free of what this check looks for.
P5 DR & Backup — not measured this run — This is a true statement about the repository that carries nothing for its owner to act on, so it is reported here rather than as a defect in their code. No backup/snapshot/replication config, RTO/RPO or restore-procedure documentation was found — and no production persistence was detected either (no data-access packages, no data-store services, no database resources), so there is nothing in this repository whose loss a DR control would recover. If this system's data lives in a platform or ops repo we can't see, that's where the DR evidence belongs.
S1 Web-Security Posture — not measured this run — Watchdog could not measure this here. That is a gap on our side — a collector, parser or image we have not built yet — and it is neither a defect in this repository nor evidence that the check would have passed. These web-security controls are read from declarative annotations, request middleware, entity/column names and guard methods in a C# source model, and none was loaded on this run, so there was nothing to gather. That is a gap in this analyzer's language reach — not a finding that the repository lacks web-security controls.
X4 Structured logging — measured, with a gap in what it reached — Watchdog measured this, but not all of it. What it did not reach is a gap on our side — a collector, parser or image we have not built yet — so the numbers on that dimension cover less than the repository, and the part left out is not evidence that it would have passed. 16 further occurrence(s) are not listed individually; the score already reflects all 56.
Repo exclusion declarations (.gitattributes linguist-generated/vendored, .editorconfig generated_code): none declared — every source file was scored.
Limitations & what we did not check
Watchdog assesses the repository exactly as committed, and only the repository. By design it does not reach outside the source tree: the live cloud account, the running CI/CD pipeline, the host's branch-protection and approval rules, the production configuration, or a restore actually exercised against a backup are all out of scope. That boundary is a feature, not a gap — a repo-relative, deterministic scan re-runs identically on any commit and every finding opens at a real file and line, where a live audit can neither be reproduced nor traced. The visible consequence is that controls which leave no in-repo evidence are reported as "not evidenced" and excluded from the score rather than awarded a number a static scan cannot justify.
Per-dimension blind spots
For each dimension that was measured, what a static, repo-only scan structurally cannot see — the honest edge of the measurement, not a failure of it.
D1 Cyclomatic Complexity: Cyclomatic complexity counts branches statically — it cannot tell an essential decision tree from accidental tangle, nor see complexity that lives in data or configuration (large switch-case token tables, DSL lexers/parsers, data-as-code rule tables) rather than control flow: a tokenizer's many single-character cases read as high complexity though each branch is trivial.
D2 Cognitive Complexity: Cognitive-complexity heuristics approximate how hard code is to follow; genuine domain difficulty and well-named intent that eases reading are not captured.
D3 God Classes: "God class" is sized by members and responsibilities visible in the type — a deliberately broad facade over a coherent subsystem can read the same as an accidental grab-bag. For front-end JS the file-length check is cohesion-aware (a single-responsibility module — one class/IIFE — earns a 3× threshold), but cohesion is approximated from top-level declarations, not true dependency structure.
D4 Code Duplication: Duplication is token-similarity — an in-process token-stream comparison over sliding windows, with type-aware normalization — so it finds copy-paste, not semantic duplication expressed differently. Committed machine-written code (scaffolded migrations, designer/codegen output, protobuf/OpenAPI stubs, model snapshots) is EXCLUDED — its repetition is the tool's, not the team's — so the score reflects hand-written duplication only.
D5 Coupling: Coupling is measured between projects/assemblies — runtime coupling through DI, reflection, messaging or shared databases is invisible to a static reference graph.
D6 Cohesion (LCOM4): LCOM4 cohesion is syntactic — it infers connectivity from which methods touch which fields/methods by name, not from real runtime behaviour or intent.
D9 Test Distribution: The test-pyramid shape is inferred from project/folder naming and references, with a single test host bucketed per-file by its path tier and content signals — a suite that names tiers unconventionally and gives no per-file signal can still be mis-bucketed.
D11 Test Reliability: Flakiness is inferred from history/markers — Watchdog runs the suite once (for coverage), not the repeated runs under varied conditions that reveal nondeterminism, so a flaky test never recorded as failing is invisible here.
D13 REDACTED Scanning: REDACTED detection is signature- and entropy-based on the current tree — a secret that does not match a known pattern, or one already rotated, will not be flagged (a clean scan is "nothing matched", not "no secrets exist").
D14 License Compliance: License compatibility is checked against declared package metadata and a policy — mislabelled or missing license metadata, and obligations that depend on how you distribute, are not resolved here.
D15 Churn × Complexity Hotspots: Churn hotspots come from git history — a freshly imported or squashed repository has no churn signal, and recent rewrites can mask a historically risky file.
D17 Explicit Debt: Acknowledged-debt signals (TODO/FIXME, suppressions, dead code) are textual — undocumented debt that nobody marked, and debt that lives in design rather than annotations, is invisible. Committed machine-written code (scaffolded migrations, designer/codegen output, generated stubs) is excluded — it is never the team's dead code to delete.
D19 Documentation Quality: Documentation quality publishes two things only: a runnable or packaged unit with no example, and a multi-project root README that maps none of its parts. It does not judge clarity, and its other checks (overview, licence, boilerplate, structure, empty sections, XML-doc coverage) are withheld after adjudication: their candidates are counted in the card's gaps, not shown as rows. A false statement in a document is M4's; units are .NET projects only.
D26 Project Cohesion: Project focus is sized from members/namespaces per project — a project that is broad by deliberate design reads the same as one that has sprawled.
D28 Secrets (history): Secret-history scanning sweeps the git log for known patterns — a secret that predates the available history, or never matched a signature, is not found (clean means "nothing matched in the history we can see").
D29 Static Analysis (SAST): SAST findings are pattern-based (semgrep) — it finds classes of bug it has rules for; logic flaws, auth/authorization gaps and issues needing runtime context are out of reach (and clean means "no rule matched").
D30 Dependency Vulnerabilities: CVE matching depends on accurate package/version metadata and on the advisory databases — a vulnerability with no published advisory, or in code not declared as a dependency, is not seen. Coverage needs a RESOLVED graph: an unpinned requirements.txt, or a pom without a resolved build, yields partial coverage rather than a clean verdict. An ecosystem the analyzer cannot scan is reported as unmeasured, never as clean.
D34 Knowledge Freshness: Freshness is decayed commit RECENCY, not comprehension — code read often but rarely committed reads as orphaned, and stable code that genuinely needs no changes is penalised the same as forgotten code; bot/squash commits distort it like the bus factor.
D35 Change Coupling: Change coupling is co-change in COMMITS — files split across separate commits, or coupled only through a shared config/build step, read as uncoupled, and a sweeping commit (rename/format) is excluded so it doesn't couple everything. It shows that files change together, not WHY: a high coupling can be a healthy cohesive pair as readily as a hidden leak.
D43 Malicious Dependencies: Only packages some vulnerability database has already NAMED as malicious are seen — a compromise published in the last hours, or never reported at all, is invisible here, and this dimension reading 10 is not evidence that a dependency is trustworthy. There is no typosquat or dependency-confusion analysis: a package nobody has reported is simply absent from the feeds. Coverage is the dependency scan's: an ecosystem that could not be scanned is disclosed as unmeasured, never as clean.
D44 Platform End-of-Life: The support table is FROZEN, so it goes out of date by losing RECALL: a release that ended support after the table was written is missed until the table is refreshed, and this dimension reading 10 is not evidence that a platform is current. Only platforms the repository DECLARES in a place this pass reads are seen — a runtime named only in a Dockerfile (D31's subject), in a CI workflow (D29's), or in a file this pass does not parse (go.mod, a Gemfile ruby directive) is invisible here, which is why a repository declaring none of them abstains rather than scoring. Only frameworks with a PUBLISHED support policy are tracked: React, Flask and Express publish none, so their age cannot be judged and their absence from a report is not a statement that they are supported.
AX10 Code composition: Role is inferred from namespace/folder convention, not semantics — a domain concept living in a folder named "Services" reads as application, and the split is lines-of-code, not business value. The business-logic-share score is a SOFT, FLOORED signal: it contributes to the Architecture lens but is floored at the Critical gate, so an infrastructure-heavy design (a gateway, an ETL, a driver) is legitimately low without being nuked to zero.
DM6 Domain ↔ infrastructure boundary: Infrastructure reached through a hand-rolled wrapper, a domain-named facade, reflection, or a string-keyed service locator resolves to a non-infra type and isn't seen; the body scan is symbol resolution over syntax, not full dataflow. A clean result means "no resolved infra reference in a domain body", not a proof of purity.
P4 Deployment & Rollback: Approval/branch-protection rules live in repository settings the scan cannot see — only their in-repo evidence (config files, workflows) is checked, so a control enforced purely in the host's settings reads as "not evidenced".
P6 Release Hygiene: Rollback/observability controls are inferred from repo artefacts (pipelines, dashboards-as-code) — controls configured in external tooling, with no in-repo trace, cannot be credited.
What it measures: How tangled the control flow is — methods with many branches are hard to test and change.
Method: Cyclomatic complexity per method (1 + decision points), computed exhaustively across production source; test projects separated by convention. Deterministic.
+ 46 more group(s) — more in Appendix A; the complete list is findings.md.
What to do
Bring the 3 bodies over 60 down to 60 or less in Cyclomatic Complexity — start with PENMANT5Tokenizer._fix_and_make_graph (cyclomatic 75), AMRBartTokenizer._fix_and_make_graph (cyclomatic 74), PENMANBartTokenizer._fix_and_make_graph (cyclomatic 74). — This score is capped by its worst body, so a finding fixed alone moves it by almost nothing — the next one down takes its place. Refactoring these 3 together lifts Cyclomatic Complexity from 6.3 to about 6.5/10, projected with the scoring formula itself and assuming each lands exactly at 60; a cleaner split scores higher.
Enforce Cyclomatic Complexity in CI to reach Verified (currently Documented). — Hardens enforcement from Documented toward Prevented — provenance only; does not change the score.
Detailed fixes: d1_recommendation.md · top locations in Appendix A, every location in findings.md.
What it measures: How hard the code is for a person to follow, beyond raw branching.
Method: Cognitive complexity per method (Sonar-style nesting-penalized score), computed exhaustively over production code, excluding test projects. Deterministic.
+ 144 more group(s) — more in Appendix A; the complete list is findings.md.
What to do
Bring the 5 bodies over 120 down to 120 or less in Cognitive Complexity — start with Document.to_pretty (cognitive 201), TransformerSequenceTokenizer.__call__ (cognitive 162), PENMANT5Tokenizer._fix_and_make_graph (cognitive 143). — This score is capped by its worst body, so a finding fixed alone moves it by almost nothing — the next one down takes its place. Refactoring these 5 together lifts Cognitive Complexity from 2.1 to about 3.2/10, projected with the scoring formula itself and assuming each lands exactly at 120; a cleaner split scores higher.
Enforce Cognitive Complexity in CI to reach Verified (currently Documented). — Hardens enforcement from Documented toward Prevented — provenance only; does not change the score.
Detailed fixes: d2_recommendation.md · top locations in Appendix A, every location in findings.md.
Do you agree with this assessment?
D3 · God Classes8.8 / 10Strong✓ Tool-verified
What it measures: Over-large classes that try to do too much ("god classes").
Method: God-class detection by line and method-count thresholds per logical type (partial classes unified), filtered for generated code and registration/contract false positives. Deterministic.
Resolve the 11 FileTooLong finding(s) in God Classes — start with ud210.py, ud27.py, conll_tf.py. — One of this dimension's main actionable groups (11 warning-level).
Resolve the 4 TooManyMethods finding(s) in God Classes — start with keras_component.py, multi_task_learning.py, biaffine_dep.py. — One of this dimension's main actionable groups (4 warning-level).
Enforce God Classes in CI to reach Verified (currently Documented). — Hardens enforcement from Documented toward Prevented — provenance only; does not change the score.
Detailed fixes: d3_recommendation.md · top locations in Appendix A, every location in findings.md.
What it measures: Copy-pasted code that should be shared instead.
Method: Code duplication via token-stream sliding windows with type-aware normalization (locals masked, type names preserved), density-scored per KLoC of production code. Deterministic.
126 duplicated block group(s) detected. A further 11 rows report members as variants of one another; they aggregate block groups already counted above and are not themselves counted.
+ 66 more group(s) — more in Appendix A; the complete list is findings.md.
What to do
Resolve the 12 Duplicated block (6 lines × 2) finding(s) in Code Duplication — start with modeling_bart.py, tokenization_bart.py, bmes_tf.py. — One of this dimension's main actionable groups (12 warning-level).
Resolve the 10 Duplicated block (9 lines × 2) finding(s) in Code Duplication — start with conll_tf.py (2), tokenization_bart.py, alg.py. — One of this dimension's main actionable groups (10 warning-level).
Resolve the 10 Duplicated block (7 lines × 2) finding(s) in Code Duplication — start with dep.py (2), tokenization_bart.py, alg.py. — One of this dimension's main actionable groups (10 warning-level).
Enforce Code Duplication in CI to reach Verified (currently Documented). — Hardens enforcement from Documented toward Verified — provenance only; does not change the score.
Detailed fixes: d4_recommendation.md · top locations in Appendix A, every location in findings.md.
Do you agree with this assessment?
D5 · Coupling10.0 / 10Exemplary✓ Tool-verified
What it measures: Whether volatile projects sit underneath others that depend on them (so their churn ripples upward), and whether project dependencies form cycles. A widely-depended-on but stable shared/kernel project is healthy, not penalised.
Method: Dependency cycles via elementary-DFS over real .csproj references, plus Martin instability (afferent/efferent) per project. Exhaustive over the reference graph, deterministic.
Coverage: Exhaustive · type-level: afferent/efferent coupling + cycles computed over every production type — the population is all types, not a name convention.
5 production modules (Maven+Python), 0 dependency cycle(s), 0 unstable depended-on module(s). Read from the build's own module declarations; 0 module(s) off the main sequence.
What it measures: Whether a class's methods are focused on a single responsibility.
Method: LCOM4 cohesion per production class with at least two methods: connected components of methods sharing state or calls, computed syntactically. Deterministic, not a proxy.
Coverage: Exhaustive · type-level: LCOM4 cohesion computed over every production class — the population is all types, not a name convention.
Resolve the 3 Low cohesion finding(s) in Cohesion (LCOM4) — start with transform_tf.py, word2vec.py, conll_tf.py. — One of this dimension's main actionable groups (3 warning-level).
Enforce Cohesion (LCOM4) in CI to reach Verified (currently Documented). — Hardens enforcement from Documented toward Verified — provenance only; does not change the score.
Detailed fixes: d6_recommendation.md · top locations in Appendix A, every location in findings.md.
Do you agree with this assessment?
D9 · Test Distribution10.0 / 10Exemplary✓ Tool-verified
What it measures: Whether the test suite has a healthy mix of unit / integration / end-to-end tests.
Method: Test projects classified (Unit/Integration/BDD/E2E) from compiled metadata; test methods counted exhaustively across projects with placement-agnostic disk fallback. Deterministic.
52 test methods: 52 unit, 0 integration, 0 BDD, 0 e2e. The Python suite contributes 33 test function(s) across 8 file(s) declaring at least one — every `def test…` in a file pytest or unittest would collect, which is those frameworks' own definition of a case; a parametrize table counts once, so this is a floor. Its tier split is read from file names and paths only.
✓ On the Gold path — maintain.
Detailed fixes: d9_recommendation.md.
Do you agree with this assessment?
D11 · Test Reliability10.0 / 10Exemplary✓ Tool-verified
What it measures: Whether the tests pass reliably, with no flakiness.
Method: Suite re-run N times within tiered wall-clock budgets (unit to e2e); tests failing non-deterministically across runs flagged; guarded tests retried when #if guards detected.
What it measures: Whether any secrets (keys, tokens, passwords) have leaked into the code.
Method: In-process native secret scanner (entropy plus signature patterns) across all tracked files; no external tool. A clean result is a measured 10, not no-data zero. Deterministic.
What it measures: Whether the licenses of third-party packages are compatible with your policy.
Method: Third-party package licenses resolved from declared package metadata and checked against the configured policy (allow/deny/copyleft). Deterministic; clean = no incompatible license found at metadata depth.
0 of 1 declared JVM dependency artifact(s) use a banned license. Licences were resolved from the POMs Maven Central publishes, over the coordinates this repository's 1 Maven/Gradle manifest(s) declare with a version resolvable from the manifest itself — the same declaration-site closure this analysis reads for dependency currency. The JVM has no lockfile a checkout is guaranteed to carry, so this is deliberately NOT the transitive closure. A further 1 coordinate(s) are not published to Maven Central — another repository (Google's Maven hosts androidx), a SNAPSHOT, or a private artifact — so their licences are not readable from here and they are outside this verdict.
What it measures: Files that change often and are also complex — the riskiest hotspots.
Method: Per production file churn times cyclomatic complexity over a rolling window, computed from git and Roslyn/JS/Razor analysis. Exhaustive, deterministic per commit date.
What it measures: Acknowledged debt left in the code — TODOs, dead code, suppressed warnings.
Method: Roslyn syntactic debt markers (suppressions/TODO/FIXME/HACK/empty-catch/commented-code/Obsolete) plus SymbolFinder dead-code analysis; weighted-debt-per-KLoC density deducted 2.0x per unit. Deterministic, exhaustive.
9 deducted task-comment markers across 50355 LoC (0.0/KLoC) → score 10.0. Task comments only: this repository's language is read without a compiler, so D17's suppression, dead-code and commented-out-code arms did not run and this score counts fewer marker kinds than a .NET repository's would.
Resolve the 8 TodoComment finding(s) in Explicit Debt — start with REDACTED (3), bert_tok.py, transformer_tokenizer.py. — One of this dimension's main actionable groups (8 warning-level).
Resolve the 1 FixmeComment finding(s) in Explicit Debt — start with srl_eval_utils.py. — One of this dimension's main actionable groups (1 warning-level).
Enforce Explicit Debt in CI to reach Verified (currently Documented). — Hardens enforcement from Documented toward Prevented — provenance only; does not change the score.
Detailed fixes: d17_recommendation.md · top locations in Appendix A, every location in findings.md.
What it measures: Whether the project's documentation is clear, complete, and useful.
Method: Every README, docs/ document and root guide parsed by its markup and measured in Unicode letters. Published: a runnable or packaged .NET project whose README (or a document it links, or a root guide) shows no example, and a root README naming none of three or more projects. Overview, licence, boilerplate, wall-of-text, empty-section and XML-doc rows are measured and withheld (counted in gaps) after adjudication. Deterministic, exhaustive, advisory.
1 of 5 build units (Maven, Python) flagged as possibly oversized/incoherent.
Projects may be oversized for their cohesion
What to do
Resolve the 1 Projects may be oversized for their cohesion finding(s) in Project Cohesion. — One of this dimension's main actionable groups (1 recommendation-level).
Detailed fixes: d26_recommendation.md · top locations in Appendix A, every location in findings.md.
What it measures: Whether any secrets were ever committed — scanned across the full git history, not just now.
Method: REDACTED scan via TWO gitleaks detect passes in an isolated checkout — the full git history, then a second --no-git pass over the working tree as it stands — merged and de-duplicated by (rule, file, line); each match flagged High. Both invocations are recorded in the audit trail. Exhaustive; when the tool is absent, or when its output cannot be parsed into the expected shape, the dimension is WITHHELD as an explicit measurement gap on our side — unscored and excluded from the lens, never a hedged middling score.
1 finding(s): 0 critical, 1 high, 0 medium, 0 low. Remediation for historically-committed secrets is credential rotation — they remain in history regardless of later deletion.
REDACTED
REDACTED
What to do
Resolve the 1 REDACTED finding(s) in Secrets (history) — start with REDACTED. — One of this dimension's main actionable groups (1 issue-level).
Resolve the 1 Rotate the exposed credentials finding(s) in Secrets (history). — One of this dimension's main actionable groups (1 recommendation-level).
Detailed fixes: d28_recommendation.md · top locations in Appendix A, every location in findings.md.
What it measures: Real static-analysis (SAST) findings — likely security bugs in the code, any language.
Method: Polyglot static analysis via semgrep across the repo using the pinned, image-baked p/security-audit + p/owasp-top-ten rulesets (no scan-time registry fetch); severity rules (ERROR/WARNING/INFO) map to a full-band severity-weighted score. Exhaustive, deterministic; degrades on parse failure.
Coverage: semgrep pattern rules over all files — exhaustive for the rule set, blind to classes of bug without a rule (clean = no rule matched).
18 finding(s): 0 critical, 12 high, 3 medium, 3 low. 4 unpinned-GitHub-Actions row(s) are reported here but scored by D36 (supply-chain provenance), which measures that posture as `pinned_actions` — one pinning decision is charged once, not once per lens.
REDACTED
REDACTED
REDACTED
REDACTED
REDACTED
+ 2 more group(s) — more in Appendix A; the complete list is findings.md.
What to do
Resolve the 5 REDACTED finding(s) in Static Analysis (SAST) — start with REDACTED (2), REDACTED, REDACTED. — One of this dimension's main actionable groups (5 issue-level).
Resolve the 2 REDACTED finding(s) in Static Analysis (SAST) — start with REDACTED (2). — One of this dimension's main actionable groups (2 issue-level).
No action in Static Analysis (SAST) — all 4 REDACTED finding(s) are reported here at file:line but scored by D36 (supply-chain provenance), so none is charged to this dimension. — One of this dimension's main actionable groups (4 issue-level, 0 of them charged here).
Detailed fixes: d29_recommendation.md · top locations in Appendix A, every location in findings.md.
What it measures: Whether any dependency has a known published vulnerability (CVE), direct or transitive, in ANY ecosystem the repository declares — Dart pub, Elixir and Erlang via Hex, Go modules, Java and Kotlin via Maven/Gradle, JavaScript/npm, .NET/NuGet, PHP/Composer, Python/PyPI, RubyGems, Rust/Cargo and Swift.
Method: Dependency-CVE scan across every ecosystem the repository declares, scored ONCE. Three sources are unioned and deduplicated by advisory identity (rule id + alias closure, CVE<->GHSA) scoped to package+version, keeping the worst severity: `osv-scanner --recursive` over osv.dev for Dart pub, Elixir/Hex (and Erlang, whose `rebar.lock` syft first converts to a CycloneDX SBOM the scanner reads, with rows attributed back to the lock), Go, Java and Kotlin via Maven/Gradle (and Scala, whose sbt build's pinned direct declarations are written into a CycloneDX SBOM the scanner reads, with rows attributed back to the build file), npm, PHP/Composer, Python/PyPI, RubyGems, Rust/Cargo and Swift; `trivy fs --scanners vuln` for npm lockfiles; and `dotnet list package --vulnerable --include-transitive` for NuGet (with per-advisory collapse of the project x target-framework fan-out), plus a DECLARED-dependency arm that resolves a published gem's gemspec against rubygems.org where no Gemfile.lock is committed. `SeverityScore(c,h,m,l, normalizer 8.0)`. NotApplicable only when NO ecosystem is readable; if any applicable ecosystem could not be scanned the findings are REPORTED and the score is withheld. Supersedes the npm and OSV arms, retired 2026-09-05.
Resolve the 1 High CVE finding(s) in Dependency Vulnerabilities — start with REDACTED. — One of this dimension's main actionable groups (1 issue-level).
Detailed fixes: d30_recommendation.md · top locations in Appendix A, every location in findings.md.
What it measures: Whether anyone still has living knowledge of each file, or it has been orphaned — last understood long ago by someone now gone quiet. The sibling of the bus factor: D16 asks who owns it, D34 asks whether anyone still knows it.
Method: File orphaning as total living-knowledge decay below one focused-commit's worth within a year, computed per-file from the D16 decay model. Exhaustive, deterministic over fixed history.
Every significant source file has living knowledge — recently and meaningfully worked. Counted over 206 of the 475 production source files in this repository: the rest are under the ~2,400-byte size floor this dimension measures over.
What it measures: Whether files that change together actually belong together — pairs that repeatedly co-change in git history despite having no explicit code dependency, surfacing the hidden/logical coupling (and boundaries in the wrong place) a static scan can't see.
Method: Pairwise co-occurrence over the per-commit file sets in git history (production source only — tests and generated dropped): Degree-of-Coupling = shared ÷ min individual revisions, reported above noise floors (each file ≥10 revisions, ≥5 shared commits, ≥50% strength); sweeping commits excluded. Deterministic over fixed history.
Coverage: Population: PRODUCTION source files only — test and generated files are dropped before pairing, so a class co-changing with its own test (trivially ~100%) can't drown the real production↔production coupling. Pairs ranked by Degree-of-Coupling. A non-source file is never a coupling PARTICIPANT either: documentation, schemas, config and data files are dropped with the rest, so a code↔docs pair — a command and the reference page that restates it — is not reported however strongly the two co-change; nor is coupling that runs THROUGH a build step or config file.
Resolve the 2 Change coupling finding(s) in Change Coupling — start with REDACTED, document.py. — One of this dimension's main actionable groups (2 warning-level).
Resolve the 1 Change-coupling hub finding(s) in Change Coupling — start with version.py. — One of this dimension's main actionable groups (1 warning-level).
Detailed fixes: d35_recommendation.md · top locations in Appendix A, every location in findings.md.
What it measures: Whether the build pipeline provides supply-chain integrity — generated provenance/attestation, signed artifacts (cosign/sigstore), an SBOM, and pinned build actions. Presence of the configuration, not a runtime guarantee.
Method: Supply-chain provenance/signing read deterministically from CI/build config (.github/workflows, .gitlab-ci.yml, azure-pipelines, Jenkinsfile, .circleci) + the release surface: four signals — generated provenance/attestation (SLSA/in-toto/actions-attest), artifact signing (cosign/sigstore/gitsign), an SBOM (syft/sbom-action/*.spdx.json/*.cdx.json), and SHA-pinned build actions — scored 10·present/denom. NotApplicable without a build pipeline. Detects configuration presence, not runtime enforcement.
What it measures: Whether any dependency the repository declares is published as MALICIOUS rather than merely vulnerable — a package that is an attacker's work, in any ecosystem osv-scanner reads. Scored apart from D30 because the answer is binary: there is no safe version to upgrade to, and the fix is to remove the package and rotate every credential it could have read.
Method: The same dependency scan D30 reads, partitioned on the scanner's own classification rather than rescanned: a row is MALICIOUS when its id is in the `MAL-` space (the ossf/malicious-packages feed) OR its `database_specific.cwe_ids` carries `CWE-506` ("Embedded Malicious Code"). Both channels are structural; the summary text is deliberately NOT read, because a malicious-package record whose summary says only "Critical severity vulnerability" is a real shape ([GHSA redacted]) and a text matcher misses it. Scored BINARY: any surviving row is 0, whatever its severity and however many CVEs sit beside it — a hostile dependency is not a quantity. Applicability and degradation are D30's: NotApplicable only when no ecosystem is readable, and an unscannable ecosystem degrades rather than reading clean. SCORED, not informational.
What it measures: Whether anyone still ships security patches for the platform this repository RUNS ON — the runtime it pins and the framework majors its own constraints hold it to. Separate from D12 because the question differs: a current Django on an end-of-life Python is perfectly up to date and completely unsupported, and the fix is a migration rather than a version bump. What the repository says it merely SUPPORTS is never charged.
Method: End-of-life PLATFORM read from the repository's own declarations and graded against a FROZEN, dated table of vendor support dates — no network, no feed, no API, so this dimension answers identically inside a closed scan fence. Two subjects: a RUNTIME the project pins (a single or all-end-of-life TargetFramework, a .nvmrc or .python-version, a requires-python CAP) and a FRAMEWORK major a dependency constraint cannot move off (a caret, tilde or exact version; `vue@^2.7.16` pins Vue 2). A FLOOR is deliberately never charged — `requires-python = ">=3.8"` states what a package SUPPORTS, not what it runs on — and a multi-target project is charged only when EVERY target is out of support. Runtime 4.0/product capped 8.0, framework 1.5 capped 4.5. The table is safe to freeze because a statement about support that ended in the past cannot become false: it loses recall as it ages, never precision, and a test asserts every entry predates the freeze date. Disjoint from D31 (a container image's OS layer) and D29 (the toolchain a CI workflow installs). Abstains when the repository declares no platform this pass reads — never scores it clean.
0 end-of-life runtime(s) and 0 end-of-life framework(s), read from 0 platform declaration(s) and 4 dependency declaration(s). This dimension reads what the repository says about ITSELF — a pinned target framework, a version file, a capped requires-python, a Rust toolchain pin, a framework major a constraint cannot move off. A FLOOR is deliberately never charged: `requires-python = ">=3.8"` states what the package SUPPORTS, not what it runs on, and a well-maintained library declares exactly that while running its own CI on a current release. The end-of-life facts are FROZEN and dated, so this dimension needs no network and answers identically inside a closed scan fence; as the table ages it loses recall and never precision, because a statement about support that ended in the past cannot become false. The OS layer of a container image is D31's question and the toolchain a CI workflow installs is D29's; this row is neither.
Other · Architecture — How the codebase splits by code ROLE — domain, application, infrastructure, test, generated. The significance map behind the knowledge/coupling weighting, and a DDD signal in its own right: a thin domain core under fat infrastructure is the anemic-domain smell, quantified. How each file's role is decided, because the split is only as good as that: a generated name or a build-output tree makes it Generated, a test project makes it Test, and otherwise the file's NAMESPACE and PATH words are matched against fixed vocabularies in a fixed ORDER — domain, then infrastructure, then application — so a file whose words hit two layers is counted under the earlier one. A production file matching none of them counts as application, so that share reads 'application or unclassified' rather than a measured application layer. Roles come from naming convention, never from what the code does. On this repository the split was taken from the source tree on disk rather than from a loaded .NET workspace, so a file's role is decided by its PATH segments alone — no declared namespace was available to add to the evidence — and generated output is excluded from the census entirely rather than counted as a generated share.
Method: Roslyn line-count by code ROLE: every source file classified Domain/Application/Infrastructure/Test/Generated by namespace + path convention (the shared CodeRoleClassifier), then significant lines summed per role. Deterministic; the advisory score is the business-logic (domain+application) share of production code.
Coverage: Population: ALL source files, each bucketed into ONE of five roles (Domain/Application/Infrastructure/Test/Generated) by namespace + path convention — a file whose layer isn't named in the convention falls to Application (the neutral default), and the split is line-count, not semantic depth or business value.
What to do
The domain core is a small share of production code, but most of the rest matched no layer vocabulary at all — so this is not yet an anemic-domain finding. The namespace/path convention could not place that code, which makes the composition above a statement about the naming, not about the design. Name the layers (or check that the repository's conventions differ from the ones this check knows) before reading a thin domain into it.
Other · Architecture — Whether the project-reference graph is acyclic (cycles block independent build/deploy and signal eroding boundaries).
Method: Project reference cycles via elementary-DFS over real .csproj references, using the engine shared with D5/D7; cyclic versus acyclic. Exhaustive, deterministic.
Other · Architecture — Whether dependencies point inward (Domain ← Application ← Infrastructure/Web) — the clean-architecture dependency rule, checked across the project graph.
Method: Layer violations by name-segment inference (Domain/Core to Application to Infrastructure/Web) over the project-reference graph. Exhaustive over all projects, deterministic.
Other · Domain Modelling — Whether the domain layer stays free of infrastructure dependencies — a domain aggregate fused to a persistence ORM (Django/SQLAlchemy) on its own declaration (active-record) couples the domain to infrastructure. The clean-architecture dependency rule.
Method: Roslyn (DDD-gated): domain-layer types scanned for infrastructure usage in member SIGNATURES and inside method/accessor BODIES — resolved calls and object-creations into EF/Marten/HTTP/Mongo/Redis/message-bus types (not just a namespace allowlist). Deterministic, symbol-resolved, exhaustive over domain-layer bodies, DDD-native.
Coverage: Domain layer identified by NAMESPACE heuristic; infrastructure then resolved by symbol in member SIGNATURES and method/accessor BODIES — rename the layer and the check evaporates.
Maturity · Maturity — Whether the repo and its projects have a README, and whether it's substantive and current.
Method: Filesystem scan: README presence, word count, and headings for depth; git history for staleness. Exhaustive across root and project dirs, deterministic.
What to do
Add a build/run (quick start) section to the root README — the first thing a newcomer needs.
Explain in the README how to run the tests. Add a section called Testing to the README in the repository's top folder, saying how to run the test suite.
Add an 'Architecture' / 'How it works' section to the root README — the high-level shape.
Add a README to the 1 of 1 project(s) that lack one — worth up to 2 pts.
Maturity · Maturity — Whether key decisions (ADRs) and the high-level shape (C4/diagrams) are written down.
Method: Filesystem scan: ADR folder/naming conventions or content, plus Mermaid/PlantUML/C4/architecture.md discovery. Exhaustive, deterministic.
No Architecture Decision Records found — no conventional ADR directory, no numbered `NNNN-title` documents in any markup this check reads, and nothing ADR-shaped by content. Design rationale recorded elsewhere (a design-notes tree, a mailing list, pull-request discussion) is not visible to this check and is not re-findable per decision, so a future maintainer cannot ask why one choice was made and get an answer.
What to do
Record your significant design decisions. Write one short document per decision, giving its date, the situation, the decision and its consequences, and keep them together with your other design documents. A docs/adr folder with each file named by number and title is the easiest place to find them.
Maturity · Maturity — Whether the repo is organised deliberately — src/test separation and consistent project naming.
Method: Filesystem scan: src/test folder separation and namespace-prefix consistency (majority RootNamespace agreement). Exhaustive across projects, deterministic.
Production code isn't grouped under a src/ folder — it's spread across several top-level directories, so there's no one place that says 'this is the product'.
What to do
Group production code under src/ (or split deliberately, e.g. backend/ + frontend/) so production and tooling code aren't mixed at the root.
Readiness · Readiness — Whether an automated pipeline builds and tests every change.
Method: Filesystem scan: CI workflow files (.github/workflows, .gitlab-ci.yml, etc.) for build and test stages. Exhaustive, deterministic.
Do you agree with this assessment?
P10 · Library API & versioning6.0 / 10Adequate✓ Tool-verified
Readiness · Readiness — For a library: a deliberate (small) public API surface and explicit semantic versioning so consumers can depend on it safely.
Method: Roslyn scan: public API surface area and semantic-versioning markers (SemVer attributes, changelog entries) for libraries; off .NET, a library is the ecosystem's publication act (an npm package that is not private and names an entry point, a PyPI distribution with a build system, a Rust library crate, a Maven/Gradle module that publishes, a Go module with no package main, a gemspec, a Composer library, a SwiftPM library product, a pub.dev or Hex package), its surface is the share of types the language model records as public (Rust, Swift, Java, Kotlin, Go, Dart; not measured where the model records no type visibility or, as in TypeScript, only module-level export), and its version is read from the manifest, a semver CHANGELOG, release tooling or semver git tags. Exhaustive, deterministic.
11/11 types (100%) declared in the published library are public. For a library, every public type is a stability contract — keep implementation types off the surface and expose only the intended API.
What to do
Keep implementation types off the public surface (Java: package-private (no modifier), or leave its package out of `module-info` exports) so internals can change without breaking consumers.
Readiness · Readiness — Whether SAST, secret/dependency scanning and performance benchmarking are wired in (presence, not runtime).
Method: Filesystem scan: SAST configuration, dependency-update automation, secret scanning, and a benchmark harness or benchmark step — in this repository's own ecosystem. Exhaustive, deterministic.
No static application security testing detected. For this repository's stack, add bandit, `semgrep --config=p/python`, or CodeQL's python pack as a CI step. What was searched, so you can tell an absence from a miss: the 2244 CI workflow file(s) in this repository, and the scanner and linter configuration checked in beside them. A scan that runs outside CI, one configured in your forge's web UI rather than in a committed file, or a tool whose name is none of those this check carries, is not seen — if that is your case the row is wrong, and saying so is more useful than adding a second scanner.
What to do
Add a SAST step to CI running what this repository's stack ships: bandit, `semgrep --config=p/python`, or CodeQL's python pack — so a security regression fails the build instead of landing.
Enable Dependabot/Renovate or a dependency-review gate.
Add gitleaks/trufflehog in CI to block PRs that introduce committed secrets.
Readiness · Readiness — Whether releases are automated and safely reversible (probes, rolling updates, approval gates) — from manifests/pipeline files, not the live environment.
Method: Filesystem scan: deployment manifests/IaC (K8s YAML, Helm, Terraform) for rolling updates, probes, approval gates, migration hooks. Exhaustive, deterministic.
The release is automated and no gate that pauses it for a human is DECLARED IN THIS REPOSITORY'S PIPELINE FILES. What was read: every file under `.github/workflows/`, `.forgejo/workflows/`, `.gitea/workflows/`, `.azuredevops/` and `.azure-pipelines/`, plus `.gitlab-ci*` and `azure-pipelines*` — with comment text stripped, so documenting a gate is not declaring one. What would have counted: GitLab's `when: manual`, CircleCI's `type: approval`, an Azure `ManualValidation@` task or an `approvals:` block, a Jenkins `input` step, a `uses:` step naming an approval action, an `environment:` paired with `reviewers` / `required_reviewers` / `protection` / `wait-timer` / `deployment_branch_policy`, a draft-release step, a `workflow_dispatch` promotion, or a release-event gate. ★ What this cannot see, because none of it is a file: a GitHub environment whose required reviewers are configured in repo SETTINGS, a branch protection rule, or an organisation deployment policy — all of them real, enforced gates that live outside the repository. If yours is one of those, this row is wrong and nothing in the tree could have told us. Otherwise: whatever the release trigger points at is published to users unreviewed, so a mistagged or unverified commit ships and the only remedy is a follow-up release.
What to do
Nothing pauses a release for a human: publish as a draft release (or gate the release job on a protected tag/manual dispatch) so a bad build can be stopped before users can download it.
Readiness · Readiness — Whether releases are traceable — a maintained changelog and explicit version stamping.
Method: Filesystem scan: changelog file presence and version tags in csproj or git tags. Exhaustive, deterministic.
No CHANGELOG/HISTORY/RELEASES file — what shipped when isn't easy to reconstruct for support or audit. (Versioning/tagging makes releases traceable, but a changelog records the what.)
What to do
Keep a changelog (e.g. Keep-a-Changelog) recording what shipped in each release.
Other · Code Health — Whether the code avoids sync-over-async (deadlock-prone blocking on tasks) and async void.
Method: Roslyn syntax scan: async methods scanned for .Wait()/.GetAwaiter().GetResult() and async-void outside event handlers. Deterministic, hard fact per invocation.
Other · Code Health — Whether any branch is dead by construction — a switch arm whose label can never equal a case-normalised subject, or an `else if` whose predicate the arm above has already swallowed.
Method: Roslyn syntax + semantics: switch labels compared against the subject's own case normaliser, and if/else-if chains checked for a literal an earlier arm's containment test already swallows. Deterministic, provable per finding. Advisory.
Do you agree with this assessment?
X13 · Undrained process stream10.0 / 10Exemplary○ Nothing flagged
Other · Code Health — Whether a child process that has BOTH standard streams redirected drains both — reading one to the end while the other is never read deadlocks once the child fills the unread pipe.
Method: Roslyn syntax + semantics: ProcessStartInfo launches with both streams redirected, checked for a drain of each stream across the enclosing type. On a repository with no .NET source it reads JavaScript/TypeScript off the engine’s own token stream (test, vendored and minified paths not) with the same rule: a `spawn` imported from `child_process` whose stdout and stderr are both pipes (no options, no `stdio`, or `stdio` of `'pipe'`), bound to a local that never leaves its scope, where exactly one of the two streams is read, the other never, and the child’s `close`/`exit` (or the read stream’s end) is awaited; a shell redirect in the call’s arguments or a `kill` of the child suppresses it. The same rule reads Python’s `subprocess.Popen`, Dart’s `Process.start`, Ruby’s `Open3.popen3`, and Scala’s JDK process starts (`new ProcessBuilder(…).start()` with no `redirectErrorStream`/`redirectError`/`redirectOutput`/`inheritIO`, `Runtime.getRuntime.exec`) where one of `getInputStream`/`getErrorStream` is read and the other never while `waitFor()` is called or the read stream is read to its end. Deterministic, provable per finding. Advisory.
Other · Security — Whether a hand-rolled public/private IP check can be walked past — a method that unwraps IPv4-mapped IPv6 but returns the opposite verdict for the same host written as IPv4-compatible, 6to4 or NAT64.
Method: Roslyn syntax + semantics: methods that unwrap IPv4-mapped IPv6 and hand-roll IPv4 range carve-outs, checked for whether the IPv6 branch also accounts for the IPv4-compatible, 6to4 and NAT64 embeddings. Deterministic, provable per finding. Advisory.
Do you agree with this assessment?
X15 · Unvalidated length from an untrusted reader10.0 / 10Exemplary○ Nothing flagged
Other · Security — Whether a length read out of the stream being parsed is bounded before it is allocated or read — an unchecked count taken from the input lets the input choose the allocation.
Method: Roslyn syntax + semantics: integer lengths read from a BinaryReader and spent on a bulk read or an array allocation, checked for any comparison or bounding call on the value anywhere in the method. Deterministic, provable per finding. Advisory.
Other · Code Health — Whether a loop that shortens a string until it fits a length budget has a floor — one with none grinds the value down to the empty string, or past it into a negative-length `Substring`.
Method: Roslyn syntax + semantics: while/do loops whose body's only effect on a string is to drop its last character, checked for whether anything — a direct comparison on the length, a body guard, a break — bounds that length below. On a repository with no .NET source it reads JavaScript/TypeScript off the engine’s own token stream (test, vendored and minified paths not) with the same rule: a `while`/`do` loop whose body’s one assignment to a value is `x = x.slice(0, -1)` or `x = x.slice|substring|substr(0, x.length - 1)`, driven by a condition that reads `x.length` only as a term of a larger expression — never compared directly, never tested for truthiness, and with no other read of `x` — and whose body has no `break`, `return`, `throw` or `if` naming `x`. Deterministic, provable per finding. Advisory.
Do you agree with this assessment?
X17 · Uncapped recursion over a caller-supplied document3.5 / 10Weak✓ Tool-verified
Other · Security — Whether a walk that recurses through a JSON/XML tree handed in by its caller bounds how deep it will go — an uncapped walk lets the document's nesting choose the stack depth, and the resulting StackOverflowException cannot be caught.
Method: Roslyn syntax + semantics: methods that take a JSON/XML document node and call themselves with a child of it, reachable from an externally-callable member of the same type that accepts a document, checked for any depth parameter, descent counter or threaded arithmetic anywhere in the walk. On a repository with no .NET source it reads TypeScript off the engine’s own token stream (test, vendored, generated and minified paths and `.d.ts` not) with the same rule: a function or method with a parameter typed as a document value — `unknown`, `any`, `object`, `Record<string, unknown|any>`, a JSON alias (`JsonValue`, `JSONObject`, …) or a DOM node global the file does not rebind — that calls itself (bare, or through `this` for a method, or hands itself to a call over the value as in `value.map(walk)`) with something it took out of that value and never through an ancestor accessor such as `closest()` or `parentElement`, that is exported or reached from an exported function (a public method of the same exported class) taking such a value, and that names no depth, level, nesting, recursion, remaining or budget anywhere and threads no `+`/`-` arithmetic through a self-call. Plain JavaScript is not read: with no annotation nothing tells a parsed document from a tree the code built itself. Deterministic, provable per finding. Advisory.
`from_config` is public and takes `config: dict` from its caller, and it calls itself at line 27 (`self.from_config`) with a piece of `config` (and at line 32) — and nowhere in `from_config` is there a depth parameter, a descent counter, or any other cap. The walk therefore spends one stack frame per LEVEL of the value it is handed, so the nesting of the caller's document — not this code — decides how deep the stack goes. Python's parsers accept far more nesting than its stack allows: `json.loads` turns 3,000 bytes of brackets into 1,500 levels and `xml.etree.ElementTree` parses tens of thousands, while a walk like this one raises `RecursionError` once it passes the interpreter's default limit of 1,000 frames. Wherever that exception lands, the work fails: a request errors, and outside any `try` the worker or the program exits with it. Raising `sys.setrecursionlimit` only moves the number the document has to beat. Thread a depth argument through `from_config` and reject the value past a fixed maximum, or rewrite the descent with an explicit stack. — plugins/hanlp_common/hanlp_common/configurable.py:27
What to do
Each finding names the externally-callable member that accepts the document and the self-call that descends, and states that no depth parameter or counter exists anywhere in the walk. Confirm from the two, then thread a depth argument through the walk and reject documents past a fixed maximum, or rewrite it with an explicit stack.
Other · Code Health — Whether a type's disposal matches what it OWNS — releasing what it created, leaving alone what it was handed, and not declaring a finalizer for state that has nothing unmanaged to finalize.
Method: Roslyn syntax + semantics: every assignment to a disposable field is read to decide whether the type CREATED the value or was handed it, and the type's disposal is checked against that answer — an injected interface it disposes, a value it constructed and never releases, a finalizer on a type holding nothing unmanaged, and a disposable local whose every reference is a plain member read. A value handed to a container that disposes its contents (a parent control's `Controls` collection, a component `IContainer`) is released by that container and is not reported; generated code is out of population. On a repository with no .NET source the same ownership questions are read in JavaScript/TypeScript off the engine’s own token stream (test, vendored, generated and minified paths not): a class declaring `dispose()`, `[Symbol.dispose]()` or `[Symbol.asyncDispose]()` that disposes a field it was handed through a constructor parameter typed as a repository interface or resolved by a dependency-injection container; that assigns a field only ever from `new X(…)` of a disposable class and neither releases it anywhere in the class nor names it in its disposal member or a method that member calls, nor hands it to anything else; and a `const`/`let` local built from literals only whose every reference opens a statement operating on a non-release member of it. A class is disposable when every repository declaration of its name declares or inherits a disposal member, or when it is a documented library disposable (`vscode` EventEmitter, CancellationTokenSource and Disposable; three.js geometries, materials, textures, render targets, renderers, controls and composers). The finalizer arm has no JavaScript counterpart: a class cannot declare one. Dart is read for the owned-and-never-released arm: a Flutter `State` (or a class declaring `dispose()`/`close()`) whose library-private field is written only with direct creations of a framework controller, `FocusNode`, `StreamController` or periodic `Timer` (or a `.listen(…)` into a field declared `StreamSubscription`), which nothing in the class or its library releases (`dispose`/`close`/`cancel`) and which is never handed on except as a named argument. Ruby is read for the owned-and-never-released arm and the abandoned-local arm: a class declaring an instance `close`, `dispose`, `shutdown`, `stop`, `disconnect` or `terminate` whose instance variable is written only with a blockless core descriptor creation (`File.open`/`new`, `Tempfile`, `Dir.open`, `IO.popen`, the `socket` and `openssl` socket classes) that no method of the class closes and that is never handed on (passed, returned, stored, exposed by `attr_reader`, delegated) — a class whose superclass or mixin lies outside the repository is not judged — and a blockless creation read through and dropped (`File.open(path).read`) or held in a local that is never closed, returned, passed or stored (a `&:read` block-pass, an `autoclose: false` borrow and a `fileno` hand-over are not abandonment). PHP is read for the same two arms: a class (not a trait, and not one using a trait) declaring an instance `close`, `dispose`, `shutdown`, `stop`, `disconnect` or `terminate` whose `private` property (or a `final` class's `protected` one) is written only with a direct core `fopen`, `tmpfile`, `popen`, `proc_open`, `fsockopen`, `stream_socket_client`/`server` or `opendir` that nothing in the class closes (`fclose`, `pclose`, `closedir`, `unset`, a reset to `null`) and that is never handed on, and a handle held in a function's local, or read through as the argument of a reader, that is never closed, returned, passed, captured or stored; a `php://` stream, `curl_init`, `SplFileObject`, `PDO` and a function the repository declares under the core name are not read. Deterministic, provable per finding. Advisory.
Other · Code Health — Whether a method that temporarily changes state belonging to the whole process — the working directory, an environment variable — puts it back on EVERY path: a restore reached only when nothing throws leaks the change to the rest of the process.
Method: Roslyn syntax + semantics: method bodies that write the process working directory or an environment variable and write it back in the same body, checked for whether that restore sits in a `finally`/`catch` or only on the straight-line path. On a repository with no .NET source the same rule reads production JavaScript/TypeScript off the engine’s own token stream (test, vendored and minified paths not): `process.chdir`, `process.env.NAME =`/`["NAME"] =` and `delete process.env.NAME`, and Deno’s `Deno.chdir`/`Deno.env.set`/`Deno.env.delete`, paired per function body (a nested function or arrow is its own body, and module top-level code is none), where the last write puts back a local the body captured from the same global or deletes a variable the first write set, with at least one statement between them; a write in a `catch`/`finally` of that body silences it. Deterministic, provable per finding. Advisory.
Other · Code Health — Whether an argument guard throws the exception its own condition describes — a guard that rejects a value for being EMPTY and reports it as `NullPointerException` (Java, Kotlin, Scala), `ArgumentError.notNull` (Dart) or `ArgumentNullException` (F#'s `nullArg`) tells the caller a parameter was null when it provably was not.
Method: Roslyn syntax: `throw new ArgumentNullException(nameof(p))` statements controlled by an `if`, whose condition is read for a test that is true of a NON-null `p` — an emptiness test that dereferences it (`p.Count == 0`, `!p.Any()`) or a BCL predicate documented true of the empty value (`string.IsNullOrEmpty(p)`). Deterministic, provable per finding. Advisory.
Other · Code Health — Whether a `when` guard is free of side effects — a guard that increments a counter or assigns while deciding whether its arm matches applies that change during PATTERN MATCHING, on an arm that may not be selected, and skips it entirely when a short-circuit to its left answers first.
Method: Roslyn syntax: `when` guards on case labels and switch-expression arms, read for a mutation (`++`/`--`/assignment) sitting in a position the guard's own `&&`/`||`/`??`/`?:`/`?.` can skip. Deterministic, provable per finding. Advisory.
Other · Code Health — Whether a method that TAKES a lock or semaphore and gives it back from a flag-guarded `finally` returns the value that flag implies — reporting success while the guard hands the primitive back admits a second caller the exclusion was there to keep out, and reporting failure while the guard keeps it leaves nothing to ever give it back.
Method: Roslyn syntax: `try` statements whose `finally` releases a synchronisation primitive under a bare local-bool guard, where the method also TOOK that same primitive before the `try`, checked for a `return` of a bool literal whose value disagrees with the flag state the method's own straight-line assignments put it in. Deterministic, provable per finding. Advisory.
Other · Code Health — Whether the work a diagnostic log line costs is paid only when that line is wanted — C# evaluates a call's arguments BEFORE the call, so a trace/debug message joined or projected out of a collection is built in full on every pass, and then discarded by a sink the shipped configuration leaves switched off.
Method: Roslyn syntax: log calls at a diagnostic level (a `Log`-prefixed method naming Trace/Debug/Verbose, or a bare `Debug`/`Trace`/`Verbose` on a receiver named for a logger), whose argument list is read for a call whose cost scales with a sequence — a LINQ operator, a materialisation, `string.Join`, a serializer — with no enclosing level check or conditional-compilation region. On a repository with no .NET source it reads JavaScript/TypeScript off the engine’s own token stream (test, vendored and minified paths not) with the same rule: a `debug`/`trace`/`verbose` call on a receiver named for a logger, or a `log`-prefixed method naming the level, whose argument calls an array operator (`map`/`filter`/`reduce`/`sort`/…), `Array.from`, `Object.keys/values/entries`, `JSON.stringify`/`util.inspect` over anything but a literal, or an array `join` — outside any arrow or function passed as an argument, which the logger calls only when the level is on — with no enclosing `if`, `&&` or `?:` whose condition names a level, a level string, or the `NODE_ENV`/`__DEV__`/`DEV` build switch. Deterministic, provable per finding. Advisory.
Do you agree with this assessment?
X24 · Document value interpolated into markup unescaped10.0 / 10Exemplary○ Nothing flagged
Other · Security — Whether text read out of the document being converted is escaped before it is written into generated markup — a value the document's author chose, interpolated into an attribute the surrounding literal delimits, can close that attribute and open another.
Method: Roslyn semantic model over the whole compilation: a string-typed `Value`/`InnerText`/`InnerXml`/`Text` member declared inside `DocumentFormat.OpenXml` or `System.Xml` is a taint SOURCE, propagated through assignments, returns, arguments, tuple elements and string composition to its transitive closure, then read at interpolated-string holes that sit in a markup position the surrounding literal itself delimits. Escaper/encoder calls and enclosing validator conditions cut the flow. Flow- and container-insensitive by construction. A second arm needs no provenance at all and reports a type that CONTRADICTS ITSELF — the same expression escaped at one delimited markup hole and interpolated raw at another hole in the same markup position of the same type, which the type's own escaping proves is a defect without knowing where the value came from. On a repository with no .NET source it reads JavaScript/TypeScript off the token stream with the same rule: a DOM read of raw document text (`getAttribute`, `textContent`, `innerText`, `nodeValue`) is the source, propagated through local bindings and string composition, and judged at template-literal and concatenation holes in the same two delimited markup positions; escapers and validating conditions cut it, and documentation-site, test, vendored and minified scripts are not read. Deterministic, provable per finding. Advisory.
Other · Code Health — Whether a value the caller is invited to supply is the value the type actually uses — a constructor parameter stored in a private field that nothing ever reads while the default it was given is spelled out a second time at the site that should have read it, a keyed lookup that falls back to a different setting than the one its key names while the same type falls back to the matching one for that same key, or a culture-sensitive parse given no format provider by a type that feeds its own settable culture to the same kind of parse elsewhere, or a number an entry point writes to standard output under the host's culture while the same entry point formats with the invariant culture elsewhere. Either way, every caller who supplies a value silently gets something else.
Method: Roslyn syntax: private instance fields of a non-partial type assigned in a constructor from one of its own parameters with a `??` fallback, checked for whether anything in the type body reads the field and whether that same fallback expression is spelled out again outside the constructor; and `??` fallbacks onto a member access from a lookup call carrying exactly one string literal, grouped by that key across the type and checked for a fallback member whose folded name disagrees with the key while a sibling site for the same key agrees with it. On a repository with no .NET source the first two arms read JavaScript/TypeScript off the engine’s own token stream (tests included, bundles and vendored paths not): a `#x`, `private` or `private` parameter-property instance field filled in the constructor from a parameter (or one member of one) through `??`/`||` or a parameter default, never read anywhere in the file by name, whose constructed default is spelled again in the class body; and `lookup("key") ?? s.member` grouped by key per class, or per module outside every class. The culture arm has no JavaScript counterpart: its parses take no locale. Deterministic, provable per finding. Advisory.
Other · Code Health — Whether a value handed from a callback to the body that waits on it crosses on something built to be crossed — a `Queue<T>`/`List<T>`/`Dictionary<K,V>` written inside an event handler and read back outside it is mutated by two flows at once, and the semaphore or completion source beside it orders how MANY items exist while leaving the collection's own head, tail and backing array unprotected.
Method: Roslyn syntax: method, accessor, local-function and lambda bodies that declare BOTH a non-thread-safe generic collection (`Queue`/`Stack`/`List`/`Dictionary`/`HashSet`/`Sorted*`/`LinkedList`) and a synchronisation primitive (`SemaphoreSlim`/`TaskCompletionSource`/`ManualResetEvent(Slim)`/`AutoResetEvent`/`CountdownEvent`) as locals, then read for a `+=`-registered lambda that raises that primitive while the body outside every lambda waits on it — and, in that scope, a mutating call on the collection inside the lambda paired with a mention of it outside. Any `lock` in the scope abstains it. On a repository with no .NET source the same handoff is read in Java off the engine’s own token stream (test source sets, vendored and demonstration paths not): an `ArrayList`/`LinkedList`/`ArrayDeque`/`PriorityQueue`/`Hash*`/`LinkedHash*`/`Tree*` local and a `CountDownLatch`/`Semaphore`/`CompletableFuture` local, a lambda or anonymous class that raises the primitive and mutates the collection, and a wait outside it; any `synchronized` or `lock()` abstains the body. Because each of those primitives orders what the callback wrote before raising it, only a touch that provably overlaps the callback is convicted: one after the registration and before the next wait, or one in a loop registered-before, waiting on every pass and not declaring the collection. Deterministic, provable per finding. Advisory.
Do you agree with this assessment?
X27 · Collection changed while being enumerated10.0 / 10Exemplary○ Nothing flagged
Other · Code Health — Whether an enhanced `for` leaves the collection it is walking alone — a body that adds to or removes from the very collection the loop is iterating invalidates the fail-fast iterator it is holding, so the next `next()` throws `ConcurrentModificationException` and the remaining elements are never seen. Whether a `for` loop leaves the `dict`, `set` or `deque` it is walking alone — CPython's iterators over those check that the collection has not changed size since the loop began, so a body that removes from (or empties) the very collection the loop is walking makes the next step raise `RuntimeError: dictionary changed size during iteration` and the remaining items are never seen.
Method: Roslyn syntax + semantics: `foreach` statements whose body calls a structural mutator (`Add`/`Remove`/`Clear`/`Insert`/…) on the very expression the loop is enumerating. Two arms. ARM A — the source is a concrete fragile BCL collection, or a live `Keys`/`Values` view over one, and the mutator resolves to that same collection's own member; concurrent and immutable collections and arrays are outside the population by construction, since their enumerators survive a structural change. ARM B — the source is an argument-less accessor CALL on a receiver whose body is in source: the accessor must return a stored field VERBATIM and a sibling member must structurally change that same field, both read off the implementations rather than from the members' names. A mutation the loop provably exits immediately after (`break`/`return`/`throw`/`goto`), or one written inside a nested loop or a lambda, is counted and never reported. Deterministic, provable per finding. Advisory.
Do you agree with this assessment?
X28 · Index access outside its own emptiness guard10.0 / 10Exemplary○ Nothing flagged
Other · Code Health — Whether a condition that tests a value for emptiness indexes that same value only where the test holds — an `||` written one parenthesis too far to the left leaves an index access outside the guard beside it, so the empty case the guard exists to anticipate reaches the index and throws.
Method: Roslyn syntax only, no semantic model: the OUTERMOST `&&`/`||` of every boolean condition, read for a symbol the condition tests for emptiness (`string.IsNullOrEmpty`/`IsNullOrWhiteSpace`, a `Length`/`Count` comparison against a literal, `Any()`, a `Length`/`Count` pattern, or a comparison against `""`) and ALSO indexes. Each `symbol[...]` access is placed by a boolean-reachability walk from the access up to the outermost connective: an access is COVERED when some enclosing step has it in the right operand and the left operand, under the truth value that step forces, proves the symbol non-empty — a recursion over `&&`/`||` whose true- and false-directions are asymmetric. A finding needs BOTH an uncovered access and a covered one on the same symbol in the same condition, which is the agreeing twin that separates a misplaced parenthesis from an unrelated length test. Bare index accesses with no emptiness test in the condition are neither counted nor reported; a non-identifier receiver and a lambda nested inside the condition are outside the population. Deterministic, provable per finding. Advisory.
Do you agree with this assessment?
X29 · Per-element action decided by a fixed element10.0 / 10Exemplary○ Nothing flagged
Other · Code Health — Whether a decision taken once per element is taken ABOUT that element — a test inside a counted loop that reads a fixed subscript of the very collection its guarded statement indexes by the loop variable applies element zero's answer to all of them, so the elements that differ from it are all handled wrongly, and in the same direction.
Method: Roslyn syntax only, no semantic model: every `for` statement declaring exactly ONE loop variable, and every `if` inside its body that is not under a nested loop or a lambda. A site enters the population when the `if`’s condition never mentions the loop variable while the statement it guards indexes some collection by that variable ALONE (`c[i]`; `c[i + 1]` and `c[i, j]` are outside it). A finding additionally needs the AGREEING TWIN at the same-collection grain: the condition must read THAT SAME collection at a subscript that does not move — written into the condition, or reached through a local declared BEFORE the loop, so an alias bound inside the body is not followed. Both collection expressions must be simple identifiers. On a repository with no .NET source the same rule reads JavaScript/TypeScript off the engine’s own token stream (tests included, bundles and vendored paths not): a `for (let|var|const x = …; …; …)` with one declarator and a braced body, an alias followed only when it is declared before the loop in a block that encloses it and never assigned inside the loop. Deterministic, provable per finding. Advisory.
Other · Code Health — Whether exceptions are handled rather than silently swallowed or rethrown with lost stack traces.
Method: Roslyn syntax scan: every catch clause counted; empty catches and bare rethrows flagged. Population is all catch clauses, not estimated. Deterministic, hard fact.
An empty catch block silently discards the error — failures vanish with no log and no rethrow. Log it, handle it, or don't catch it. (×6) — plugins/hanlp_restful_java/src/main/java/com/hankcs/hanlp/restful/HanLPClient.java:611, hanlp/utils/component_util.py:186, hanlp/components/amr/seq2seq/seq2seq_amr_parser.py:250, …
`except Exception` takes every exception and records none of it — the body neither logs it, rethrows it, nor even names it, so the failure is discarded as completely as by an empty catch and only the substituted value survives. Log it through whatever this codebase already uses to report problems, narrow the catch to the exception this call can actually raise, or say in a comment on the catch why the failure genuinely cannot matter. (×3) — hanlp/utils/component_util.py:130, hanlp/components/amr/amrbart/model_interface/tokenization_bart.py:144, hanlp/components/amr/amrbart/model_interface/tokenization_bart.py:166
`except` takes every exception and records none of it — the body neither logs it, rethrows it, nor even names it, so the failure is discarded as completely as by an empty catch and only the substituted value survives. Log it through whatever this codebase already uses to report problems, narrow the catch to the exception this call can actually raise, or say in a comment on the catch why the failure genuinely cannot matter. (×6) — hanlp/utils/component_util.py:145, REDACTED:536, hanlp/optimizers/adamw/optimization.py:100, …
What to do
Swallowed exception (empty catch)
Do you agree with this assessment?
X30 · Support guard that admits what it rejects10.0 / 10Exemplary○ Nothing flagged
Other · Code Health — Whether a guard written as a NEGATED `||` says what its author meant — `!(a || b || x != k)` is `!a && !b && x == k` by De Morgan, so a bail-out that mixes capabilities the code needs with a fault it refuses turns inside out: it fires only where the capabilities are ABSENT, and lets every value the fault term names walk straight into the body that cannot handle it.
Method: Roslyn syntax only, no semantic model: every logical-not whose operand is a parenthesised `||` chain of two or more disjuncts, flattened (a left-nested `a || b || c` read once would see `(a || b)` as one disjunct). A site enters the population on that shape alone. A finding additionally needs the disjuncts to DISAGREE in polarity: at least one bare boolean read — an identifier or member access, never an invocation, which is a predicate rather than a capability flag — and at least one `x != <constant>`, the only form that negates into an exact-value pin (`== null` negates into a looser requirement and is outside the fault set). Consistently-polarised disjunctions, all-fault or all-capability, are counted and never reported; a negated `&&` is outside the population entirely. No same-receiver gate: it was measured to cost a real defect and remove no false positive. Deterministic, provable per finding. Advisory.
Other · Code Health — Whether log calls use message templates (queryable) rather than interpolated strings.
Method: Roslyn syntax scan: every log call-site counted; interpolated-string first-argument violations flagged. Population is all log calls, not estimated. On a repository with no .NET source it reads JavaScript/TypeScript off the engine’s own token stream (test, vendored and minified paths not) with the same rule: a `trace`/`debug`/`info`/`warn`/`error`/`fatal`/`verbose`/`silly`/`http`/`log` call on a receiver named for a logger, in a package whose own or an enclosing `package.json` declares a logger that carries values as fields (pino, pino-http, nestjs-pino, Fastify, winston, bunyan, tslog, LogTape, roarr), whose first deciding argument is a template literal with substitutions; an object-literal fields argument is stepped over, a plain string message clears the call. Python is read the same way: a `debug`/`info`/`warning`/`error`/`exception`/`critical`/`log` call on a logger in a file that imports the standard library’s `logging` or structlog (loguru, which formats at the call, is not judged), whose message is a whole f-string with a non-constant field; a `%`-formatted, `.format()`-ed or concatenated message is not convicted, as in C#. PHP is read the same way: a PSR-3 level call (`emergency`…`debug`, `log`) on a logger (`$logger`, `$this->logger`, Laravel’s `Log::`, `logger()`, `\Drupal::logger()`) in a package whose `composer.json` requires a PSR-3 logger or a framework that hands one out (psr/log, Monolog, Laravel, Symfony, Drupal, CakePHP), or a file that imports one, whose message is a double-quoted string interpolating a variable; a single-quoted message clears the call, and `sprintf` or `.` concatenation is not convicted, as in C#. Dart is read the same way: a `trace`/`debug`/`info`/`warn`/`error`/`fatal` call on Sentry’s structured logger (`Sentry.logger`) whose message is a whole string literal interpolating a value (`$name`, `${…}`); the `Sentry.logger.fmt` template API clears the call. Dart that logs only through loggers with one message and no fields (package:logging, package:logger, talker, dart:developer) is not applicable rather than a reach gap. Ruby is read the same way: a `trace`/`debug`/`info`/`warn`/`error`/`fatal` call on the console gem’s `Console` (declared by a Gemfile, gemspec or Gemfile.lock, or required) or, where SemanticLogger, Ougai or dry-logger is declared or required, on a receiver named for a logger (`logger`, `@logger`, `Rails.logger`, `log`), whose first deciding argument is, as the whole argument, a double-quoted string interpolating a non-constant value (`#{…}`, `#@ivar`); a plain literal clears the call, keyword arguments are the fields, and a block, `%`-format or `+` concatenation is not convicted. Ruby that logs only through the standard Logger (Rails.logger) is not applicable rather than a reach gap. Deterministic.
`debug` at line 339 is handed an f-string as its message, so the logger receives one finished string — the values in it cannot be queried, filtered or aggregated as fields, every occurrence reads as a different message, and the string is built even when the record is filtered out. Pass the values beside a constant message instead: `logger.debug("order %s moved to %s", order_id, status)` with the standard library's `logging`, which keeps the message and its arguments apart on the record, or `log.debug("order moved", order_id=order_id, status=status)` with structlog. — REDACTED:339
`info` at line 52 is handed an f-string as its message, so the logger receives one finished string — the values in it cannot be queried, filtered or aggregated as fields, every occurrence reads as a different message, and the string is built even when the record is filtered out. Pass the values beside a constant message instead: `logger.info("order %s moved to %s", order_id, status)` with the standard library's `logging`, which keeps the message and its arguments apart on the record, or `log.info("order moved", order_id=order_id, status=status)` with structlog. — hanlp/components/taggers/rnn_tagger.py:52
`info` at line 76 is handed an f-string as its message, so the logger receives one finished string — the values in it cannot be queried, filtered or aggregated as fields, every occurrence reads as a different message, and the string is built even when the record is filtered out. Pass the values beside a constant message instead: `logger.info("order %s moved to %s", order_id, status)` with the standard library's `logging`, which keeps the message and its arguments apart on the record, or `log.info("order moved", order_id=order_id, status=status)` with structlog. (×2) — hanlp/components/taggers/rnn_tagger.py:76, hanlp/components/parsers/constituency/crf_constituency_parser.py:76
`info` at line 77 is handed an f-string as its message, so the logger receives one finished string — the values in it cannot be queried, filtered or aggregated as fields, every occurrence reads as a different message, and the string is built even when the record is filtered out. Pass the values beside a constant message instead: `logger.info("order %s moved to %s", order_id, status)` with the standard library's `logging`, which keeps the message and its arguments apart on the record, or `log.info("order moved", order_id=order_id, status=status)` with structlog. (×2) — hanlp/components/taggers/rnn_tagger.py:77, hanlp/components/parsers/constituency/crf_constituency_parser.py:77
`info` at line 75 is handed an f-string as its message, so the logger receives one finished string — the values in it cannot be queried, filtered or aggregated as fields, every occurrence reads as a different message, and the string is built even when the record is filtered out. Pass the values beside a constant message instead: `logger.info("order %s moved to %s", order_id, status)` with the standard library's `logging`, which keeps the message and its arguments apart on the record, or `log.info("order moved", order_id=order_id, status=status)` with structlog. — hanlp/components/taggers/tagger.py:75
`info` at line 96 is handed an f-string as its message, so the logger receives one finished string — the values in it cannot be queried, filtered or aggregated as fields, every occurrence reads as a different message, and the string is built even when the record is filtered out. Pass the values beside a constant message instead: `logger.info("order %s moved to %s", order_id, status)` with the standard library's `logging`, which keeps the message and its arguments apart on the record, or `log.info("order moved", order_id=order_id, status=status)` with structlog. — hanlp/components/taggers/tagger.py:96
`info` at line 97 is handed an f-string as its message, so the logger receives one finished string — the values in it cannot be queried, filtered or aggregated as fields, every occurrence reads as a different message, and the string is built even when the record is filtered out. Pass the values beside a constant message instead: `logger.info("order %s moved to %s", order_id, status)` with the standard library's `logging`, which keeps the message and its arguments apart on the record, or `log.info("order moved", order_id=order_id, status=status)` with structlog. (×3) — hanlp/components/taggers/tagger.py:97, hanlp/components/srl/span_rank/span_rank.py:97, hanlp/components/srl/span_bio/span_bio.py:97
`info` at line 98 is handed an f-string as its message, so the logger receives one finished string — the values in it cannot be queried, filtered or aggregated as fields, every occurrence reads as a different message, and the string is built even when the record is filtered out. Pass the values beside a constant message instead: `logger.info("order %s moved to %s", order_id, status)` with the standard library's `logging`, which keeps the message and its arguments apart on the record, or `log.info("order moved", order_id=order_id, status=status)` with structlog. (×3) — hanlp/components/taggers/tagger.py:98, hanlp/components/sts/transformer_sts.py:98, hanlp/components/ner/biaffine_ner/biaffine_ner.py:98
`info` at line 160 is handed an f-string as its message, so the logger receives one finished string — the values in it cannot be queried, filtered or aggregated as fields, every occurrence reads as a different message, and the string is built even when the record is filtered out. Pass the values beside a constant message instead: `logger.info("order %s moved to %s", order_id, status)` with the standard library's `logging`, which keeps the message and its arguments apart on the record, or `log.info("order moved", order_id=order_id, status=status)` with structlog. — hanlp/components/taggers/transformers/transformer_tagger.py:160
`info` at line 181 is handed an f-string as its message, so the logger receives one finished string — the values in it cannot be queried, filtered or aggregated as fields, every occurrence reads as a different message, and the string is built even when the record is filtered out. Pass the values beside a constant message instead: `logger.info("order %s moved to %s", order_id, status)` with the standard library's `logging`, which keeps the message and its arguments apart on the record, or `log.info("order moved", order_id=order_id, status=status)` with structlog. — hanlp/components/taggers/transformers/transformer_tagger.py:181
`info` at line 86 is handed an f-string as its message, so the logger receives one finished string — the values in it cannot be queried, filtered or aggregated as fields, every occurrence reads as a different message, and the string is built even when the record is filtered out. Pass the values beside a constant message instead: `logger.info("order %s moved to %s", order_id, status)` with the standard library's `logging`, which keeps the message and its arguments apart on the record, or `log.info("order moved", order_id=order_id, status=status)` with structlog. — hanlp/components/sts/transformer_sts.py:86
`info` at line 118 is handed an f-string as its message, so the logger receives one finished string — the values in it cannot be queried, filtered or aggregated as fields, every occurrence reads as a different message, and the string is built even when the record is filtered out. Pass the values beside a constant message instead: `logger.info("order %s moved to %s", order_id, status)` with the standard library's `logging`, which keeps the message and its arguments apart on the record, or `log.info("order moved", order_id=order_id, status=status)` with structlog. — hanlp/components/srl/span_bio/span_bio.py:118
`info` at line 119 is handed an f-string as its message, so the logger receives one finished string — the values in it cannot be queried, filtered or aggregated as fields, every occurrence reads as a different message, and the string is built even when the record is filtered out. Pass the values beside a constant message instead: `logger.info("order %s moved to %s", order_id, status)` with the standard library's `logging`, which keeps the message and its arguments apart on the record, or `log.info("order moved", order_id=order_id, status=status)` with structlog. — hanlp/components/srl/span_bio/span_bio.py:119
`info` at line 120 is handed an f-string as its message, so the logger receives one finished string — the values in it cannot be queried, filtered or aggregated as fields, every occurrence reads as a different message, and the string is built even when the record is filtered out. Pass the values beside a constant message instead: `logger.info("order %s moved to %s", order_id, status)` with the standard library's `logging`, which keeps the message and its arguments apart on the record, or `log.info("order moved", order_id=order_id, status=status)` with structlog. — hanlp/components/srl/span_bio/span_bio.py:120
`info` at line 289 is handed an f-string as its message, so the logger receives one finished string — the values in it cannot be queried, filtered or aggregated as fields, every occurrence reads as a different message, and the string is built even when the record is filtered out. Pass the values beside a constant message instead: `logger.info("order %s moved to %s", order_id, status)` with the standard library's `logging`, which keeps the message and its arguments apart on the record, or `log.info("order moved", order_id=order_id, status=status)` with structlog. — hanlp/components/srl/span_bio/span_bio.py:289
`info` at line 222 is handed an f-string as its message, so the logger receives one finished string — the values in it cannot be queried, filtered or aggregated as fields, every occurrence reads as a different message, and the string is built even when the record is filtered out. Pass the values beside a constant message instead: `logger.info("order %s moved to %s", order_id, status)` with the standard library's `logging`, which keeps the message and its arguments apart on the record, or `log.info("order moved", order_id=order_id, status=status)` with structlog. — hanlp/components/parsers/ud/ud_parser.py:222
`info` at line 243 is handed an f-string as its message, so the logger receives one finished string — the values in it cannot be queried, filtered or aggregated as fields, every occurrence reads as a different message, and the string is built even when the record is filtered out. Pass the values beside a constant message instead: `logger.info("order %s moved to %s", order_id, status)` with the standard library's `logging`, which keeps the message and its arguments apart on the record, or `log.info("order moved", order_id=order_id, status=status)` with structlog. — hanlp/components/parsers/ud/ud_parser.py:243
`info` at line 244 is handed an f-string as its message, so the logger receives one finished string — the values in it cannot be queried, filtered or aggregated as fields, every occurrence reads as a different message, and the string is built even when the record is filtered out. Pass the values beside a constant message instead: `logger.info("order %s moved to %s", order_id, status)` with the standard library's `logging`, which keeps the message and its arguments apart on the record, or `log.info("order moved", order_id=order_id, status=status)` with structlog. — hanlp/components/parsers/ud/ud_parser.py:244
`info` at line 245 is handed an f-string as its message, so the logger receives one finished string — the values in it cannot be queried, filtered or aggregated as fields, every occurrence reads as a different message, and the string is built even when the record is filtered out. Pass the values beside a constant message instead: `logger.info("order %s moved to %s", order_id, status)` with the standard library's `logging`, which keeps the message and its arguments apart on the record, or `log.info("order moved", order_id=order_id, status=status)` with structlog. — hanlp/components/parsers/ud/ud_parser.py:245
`info` at line 55 is handed an f-string as its message, so the logger receives one finished string — the values in it cannot be queried, filtered or aggregated as fields, every occurrence reads as a different message, and the string is built even when the record is filtered out. Pass the values beside a constant message instead: `logger.info("order %s moved to %s", order_id, status)` with the standard library's `logging`, which keeps the message and its arguments apart on the record, or `log.info("order moved", order_id=order_id, status=status)` with structlog. — hanlp/components/parsers/constituency/crf_constituency_parser.py:55
`info` at line 78 is handed an f-string as its message, so the logger receives one finished string — the values in it cannot be queried, filtered or aggregated as fields, every occurrence reads as a different message, and the string is built even when the record is filtered out. Pass the values beside a constant message instead: `logger.info("order %s moved to %s", order_id, status)` with the standard library's `logging`, which keeps the message and its arguments apart on the record, or `log.info("order moved", order_id=order_id, status=status)` with structlog. — hanlp/components/parsers/constituency/crf_constituency_parser.py:78
`info` at line 148 is handed an f-string as its message, so the logger receives one finished string — the values in it cannot be queried, filtered or aggregated as fields, every occurrence reads as a different message, and the string is built even when the record is filtered out. Pass the values beside a constant message instead: `logger.info("order %s moved to %s", order_id, status)` with the standard library's `logging`, which keeps the message and its arguments apart on the record, or `log.info("order moved", order_id=order_id, status=status)` with structlog. — hanlp/components/mtl/multi_task_learning.py:148
`info` at line 286 is handed an f-string as its message, so the logger receives one finished string — the values in it cannot be queried, filtered or aggregated as fields, every occurrence reads as a different message, and the string is built even when the record is filtered out. Pass the values beside a constant message instead: `logger.info("order %s moved to %s", order_id, status)` with the standard library's `logging`, which keeps the message and its arguments apart on the record, or `log.info("order moved", order_id=order_id, status=status)` with structlog. — hanlp/components/mtl/multi_task_learning.py:286
`info` at line 307 is handed an f-string as its message, so the logger receives one finished string — the values in it cannot be queried, filtered or aggregated as fields, every occurrence reads as a different message, and the string is built even when the record is filtered out. Pass the values beside a constant message instead: `logger.info("order %s moved to %s", order_id, status)` with the standard library's `logging`, which keeps the message and its arguments apart on the record, or `log.info("order moved", order_id=order_id, status=status)` with structlog. — hanlp/components/mtl/multi_task_learning.py:307
`info` at line 112 is handed an f-string as its message, so the logger receives one finished string — the values in it cannot be queried, filtered or aggregated as fields, every occurrence reads as a different message, and the string is built even when the record is filtered out. Pass the values beside a constant message instead: `logger.info("order %s moved to %s", order_id, status)` with the standard library's `logging`, which keeps the message and its arguments apart on the record, or `log.info("order moved", order_id=order_id, status=status)` with structlog. — hanlp/components/eos/ngram.py:112
`info` at line 153 is handed an f-string as its message, so the logger receives one finished string — the values in it cannot be queried, filtered or aggregated as fields, every occurrence reads as a different message, and the string is built even when the record is filtered out. Pass the values beside a constant message instead: `logger.info("order %s moved to %s", order_id, status)` with the standard library's `logging`, which keeps the message and its arguments apart on the record, or `log.info("order moved", order_id=order_id, status=status)` with structlog. — hanlp/components/classifiers/transformer_classifier.py:153
`info` at line 283 is handed an f-string as its message, so the logger receives one finished string — the values in it cannot be queried, filtered or aggregated as fields, every occurrence reads as a different message, and the string is built even when the record is filtered out. Pass the values beside a constant message instead: `logger.info("order %s moved to %s", order_id, status)` with the standard library's `logging`, which keeps the message and its arguments apart on the record, or `log.info("order moved", order_id=order_id, status=status)` with structlog. — hanlp/components/classifiers/transformer_classifier.py:283
`info` at line 147 is handed an f-string as its message, so the logger receives one finished string — the values in it cannot be queried, filtered or aggregated as fields, every occurrence reads as a different message, and the string is built even when the record is filtered out. Pass the values beside a constant message instead: `logger.info("order %s moved to %s", order_id, status)` with the standard library's `logging`, which keeps the message and its arguments apart on the record, or `log.info("order moved", order_id=order_id, status=status)` with structlog. — hanlp/components/amr/seq2seq/seq2seq_amr_parser.py:147
`info` at line 172 is handed an f-string as its message, so the logger receives one finished string — the values in it cannot be queried, filtered or aggregated as fields, every occurrence reads as a different message, and the string is built even when the record is filtered out. Pass the values beside a constant message instead: `logger.info("order %s moved to %s", order_id, status)` with the standard library's `logging`, which keeps the message and its arguments apart on the record, or `log.info("order moved", order_id=order_id, status=status)` with structlog. — hanlp/components/amr/seq2seq/seq2seq_amr_parser.py:172
`info` at line 173 is handed an f-string as its message, so the logger receives one finished string — the values in it cannot be queried, filtered or aggregated as fields, every occurrence reads as a different message, and the string is built even when the record is filtered out. Pass the values beside a constant message instead: `logger.info("order %s moved to %s", order_id, status)` with the standard library's `logging`, which keeps the message and its arguments apart on the record, or `log.info("order moved", order_id=order_id, status=status)` with structlog. — hanlp/components/amr/seq2seq/seq2seq_amr_parser.py:173
`info` at line 174 is handed an f-string as its message, so the logger receives one finished string — the values in it cannot be queried, filtered or aggregated as fields, every occurrence reads as a different message, and the string is built even when the record is filtered out. Pass the values beside a constant message instead: `logger.info("order %s moved to %s", order_id, status)` with the standard library's `logging`, which keeps the message and its arguments apart on the record, or `log.info("order moved", order_id=order_id, status=status)` with structlog. — hanlp/components/amr/seq2seq/seq2seq_amr_parser.py:174
`info` at line 309 is handed an f-string as its message, so the logger receives one finished string — the values in it cannot be queried, filtered or aggregated as fields, every occurrence reads as a different message, and the string is built even when the record is filtered out. Pass the values beside a constant message instead: `logger.info("order %s moved to %s", order_id, status)` with the standard library's `logging`, which keeps the message and its arguments apart on the record, or `log.info("order moved", order_id=order_id, status=status)` with structlog. — hanlp/common/dataset.py:309
`warning` at line 304 is handed an f-string as its message, so the logger receives one finished string — the values in it cannot be queried, filtered or aggregated as fields, every occurrence reads as a different message, and the string is built even when the record is filtered out. Pass the values beside a constant message instead: `logger.warning("order %s moved to %s", order_id, status)` with the standard library's `logging`, which keeps the message and its arguments apart on the record, or `log.warning("order moved", order_id=order_id, status=status)` with structlog. — hanlp/common/keras_component.py:304
`info` at line 325 is handed an f-string as its message, so the logger receives one finished string — the values in it cannot be queried, filtered or aggregated as fields, every occurrence reads as a different message, and the string is built even when the record is filtered out. Pass the values beside a constant message instead: `logger.info("order %s moved to %s", order_id, status)` with the standard library's `logging`, which keeps the message and its arguments apart on the record, or `log.info("order moved", order_id=order_id, status=status)` with structlog. — hanlp/common/keras_component.py:325
Other · Code Health — Whether Java packages are null-marked and not undermined by heavy nullness suppression — a package under `@NullMarked` (or `@NonNullApi`, `@ParametersAreNonnullByDefault`, `@NonNullByDefault`, or a NullAway/Checker Framework configuration naming it) makes an unannotated reference exclude null, and `@SuppressWarnings("NullAway")`, `("nullness")`, `("null")` and `castToNonNull(…)` are what silence it.
Method: Roslyn compiler-options scan: NullableContextOptions per project; null-forgiving (!) suppression density per 1k syntax nodes. Deterministic, adoption plus suppression penalty.
0/2 of this repository's Java package directories are under a non-null default: a `package-info.java` declaring `@NullMarked` (JSpecify) or `@NonNullApi`, `@ParametersAreNonnullByDefault`, `@NonNullByDefault`, `@DefaultQualifier(NonNull.class)`; a `module-info.java` declaring `@NullMarked`; every type in it `@NullMarked`; or a build that runs NullAway with the package in `AnnotatedPackages`, or the Checker Framework's `NullnessChecker`. Elsewhere every reference admits null and no checker warns on the dereference that throws a `NullPointerException`.
What to do
Declare the default where Java declares it: `@NullMarked` (org.jspecify.annotations) in each package's `package-info.java`, or once in `module-info.java`, and run NullAway (Error Prone) or the Checker Framework so the declaration is enforced. Then annotate the real exceptions `@Nullable` and resolve the suppressions rather than adding to them: a `@SuppressWarnings("NullAway")` or `castToNonNull(x)` on a value the checker could show to be null is a `NullPointerException` waiting for the path that makes it true — where the value genuinely cannot be null, say something the checker can see instead (a null check, `Objects.requireNonNull`, a non-null field initialised in the constructor).
Do you agree with this assessment?
Reference — by lens
The score is the rank-weighted fold of these lenses (worst-heaviest), each including its meta-dimensions; a lens with a Critical contributor is capped at Fair (its band reads "gated by …") and is never the strongest area however high its average.
Unscored — 1 check(s) recorded observations but carry no score
These checks ran and found something, but they do not carry a score — either by design (an advisory check reports evidence rather than grading it) or because they could not be scored here. They are excluded from the score for that reason, not because there was nothing to see.
X10 Duplicated predicate — 22 observation(s) recorded · Advisory — this card reports evidence and never carries a score.
Not evidenced — 1 control(s) we could not find positive evidence for
These checks grade a working control, and the repository shows no evidence of one. That is deliberately not scored as a zero: a repository cannot show an ops runbook, a database TTL or an infrastructure-side audit log, so absence of evidence here is not evidence the control is missing. It is also not a statement that the check is irrelevant to this codebase — the thing it grades applies; we just could not see it. Excluded from the score either way.
P5 DR & Backup — not evidenced — repo shows no backup/RTO/RPO controls; absence of evidence is not evidence of a working control
Not included — 61 check(s) not relevant to this codebase
These checks had nothing to measure here (no tests, no git history, the codebase is small, or the architecture style doesn't apply), so they're omitted above rather than scored low.
AC1 Text alternatives — No web markup found — accessibility is not applicable to this repository.
AC2 Forms & labels — No web markup found — accessibility is not applicable to this repository.
AC3 Page structure — No web markup found — accessibility is not applicable to this repository.
AC4 Keyboard semantics — No web markup found — accessibility is not applicable to this repository.
AC5 ARIA correctness — No web markup found — accessibility is not applicable to this repository.
AC6 Visual & motion safety — No web markup found — accessibility is not applicable to this repository.
AC7 A11y enforcement — No web markup found — accessibility is not applicable to this repository.
AX1 Captive dependencies — Not applicable: this repository's Java imports no dependency-injection container and defines no container of its own — nothing that both registers and resolves bindings, and nothing that names two lifetimes — so nothing holds one lifetime's instance while handing out another's; this repository's Python imports no dependency-injection container and defines no container of its own — nothing that both registers and resolves bindings, and nothing that names two lifetimes — so nothing holds one lifetime's instance while handing out another's.
AX2 Stateful singletons — No container singleton was found, so there is no shared instance for concurrent requests to race on. No class in this Java code is a Spring, JSR-330 or CDI singleton bean. No function in this Python code is served by a threaded web framework (a Flask, Bottle or FastAPI route, a Django or Pyramid view), so no module is shared between request threads.
AX5 Architecture & structure — not assessed — architecture style/structure is computed from a project graph (projects, types, module namespaces) that was not loaded for this repository, because the repository is written in a language this check does not yet model, or because its projects failed to load. This is a gap in the analyzer, not a finding about this repository
AX6 Interface segregation — no public interfaces
AX7 Slice cohesion — not applicable — not a vertical-slice architecture
AX8 Test isolation — no test/production split to check
AX9 CQS / query purity — no CQRS query handlers detected — query purity is not applicable to this codebase
AXB1 Runtime evidence locked — no reproducible boot — This repository has nothing the runtime tiers could boot or serve — no markup, no UI framework or web-server dependency, no UI component source, no native UI project and no API definition — nothing here is a surface to boot — so runtime a11y/egress/header evidence has no subject here. Not applicable: this is neither a gap in the scan nor a finding about your code.
AXB2 Runtime readiness — Advisory — this card reports evidence and never carries a score, so there is nothing missing here.
C1 Data Protection — Not assessed: these personal data controls are read from a source model (declarative annotations, request middleware, entity/column names, guard methods) that was not loaded for this repository — because the repository is written in a language this check does not yet model, or because its projects failed to load. Absence of an idiom this check recognises is NOT evidence that this repository lacks personal data controls: it may implement them entirely in its own ecosystem. This is a gap in the analyzer's language coverage, not a finding about this repository.
C2 Access Controls — No access-control surface detected in the analyzed source — this repository registers no route and starts no server another machine can reach (no Flask/FastAPI/Starlette route, Django urlpatterns or class-based view, Tornado handler). Access control is therefore N/A here — a library or command-line tool is authorized by the application that serves it, not by itself. If this codebase grows request handlers, the dimension reactivates and a default-deny posture is expected then.
C3 Audit Trail — No personal data detected in a persisted data model — no PII-named field (email, first_name, date_of_birth, phone, …) or stored credential in a Django, SQLAlchemy, SQLModel, Peewee, Tortoise or MongoEngine model, a migration or a SQL schema — and no database or data-store client (Django ORM, SQLAlchemy, psycopg, PyMongo, …) in the source either, so this repository keeps no sensitive data whose changes an audit trail would record. If it does (through a store configured outside this repository, for instance), the trail belongs to wherever that data is stored.
C4 Data Retention — No personal data detected in a persisted data model — no PII-named field (email, first_name, date_of_birth, phone, …) or stored credential on a Django, SQLAlchemy, SQLModel, Peewee, Tortoise or mongoengine model, in a migration or in a SQL schema — and no database or data-store client (django.db, sqlalchemy, psycopg2, pymongo, sqlite3, …) imported either, so this repository keeps no personal data whose lifetime a retention limit would bound.
C5 Data-Subject Rights — No personal data detected in a persisted data model — no PII-named field (email, first_name, date_of_birth, phone, …) or stored credential on a Django, SQLAlchemy, SQLModel, Peewee, Tortoise or mongoengine model, in a migration or in a SQL schema — and no database or data-store client (django.db, sqlalchemy, psycopg2, pymongo, sqlite3, …) imported either, so this repository keeps no personal data for a data subject to erase, export or consent over.
D10 Test Quality — ~483 lines of test source are present (.py) but no test cases reached the test census for this repository — no test root we could resolve declared them, or the files we read declare no cases in a test framework we recognise — so skipped/assertion-free tests couldn't be counted. Not scored — this is a gap in the analyzer, not a finding about this repository.
D12 Dependency Hygiene — Not scored — 1 Maven/Gradle manifest(s) were read, stating a version for 2 dependency(ies) and leaving 0 to a parent or an imported BOM, but no build resolved the dependency graph, so the outdated signal was not established. NOT a finding that these dependencies are current or healthy.
D16 Bus Factor — single-maintainer repository — bus factor is not applicable
D18 Solution Shape — D18 scores the shape of a .NET solution; this repository has no .NET solution or project files, so the dimension does not apply.
D20 ADR Quality — N/A — ADRs are expected on deployable products with a user-facing host, not consumed libraries; no ADR log is required here.
D21 Naming Consistency — Naming consistency is measured on C# only; Java, Python not measured.
D22 Internal API Consistency — The exposed public-API surface could not be collected — no C#/VB projects loaded.
D23 Boundary Type-Coupling — No bounded-context organisation was detected either — neither a context-shaped layout, nor 2+ sibling source directories each declaring an aggregate root, nor 2+ projects each declaring an entity of their own. Aggregate roots were found in only one source directory (hanlp), which is a layered codebase, not two contexts. Declaring this codebase's bounded contexts (≥2) would let cross-boundary type coupling be assessed. Declare them in `.codehealth/config.yaml` at the repository root (create it if absent), mapping each context name to the module-path or namespace prefixes that belong to it — e.g. `architecture:` → `contexts:` → `Billing: ["src/billing", "Acme.Billing"]`, `Catalog: ["src/catalog", "Acme.Catalog"]`.
D24 Comment Value — D24 reads C# and Visual Basic comments only — this repository's production source is .java, .py, which was left unread. Not scored: this is a gap in the analyzer, not a verdict about this repository.
D25 ADR Conformance — no ADRs to check
D27 Navigability — D27 counts a call into the standard library or a dependency as unresolved for this repository's language (.py), so too few call sites counted as resolved to assess navigability. Not scored: this is a gap in the analyzer, not a verdict about this repository.
D31 IaC & Container Security — No Infrastructure-as-Code or container manifests found (Dockerfile, Docker Compose, Terraform, Kubernetes/Helm, CloudFormation, ARM, Bicep, Ansible); nothing to scan.
D32 Data Compliance (PII/GDPR) — No personal data was found crossing a boundary the PII/GDPR ruleset checks — nothing written to a log or console sink, placed in a URL or query string, or persisted to browser storage. That is a clean result for the LEAK surface only: this ruleset detects personal data escaping, it does not inventory the personal data a repository holds, so it is not evidence that this repository has no personal-data surface. The personal-data map (Appendix C) and the C1-C5 compliance cards are what speak to that.
D37 Vulnerability-disclosure Policy — No vulnerability-disclosure policy file found (SECURITY.md/.markdown/.rst/.txt at root or under .github/.forgejo/.gitea/docs, .well-known/security.txt). A coordinated-disclosure policy may live off-repo, so this is not evidenced rather than failed.
D39 IL Efficiency — D39 measures the IL emitted by a .NET build; this repository has no .NET solution or project files, so the dimension does not apply.
D40 Network Egress Confinement — No Kubernetes/orchestration workloads found in the repository manifests; network egress policy is a cluster-native control that may live at the platform/firewall layer, so there is nothing to assess here.
D41 Kernel & Syscall Confinement — No Kubernetes/orchestration workloads found in the repository manifests; seccomp/AppArmor/SELinux confinement is a workload-level control, so there is nothing to assess here.
D42 Runtime Threat Enforcement — No Kubernetes/orchestration workloads found in the repository manifests; runtime threat-detection and admission-control policy are cluster-level controls, so there is nothing to assess here.
D7 Architectural Integrity — no checkable ADRs, and no project-reference graph for the cycle pass to read — so this dimension makes no claim about dependency cycles in either direction (where this repository's language has an import-cycle lens, cycles are reported there). Architectural integrity not assessed
D8 Code Coverage — Coverage not included — suite not readable by the collector
ED1 Event-Driven — not scored — this repository shows none of the 3 signals this lens looks for
ED5 Idempotency — no mutating command handlers or message consumers detected — idempotency check not applicable
ES1 Event Sourcing — not scored — this repository shows none of the 3 signals this lens looks for
GD1 Unfinished & placeholder code — no source files were read — this check reads C#, F#, JavaScript/TypeScript, Dart, Ruby and Scala, and none was read for this repository's product. That is a limit of the analyzer, not a finding about your code.
IC1 Incompleteness & stubs — not analysed — these correctness checks read a source model that was not loaded for this repository, because the repository is written in a language this check does not yet model, or because its projects failed to load. This is a gap in the analyzer, not a finding about this repository
M4 Documentation accuracy — Not applicable: the 162 current-state document(s) make no claim of a kind M4 checks. 8 claim(s) of kinds M4 does not check were not checked (path 1, environment variable 7); M4 checks only links and versions.
P12 CI test-gate honesty — Reported, not scored — and nothing was matched here. The coverage check applies to any stack, and the automatic-re-run check to any GitHub-Actions workflow, but the checks for excluded tests, skipped tests and sleep-based synchronisation currently recognise only some ecosystems' test-runner idioms, so on a repository built with another stack the zeros below mean 'not checked', not 'clean'.
P2 Observability — This repository's Python, JVM source (6 module(s), 479 file(s) read) declares no entry point and bootstraps no server, and nothing here deploys a service — it is a library, run inside whatever hosts it, so production observability (structured logging, tracing/metrics, health checks) is N/A. If it grows a binary or a service, the dimension reactivates.
P7 Outbound HTTP resilience — not applicable — no HTTP server, API framework or worker entry point was found in the JVM, Python source, so there is no service whose uptime a failing dependency could take down
P8 Schema migrations — no ORM, schema-migration tool or schema auto-create was found in this repository's dependency manifests or source, so there is no database schema for this check to judge
P9 Domain vs controller coverage — no coverage report found on disk — produce a coverage report in a standard format (`coverage run -m pytest` then `coverage xml`) and commit it — a hosted scan measures a clone of the repository, so a report that exists only in a working tree, a CI runner's or your own, never reaches it; the artefact is commonly gitignored, so `git add -f` that one file (or un-ignore its path) and commit it alongside the code it measures, or wire coverage collection into CI, to enable this cross-layer check
PF1 Benchmark discipline — Not applicable: no benchmark suite was found. This check searched for a pytest-benchmark or pytest-codspeed dependency with tests taking the `benchmark` fixture, an asv.conf.json with its `time_`/`mem_`/`peakmem_` benchmarks, or a pyperf Runner, and for a `*benchmark*` script that this repository's CI runs. Benchmarks are credited as a bonus, so their absence is neither scored nor deducted.
PF2 Allocation hygiene — Not applicable: allocation awareness is rated where code engineers for it — at least 400 production lines that ship benchmarks or already use allocation-aware idioms (8 or more). This repository's Java code has 918 production line(s), 0 such use(s) and no benchmarks, so there is no allocation work to rate. The card is reward-only: its absence costs nothing.
PF3 Async & latency hygiene — Not applicable: this repository declares no async functions, so there is no asynchronous code for a blocking call to stall.
S1 Web-Security Posture — Not assessed: these web-security controls are read from a source model (declarative annotations, request middleware, entity/column names, guard methods) that was not loaded for this repository — because the repository is written in a language this check does not yet model, or because its projects failed to load. Absence of an idiom this check recognises is NOT evidence that this repository lacks web-security controls: it may implement them entirely in its own ecosystem. This is a gap in the analyzer's language coverage, not a finding about this repository.
X2 Cancellation propagation — no C# was loaded; this repository's Java and Python were read, and no request in it can be cancelled by a caller — a Java request off Android takes no token: it is cancelled through the handle it returns (Future.cancel, Call.cancel, a Disposable), ambiently (io.grpc.Context) or by interrupting its thread, not through a parameter; a Python request takes no token: asyncio cancels the task and trio or anyio the scope that made it, from outside the call, and requests, httpx, aiohttp and gRPC take a timeout, not a cancellation
X32 Type resolved by simple name across every loaded assembly — This check is about how a .NET program searches the assemblies loaded into its process for a type, and this repository contains no .NET source, so there is nothing here for it to assess. Not a gap in the analyzer and not a finding about your code.
X6 Hand-rolled structured-format parsing — Advisory — this card reports evidence and never carries a score, so there is nothing missing here.
X7 Silent fallback defaults — Advisory — this card reports evidence and never carries a score, so there is nothing missing here.
X9 Subsumed condition operand — Advisory — this card reports evidence and never carries a score, so there is nothing missing here.
Appendix A — Findings (grouped)
The findings behind the scores, grouped by severity, then by dimension and kind. The high-severity issues are enumerated in full below; items per group are capped at 25 with any overflow stated explicitly per group, never silently truncated. The complete machine-readable list of every finding (all severities) is the companion findings.md in this report's bundle.
Duplicated block (6 lines × 2) hanlp/components/amr/amrbart/model_interface/modeling_bart.py:1005— hanlp/components/amr/amrbart/model_interface/modeling_bart.py:1005-1010 | hanlp/components/amr/amrbart/model_interface/modeling_bart.py:1214-1219 — both copies are in the same file, so extract the block into one function there and call it from each site — the copies drift apart the first time only one of them is edited.
Duplicated block (6 lines × 2) hanlp/components/amr/seq2seq/dataset/tokenization_bart.py:194— hanlp/components/amr/seq2seq/dataset/tokenization_bart.py:194-199 | hanlp/components/amr/seq2seq/dataset/tokenization_t5.py:184-189 — before extracting anything, compare `hanlp/components/amr/seq2seq/dataset/tokenization_bart.py` and `hanlp/components/amr/seq2seq/dataset/tokenization_t5.py` as WHOLE FILES: this scan already matched 21 separate duplicated blocks between them, totalling at least 476 lines, which is the signature of one file having been copied from the other rather than of a helper waiting to be extracted. If that is what happened, the fix is to keep one copy and have the other call it (or delete it), which resolves this row and its siblings together — extracting one helper per block leaves the fork in place. The `return` at the foot of the matched lines is the enclosing body's own terminal exit, not an early one: it moves with them unchanged, and each site calls the extracted unit from the position that `return` occupied — no decision has to be handed back and re-acted on.
Duplicated block (6 lines × 2) hanlp/metrics/chunking/bmes_tf.py:35— hanlp/metrics/chunking/bmes_tf.py:35-42 | hanlp/metrics/f1.py:22-27 — the copies span different directories, so extracting a shared function means choosing where it lives: put it somewhere both call sites can already reach — a location they all depend on today, or a new shared one if there is none — and call it from each site; until then, every change has to be made twice. The `return` at the foot of the matched lines is the enclosing body's own terminal exit, not an early one: it moves with them unchanged, and each site calls the extracted unit from the position that `return` occupied — no decision has to be handed back and re-acted on. Note that the copies do not run to the end of the range shown: their LAST lines are different code, not the same code under different names — the matched region ends inside that line. Extract the lines above it, and read the last line of each site separately.
Duplicated block (6 lines × 2) hanlp/transform/conll_tf.py:56— hanlp/transform/conll_tf.py:56-61 | hanlp/transform/conll_tf.py:295-300 — both copies are in the same file, so extract the block into one function there and call it from each site — the copies drift apart the first time only one of them is edited.
Duplicated block (6 lines × 2) hanlp/common/keras_component.py:372— hanlp/common/keras_component.py:372-377 | hanlp/components/taggers/transformers/transformer_tagger_tf.py:75-83 — the copies span different directories, so extracting a shared function means choosing where it lives: put it somewhere both call sites can already reach — a location they all depend on today, or a new shared one if there is none — and call it from each site; until then, every change has to be made twice. The `return` at the foot of the matched lines is the enclosing body's own terminal exit, not an early one: it moves with them unchanged, and each site calls the extracted unit from the position that `return` occupied — no decision has to be handed back and re-acted on.
Duplicated block (6 lines × 2) hanlp/common/vocab.py:68— hanlp/common/vocab.py:68-73 | hanlp/common/vocab_tf.py:53-58 — the copies sit in sibling files of one directory, so a shared home is within easy reach: extract the block into a single shared function the call sites can all reach — a file they already depend on, or a new one alongside them — and call it from both call sites, so a change lands once. The matched lines also transfer control out of the body holding them, which cannot survive a move into a called unit unchanged: have the extracted unit return that decision and let each site act on it.
Duplicated block (6 lines × 2) hanlp/components/amr/amrbart/bart_amr_generation.py:125— hanlp/components/amr/amrbart/bart_amr_generation.py:125-130 | hanlp/components/amr/amrbart/bart_amr_parser.py:153-158 — before extracting anything, compare `hanlp/components/amr/amrbart/bart_amr_generation.py` and `hanlp/components/amr/amrbart/bart_amr_parser.py` as WHOLE FILES: this scan already matched 4 separate duplicated blocks between them, totalling at least 35 lines, which is the signature of one file having been copied from the other rather than of a helper waiting to be extracted. If that is what happened, the fix is to keep one copy and have the other call it (or delete it), which resolves this row and its siblings together — extracting one helper per block leaves the fork in place.
Duplicated block (6 lines × 2) hanlp/transform/txt_tf.py:19— hanlp/transform/txt_tf.py:19-24 | hanlp/utils/span_util.py:9-14 — the copies span different directories, so extracting a shared function means choosing where it lives: put it somewhere both call sites can already reach — a location they all depend on today, or a new shared one if there is none — and call it from each site; until then, every change has to be made twice. The matched lines also transfer control out of the body holding them, which cannot survive a move into a called unit unchanged: have the extracted unit return that decision and let each site act on it.
Duplicated block (6 lines × 2) hanlp/metrics/parsing/attachmentscore.py:29— hanlp/metrics/parsing/attachmentscore.py:29-34 | hanlp/metrics/parsing/labeled_score.py:11-16 — the copies sit in sibling files of one directory, so a shared home is within easy reach: extract the block into a single shared function the call sites can all reach — a file they already depend on, or a new one alongside them — and call it from both call sites, so a change lands once.
Duplicated block (6 lines × 2) hanlp/metrics/parsing/labeled_f1.py:11— hanlp/metrics/parsing/labeled_f1.py:11-16 | hanlp/metrics/parsing/labeled_f1_tf.py:10-15 — the copies sit in sibling files of one directory, so a shared home is within easy reach: extract the block into a single shared function the call sites can all reach — a file they already depend on, or a new one alongside them — and call it from both call sites, so a change lands once.
Duplicated block (6 lines × 2) hanlp/components/parsers/alg.py:655— hanlp/components/parsers/alg.py:655-660 | hanlp/components/parsers/constituency/treecrf.py:221-226 — the copies span different directories, so extracting a shared function means choosing where it lives: put it somewhere both call sites can already reach — a location they all depend on today, or a new shared one if there is none — and call it from each site; until then, every change has to be made twice.
Duplicated block (6 lines × 2) hanlp/components/srl/span_rank/layer.py:318— hanlp/components/srl/span_rank/layer.py:318-323 | hanlp/components/srl/span_rank/layer.py:333-338 — both copies are in the same file, so extract the block into one function there and call it from each site — the copies drift apart the first time only one of them is edited. ★ These copies have DRIFTED, and that is worth reading before extracting anything: just after the matched lines, `hanlp/components/srl/span_rank/layer.py:339` calls `reverse` and `hanlp/components/srl/span_rank/layer.py:324` does not — after which the two agree again for 2 more lines. One of those two behaviours is the intended one and the other is what a copy-paste left behind, so decide which BEFORE unifying them: extracting the shared part will silently settle it, and if the copy that skips the call is the wrong one, that bug is already live.
FileTooLong: ud/ud210.py hanlp/datasets/parsing/ud/ud210.py— FileTooLong — 1006 significant lines (blank, comment-only and punctuation-only lines excluded). The bar is 500 significant lines; this is 506 over it, 2.01× the bar. To reduce it, split the file along the responsibilities already in it: move each cohesive group of declarations into its own sibling file in the same module or package, so no one file has to be read whole to change one of them.
FileTooLong: ud/ud27.py hanlp/datasets/parsing/ud/ud27.py— FileTooLong — 842 significant lines (blank, comment-only and punctuation-only lines excluded). The bar is 500 significant lines; this is 342 over it, 1.68× the bar. To reduce it, split the file along the responsibilities already in it: move each cohesive group of declarations into its own sibling file in the same module or package, so no one file has to be read whole to change one of them.
FileTooLong: transform/conll_tf.py hanlp/transform/conll_tf.py— FileTooLong — 688 significant lines (blank, comment-only and punctuation-only lines excluded). The bar is 500 significant lines; this is 188 over it, 1.38× the bar. To reduce it, split the file along the responsibilities already in it: move each cohesive group of declarations into its own sibling file in the same module or package, so no one file has to be read whole to change one of them.
FileTooLong: mtl/multi_task_learning.py hanlp/components/mtl/multi_task_learning.py— FileTooLong — 683 significant lines (blank, comment-only and punctuation-only lines excluded). The bar is 500 significant lines; this is 183 over it, 1.37× the bar. To reduce it, split the file along the responsibilities already in it: move each cohesive group of declarations into its own sibling file in the same module or package, so no one file has to be read whole to change one of them.
FileTooLong: model_interface/modeling_bart.py hanlp/components/amr/amrbart/model_interface/modeling_bart.py— FileTooLong — 679 significant lines (blank, comment-only and punctuation-only lines excluded). The bar is 500 significant lines; this is 179 over it, 1.36× the bar. To reduce it, split the file along the responsibilities already in it: move each cohesive group of declarations into its own sibling file in the same module or package, so no one file has to be read whole to change one of them.
FileTooLong: dataset/tokenization_bart.py hanlp/components/amr/seq2seq/dataset/tokenization_bart.py— FileTooLong — 563 significant lines (blank, comment-only and punctuation-only lines excluded). The bar is 500 significant lines; this is 63 over it, 1.13× the bar. To reduce it, split the file along the responsibilities already in it: move each cohesive group of declarations into its own sibling file in the same module or package, so no one file has to be read whole to change one of them.
FileTooLong: dataset/tokenization_t5.py hanlp/components/amr/seq2seq/dataset/tokenization_t5.py— FileTooLong — 557 significant lines (blank, comment-only and punctuation-only lines excluded). The bar is 500 significant lines; this is 57 over it, 1.11× the bar. To reduce it, split the file along the responsibilities already in it: move each cohesive group of declarations into its own sibling file in the same module or package, so no one file has to be read whole to change one of them.
FileTooLong: ontonotes5/_utils.py hanlp/datasets/srl/ontonotes5/_utils.py— FileTooLong — 548 significant lines (blank, comment-only and punctuation-only lines excluded). The bar is 500 significant lines; this is 48 over it, 1.10× the bar. To reduce it, split the file along the responsibilities already in it: move each cohesive group of declarations into its own sibling file in the same module or package, so no one file has to be read whole to change one of them.
FileTooLong: REDACTED REDACTED— FileTooLong — 525 significant lines (blank, comment-only and punctuation-only lines excluded). The bar is 500 significant lines; this is 25 over it, 1.05× the bar. To reduce it, split the file along the responsibilities already in it: move each cohesive group of declarations into its own sibling file in the same module or package, so no one file has to be read whole to change one of them.
FileTooLong: biaffine/biaffine_dep.py hanlp/components/parsers/biaffine/biaffine_dep.py— FileTooLong — 516 significant lines (blank, comment-only and punctuation-only lines excluded). The bar is 500 significant lines; this is 16 over it, 1.03× the bar. To reduce it, split the file along the responsibilities already in it: move each cohesive group of declarations into its own sibling file in the same module or package, so no one file has to be read whole to change one of them.
FileTooLong: transform/transformer_tokenizer.py hanlp/transform/transformer_tokenizer.py— FileTooLong — 507 significant lines (blank, comment-only and punctuation-only lines excluded). The bar is 500 significant lines; this is 7 over it, 1.01× the bar. To reduce it, split the file along the responsibilities already in it: move each cohesive group of declarations into its own sibling file in the same module or package, so no one file has to be read whole to change one of them.
Duplicated block (9 lines × 2) hanlp/components/amr/seq2seq/dataset/tokenization_bart.py:25— hanlp/components/amr/seq2seq/dataset/tokenization_bart.py:25-33 | hanlp/components/amr/seq2seq/dataset/tokenization_t5.py:26-34 — before extracting anything, compare `hanlp/components/amr/seq2seq/dataset/tokenization_bart.py` and `hanlp/components/amr/seq2seq/dataset/tokenization_t5.py` as WHOLE FILES: this scan already matched 21 separate duplicated blocks between them, totalling at least 476 lines, which is the signature of one file having been copied from the other rather than of a helper waiting to be extracted. If that is what happened, the fix is to keep one copy and have the other call it (or delete it), which resolves this row and its siblings together — extracting one helper per block leaves the fork in place.
Duplicated block (9 lines × 2) hanlp/components/parsers/alg.py:618— hanlp/components/parsers/alg.py:618-626 | hanlp/components/parsers/constituency/treecrf.py:178-186 — the copies span different directories, so extracting a shared function means choosing where it lives: put it somewhere both call sites can already reach — a location they all depend on today, or a new shared one if there is none — and call it from each site; until then, every change has to be made twice.
Duplicated block (9 lines × 2) hanlp/components/parsers/constituency/crf_constituency_parser.py:227— hanlp/components/parsers/constituency/crf_constituency_parser.py:227-235 | hanlp/components/srl/span_bio/span_bio.py:259-267 — the copies span different directories, so extracting a shared function means choosing where it lives: put it somewhere both call sites can already reach — a location they all depend on today, or a new shared one if there is none — and call it from each site; until then, every change has to be made twice. The `return` at the foot of the matched lines is the enclosing body's own terminal exit, not an early one: it moves with them unchanged, and each site calls the extracted unit from the position that `return` occupied — no decision has to be handed back and re-acted on.
Duplicated block (9 lines × 2) hanlp/components/srl/span_rank/highway_variational_lstm.py:55— hanlp/components/srl/span_rank/highway_variational_lstm.py:55-63 | hanlp/components/srl/span_rank/highway_variational_lstm.py:183-191 — both copies are in the same file, so extract the block into one function there and call it from each site — the copies drift apart the first time only one of them is edited. The `return` at the foot of the matched lines is the enclosing body's own terminal exit, not an early one: it moves with them unchanged, and each site calls the extracted unit from the position that `return` occupied — no decision has to be handed back and re-acted on.
Duplicated block (9 lines × 2) hanlp/metrics/parsing/semdep_eval.py:81— hanlp/metrics/parsing/semdep_eval.py:81-89 | hanlp/metrics/parsing/semdep_eval.py:92-100 — both copies are in the same file, so extract the block into one function there and call it from each site — the copies drift apart the first time only one of them is edited.
Duplicated block (9 lines × 2) hanlp/transform/conll_tf.py:412— hanlp/transform/conll_tf.py:412-420 | hanlp/transform/conll_tf.py:752-760 — both copies are in the same file, so extract the block into one function there and call it from each site — the copies drift apart the first time only one of them is edited. The matched lines also transfer control out of the body holding them, which cannot survive a move into a called unit unchanged: have the extracted unit return that decision and let each site act on it.
Duplicated block (9 lines × 2) hanlp/transform/conll_tf.py:434— hanlp/transform/conll_tf.py:434-442 | hanlp/transform/conll_tf.py:543-551 — both copies are in the same file, so extract the block into one function there and call it from each site — the copies drift apart the first time only one of them is edited.
Duplicated block (9 lines × 2) hanlp/transform/tsv_tf.py:31— hanlp/transform/tsv_tf.py:31-39 | hanlp/transform/txt_tf.py:141-149 — the copies sit in sibling files of one directory, so a shared home is within easy reach: extract the block into a single shared function the call sites can all reach — a file they already depend on, or a new one alongside them — and call it from both call sites, so a change lands once. The `return` at the foot of the matched lines is the enclosing body's own terminal exit, not an early one: it moves with them unchanged, and each site calls the extracted unit from the position that `return` occupied — no decision has to be handed back and re-acted on.
Duplicated block (9 lines × 2) hanlp/transform/txt_tf.py:28— hanlp/transform/txt_tf.py:28-36 | hanlp/utils/span_util.py:18-26 — the copies span different directories, so extracting a shared function means choosing where it lives: put it somewhere both call sites can already reach — a location they all depend on today, or a new shared one if there is none — and call it from each site; until then, every change has to be made twice. The `return` at the foot of the matched lines is the enclosing body's own terminal exit, not an early one: it moves with them unchanged, and each site calls the extracted unit from the position that `return` occupied — no decision has to be handed back and re-acted on.
Duplicated block (9 lines × 2) hanlp/components/classifiers/transformer_classifier.py:149— hanlp/components/classifiers/transformer_classifier.py:149-157 | hanlp/components/eos/ngram.py:108-116 — the copies span different directories, so extracting a shared function means choosing where it lives: put it somewhere both call sites can already reach — a location they all depend on today, or a new shared one if there is none — and call it from each site; until then, every change has to be made twice.
Duplicated block (7 lines × 2) hanlp/components/amr/seq2seq/dataset/tokenization_bart.py:229— hanlp/components/amr/seq2seq/dataset/tokenization_bart.py:229-235 | hanlp/components/amr/seq2seq/dataset/tokenization_t5.py:219-225 — before extracting anything, compare `hanlp/components/amr/seq2seq/dataset/tokenization_bart.py` and `hanlp/components/amr/seq2seq/dataset/tokenization_t5.py` as WHOLE FILES: this scan already matched 21 separate duplicated blocks between them, totalling at least 476 lines, which is the signature of one file having been copied from the other rather than of a helper waiting to be extracted. If that is what happened, the fix is to keep one copy and have the other call it (or delete it), which resolves this row and its siblings together — extracting one helper per block leaves the fork in place. The matched lines also transfer control out of the body holding them, which cannot survive a move into a called unit unchanged: have the extracted unit return that decision and let each site act on it. ★ These copies have DRIFTED, and that is worth reading before extracting anything: just after the matched lines, `hanlp/components/amr/seq2seq/dataset/tokenization_t5.py:226` calls `print_exc` and `hanlp/components/amr/seq2seq/dataset/tokenization_bart.py:236` does not — after which the two agree again for 4 more lines. One of those two behaviours is the intended one and the other is what a copy-paste left behind, so decide which BEFORE unifying them: extracting the shared part will silently settle it, and if the copy that skips the call is the wrong one, that bug is already live.
Duplicated block (7 lines × 2) hanlp/components/mtl/tasks/dep.py:124— hanlp/components/mtl/tasks/dep.py:124-130 | hanlp/components/mtl/tasks/ud.py:83-89 — the copies sit in sibling files of one directory, so a shared home is within easy reach: extract the block into a single shared function the call sites can all reach — a file they already depend on, or a new one alongside them — and call it from both call sites, so a change lands once. Read the line range as the matched WINDOW rather than a finished unit: at `hanlp/components/mtl/tasks/dep.py:124` it does not close everything it opens, so those exact lines cannot be lifted as they stand — widen the region to the smallest complete statement or declaration that contains it, and extract that.
Duplicated block (7 lines × 2) hanlp/components/parsers/alg.py:292— hanlp/components/parsers/alg.py:292-298 | hanlp/components/parsers/constituency/treecrf.py:90-96 — the copies span different directories, so extracting a shared function means choosing where it lives: put it somewhere both call sites can already reach — a location they all depend on today, or a new shared one if there is none — and call it from each site; until then, every change has to be made twice. The matched lines also transfer control out of the body holding them, which cannot survive a move into a called unit unchanged: have the extracted unit return that decision and let each site act on it. ★ These copies have DRIFTED, and that is worth reading before extracting anything: just before the matched lines, `hanlp/components/parsers/alg.py:290` calls `new_tensor`, `unsqueeze` and `hanlp/components/parsers/constituency/treecrf.py:88` does not — after which the two agree again for 2 more lines. One of those two behaviours is the intended one and the other is what a copy-paste left behind, so decide which BEFORE unifying them: extracting the shared part will silently settle it, and if the copy that skips the call is the wrong one, that bug is already live.
Duplicated block (7 lines × 2) hanlp/components/srl/span_rank/layer.py:280— hanlp/components/srl/span_rank/layer.py:280-286 | hanlp/components/srl/span_rank/layer.py:291-297 — both copies are in the same file, so extract the block into one function there and call it from each site — the copies drift apart the first time only one of them is edited.
Duplicated block (7 lines × 2) hanlp/layers/embeddings/char_cnn_tf.py:34— hanlp/layers/embeddings/char_cnn_tf.py:34-40 | hanlp/layers/embeddings/char_rnn_tf.py:31-37 — the copies sit in sibling files of one directory, so a shared home is within easy reach: extract the block into a single shared function the call sites can all reach — a file they already depend on, or a new one alongside them — and call it from both call sites, so a change lands once.
Duplicated block (7 lines × 2) hanlp/metrics/chunking/chunking_f1.py:119— hanlp/metrics/chunking/chunking_f1.py:119-125 | hanlp/metrics/chunking/conlleval.py:280-286 — the copies sit in sibling files of one directory, so a shared home is within easy reach: extract the block into a single shared function the call sites can all reach — a file they already depend on, or a new one alongside them — and call it from both call sites, so a change lands once.
Duplicated block (7 lines × 2) hanlp/metrics/chunking/sequence_labeling.py:174— hanlp/metrics/chunking/sequence_labeling.py:174-180 | hanlp/metrics/chunking/sequence_labeling.py:373-379 — both copies are in the same file, so extract the block into one function there and call it from each site — the copies drift apart the first time only one of them is edited.
Duplicated block (7 lines × 2) hanlp/components/mtl/tasks/dep.py:141— hanlp/components/mtl/tasks/dep.py:141-147 | hanlp/components/mtl/tasks/sdp.py:141-147 — the copies sit in sibling files of one directory, so a shared home is within easy reach: extract the block into a single shared function the call sites can all reach — a file they already depend on, or a new one alongside them — and call it from both call sites, so a change lands once. The `return` at the foot of the matched lines is the enclosing body's own terminal exit, not an early one: it moves with them unchanged, and each site calls the extracted unit from the position that `return` occupied — no decision has to be handed back and re-acted on.
Duplicated block (7 lines × 2) hanlp/components/amr/amrbart/model_interface/modeling_bart.py:333— hanlp/components/amr/amrbart/model_interface/modeling_bart.py:333-339 | hanlp/components/amr/amrbart/model_interface/modeling_bart.py:453-459 — both copies are in the same file, so extract the block into one function there and call it from each site — the copies drift apart the first time only one of them is edited.
Duplicated block (7 lines × 2) plugins/hanlp_restful_java/src/main/java/com/hankcs/hanlp/restful/HanLPClient.java:421— plugins/hanlp_restful_java/src/main/java/com/hankcs/hanlp/restful/HanLPClient.java:421-427 | plugins/hanlp_restful_java/src/main/java/com/hankcs/hanlp/restful/HanLPClient.java:454-460 — both copies are in the same file, so extract the block into one function there and call it from each site — the copies drift apart the first time only one of them is edited. Each matched range is the entire body of the declaration above it, so the region is already a complete unit: move that whole declaration to the shared location and have each site call it, rather than lifting the lines out of their bodies. Any `return` inside it is the body's own exit and keeps its meaning in the moved unit.
Duplicated block (5 lines × 2) hanlp/components/amr/amrbart/bart_amr_generation.py:39— hanlp/components/amr/amrbart/bart_amr_generation.py:39-43 | hanlp/components/amr/amrbart/bart_amr_parser.py:43-47 — before extracting anything, compare `hanlp/components/amr/amrbart/bart_amr_generation.py` and `hanlp/components/amr/amrbart/bart_amr_parser.py` as WHOLE FILES: this scan already matched 4 separate duplicated blocks between them, totalling at least 35 lines, which is the signature of one file having been copied from the other rather than of a helper waiting to be extracted. If that is what happened, the fix is to keep one copy and have the other call it (or delete it), which resolves this row and its siblings together — extracting one helper per block leaves the fork in place. Read the line range as the matched WINDOW rather than a finished unit: at `hanlp/components/amr/amrbart/bart_amr_generation.py:39` it begins part-way through the construct above it, so those exact lines cannot be lifted as they stand — widen the region to the smallest complete statement or declaration that contains it, and extract that. The `return` at the foot of the matched lines is the enclosing body's own terminal exit, not an early one: it moves with them unchanged, and each site calls the extracted unit from the position that `return` occupied — no decision has to be handed back and re-acted on.
Duplicated block (5 lines × 2) REDACTED:404— REDACTED:404-408 | REDACTED:530-534 — before extracting anything, compare `REDACTED` and `REDACTED` as WHOLE FILES: this scan already matched 15 separate duplicated blocks between them, totalling at least 404 lines, which is the signature of one file having been copied from the other rather than of a helper waiting to be extracted. The two sit in different directories, so one cannot simply be deleted in favour of the other while both are reached separately: hoist the shared part into a location both already depend on and have each file call it, and retire whichever file turns out to have no caller of its own left. Extracting one helper per block leaves the fork in place. The matched lines also transfer control out of the body holding them, which cannot survive a move into a called unit unchanged: have the extracted unit return that decision and let each site act on it.
Duplicated block (5 lines × 2) hanlp/components/amr/amrbart/model_interface/tokenization_bart.py:117— hanlp/components/amr/amrbart/model_interface/tokenization_bart.py:117-121 | hanlp/components/amr/seq2seq/dataset/tokenization_bart.py:150-154 — before extracting anything, compare `hanlp/components/amr/amrbart/model_interface/tokenization_bart.py` and `hanlp/components/amr/seq2seq/dataset/tokenization_bart.py` as WHOLE FILES: this scan already matched 10 separate duplicated blocks between them, totalling at least 329 lines, which is the signature of one file having been copied from the other rather than of a helper waiting to be extracted. The two sit in different directories, so one cannot simply be deleted in favour of the other while both are reached separately: hoist the shared part into a location both already depend on and have each file call it, and retire whichever file turns out to have no caller of its own left. Extracting one helper per block leaves the fork in place. Read the line range as the matched WINDOW rather than a finished unit: at `hanlp/components/amr/amrbart/model_interface/tokenization_bart.py:117` it begins part-way through the construct above it, so those exact lines cannot be lifted as they stand — widen the region to the smallest complete statement or declaration that contains it, and extract that.
Duplicated block (5 lines × 2) hanlp/components/amr/seq2seq/dataset/tokenization_bart.py:182— hanlp/components/amr/seq2seq/dataset/tokenization_bart.py:182-186 | hanlp/components/amr/seq2seq/dataset/tokenization_t5.py:172-176 — before extracting anything, compare `hanlp/components/amr/seq2seq/dataset/tokenization_bart.py` and `hanlp/components/amr/seq2seq/dataset/tokenization_t5.py` as WHOLE FILES: this scan already matched 21 separate duplicated blocks between them, totalling at least 476 lines, which is the signature of one file having been copied from the other rather than of a helper waiting to be extracted. If that is what happened, the fix is to keep one copy and have the other call it (or delete it), which resolves this row and its siblings together — extracting one helper per block leaves the fork in place. The `return` at the foot of the matched lines is the enclosing body's own terminal exit, not an early one: it moves with them unchanged, and each site calls the extracted unit from the position that `return` occupied — no decision has to be handed back and re-acted on.
Duplicated block (5 lines × 2) hanlp/components/amr/seq2seq/seq2seq_amr_parser.py:145— hanlp/components/amr/seq2seq/seq2seq_amr_parser.py:145-149 | hanlp/components/srl/span_bio/span_bio.py:95-99 — the copies span different directories, so extracting a shared function means choosing where it lives: put it somewhere both call sites can already reach — a location they all depend on today, or a new shared one if there is none — and call it from each site; until then, every change has to be made twice.
Duplicated block (5 lines × 2) hanlp/components/distillation/losses.py:52— hanlp/components/distillation/losses.py:52-56 | hanlp/components/distillation/losses.py:77-81 — both copies are in the same file, so extract the block into one function there and call it from each site — the copies drift apart the first time only one of them is edited.
Duplicated block (5 lines × 2) hanlp/components/parsers/alg_tf.py:83— hanlp/components/parsers/alg_tf.py:83-87 | hanlp/components/parsers/biaffine_tf/alg.py:83-87 — before extracting anything, compare `hanlp/components/parsers/alg_tf.py` and `hanlp/components/parsers/biaffine_tf/alg.py` as WHOLE FILES: this scan already matched 9 separate duplicated blocks between them, totalling at least 185 lines, which is the signature of one file having been copied from the other rather than of a helper waiting to be extracted. The two sit in different directories, so one cannot simply be deleted in favour of the other while both are reached separately: hoist the shared part into a location both already depend on and have each file call it, and retire whichever file turns out to have no caller of its own left. Extracting one helper per block leaves the fork in place. The `return` at the foot of the matched lines is the enclosing body's own terminal exit, not an early one: it moves with them unchanged, and each site calls the extracted unit from the position that `return` occupied — no decision has to be handed back and re-acted on.
Duplicated block (5 lines × 2) hanlp/layers/embeddings/char_cnn_tf.py:25— hanlp/layers/embeddings/char_cnn_tf.py:25-29 | hanlp/layers/embeddings/char_rnn_tf.py:21-25 — the copies sit in sibling files of one directory, so a shared home is within easy reach: extract the block into a single shared function the call sites can all reach — a file they already depend on, or a new one alongside them — and call it from both call sites, so a change lands once.
Duplicated block (5 lines × 2) hanlp/components/ner/biaffine_ner/biaffine_ner_model.py:95— hanlp/components/ner/biaffine_ner/biaffine_ner_model.py:95-99 | hanlp/components/srl/span_rank/span_ranking_srl_model.py:281-285 — the copies span different directories, so extracting a shared function means choosing where it lives: put it somewhere both call sites can already reach — a location they all depend on today, or a new shared one if there is none — and call it from each site; until then, every change has to be made twice.
Swallowed exception (caught, then discarded) hanlp/utils/component_util.py:130— `except Exception` takes every exception and records none of it — the body neither logs it, rethrows it, nor even names it, so the failure is discarded as completely as by an empty catch and only the substituted value survives. Log it through whatever this codebase already uses to report problems, narrow the catch to the exception this call can actually raise, or say in a comment on the catch why the failure genuinely cannot matter.
Swallowed exception (caught, then discarded) hanlp/utils/component_util.py:145— `except` takes every exception and records none of it — the body neither logs it, rethrows it, nor even names it, so the failure is discarded as completely as by an empty catch and only the substituted value survives. Log it through whatever this codebase already uses to report problems, narrow the catch to the exception this call can actually raise, or say in a comment on the catch why the failure genuinely cannot matter.
Swallowed exception (caught, then discarded) REDACTED:536— `except` takes every exception and records none of it — the body neither logs it, rethrows it, nor even names it, so the failure is discarded as completely as by an empty catch and only the substituted value survives. Log it through whatever this codebase already uses to report problems, narrow the catch to the exception this call can actually raise, or say in a comment on the catch why the failure genuinely cannot matter.
Swallowed exception (caught, then discarded) hanlp/optimizers/adamw/optimization.py:100— `except` takes every exception and records none of it — the body neither logs it, rethrows it, nor even names it, so the failure is discarded as completely as by an empty catch and only the substituted value survives. Log it through whatever this codebase already uses to report problems, narrow the catch to the exception this call can actually raise, or say in a comment on the catch why the failure genuinely cannot matter.
Swallowed exception (caught, then discarded) hanlp/components/parsers/ud/lemma_edit.py:127— `except` takes every exception and records none of it — the body neither logs it, rethrows it, nor even names it, so the failure is discarded as completely as by an empty catch and only the substituted value survives. Log it through whatever this codebase already uses to report problems, narrow the catch to the exception this call can actually raise, or say in a comment on the catch why the failure genuinely cannot matter.
Swallowed exception (caught, then discarded) REDACTED:21— `except` takes every exception and records none of it — the body neither logs it, rethrows it, nor even names it, so the failure is discarded as completely as by an empty catch and only the substituted value survives. Log it through whatever this codebase already uses to report problems, narrow the catch to the exception this call can actually raise, or say in a comment on the catch why the failure genuinely cannot matter.
Swallowed exception (caught, then discarded) hanlp/components/amr/amrbart/model_interface/tokenization_bart.py:144— `except Exception` takes every exception and records none of it — the body neither logs it, rethrows it, nor even names it, so the failure is discarded as completely as by an empty catch and only the substituted value survives. Log it through whatever this codebase already uses to report problems, narrow the catch to the exception this call can actually raise, or say in a comment on the catch why the failure genuinely cannot matter.
Swallowed exception (caught, then discarded) hanlp/components/amr/amrbart/model_interface/tokenization_bart.py:166— `except Exception` takes every exception and records none of it — the body neither logs it, rethrows it, nor even names it, so the failure is discarded as completely as by an empty catch and only the substituted value survives. Log it through whatever this codebase already uses to report problems, narrow the catch to the exception this call can actually raise, or say in a comment on the catch why the failure genuinely cannot matter.
Swallowed exception (caught, then discarded) REDACTED:45— `except` takes every exception and records none of it — the body neither logs it, rethrows it, nor even names it, so the failure is discarded as completely as by an empty catch and only the substituted value survives. Log it through whatever this codebase already uses to report problems, narrow the catch to the exception this call can actually raise, or say in a comment on the catch why the failure genuinely cannot matter.
TodoComment hanlp/utils/lang/ja/bert_tok.py:78— # TODO: This doesn't work with rust tokenizers — source code is not a task system: move the work to your tracker and leave a reference instead (e.g. `# REF: #123`), so the task is planned where tasks live and the ticket links back to the code.
TodoComment hanlp/transform/transformer_tokenizer.py:73— # TODO: other fields should be properly handled too — source code is not a task system: move the work to your tracker and leave a reference instead (e.g. `# REF: #123`), so the task is planned where tasks live and the ticket links back to the code.
TodoComment hanlp/transform/txt_tf.py:58— # TODO: optimize ngram generation using https://www.tensorflow.org/api_docs/python/tf/strings/ngrams
TodoComment hanlp/layers/transformers/utils.py:43— # TODO: split token type ids in transformer_sliding_window if token type ids are not always 1 — source code is not a task system: move the work to your tracker and leave a reference instead (e.g. `# REF: #123`), so the task is planned where tasks live and the ticket links back to the code.
TodoComment hanlp/layers/crf/crf_tf.py:23— # TODO: Wrap functions in @tf.function once — source code is not a task system: move the work to your tracker and leave a reference instead (e.g. `# REF: #123`), so the task is planned where tasks live and the ticket links back to the code.
TodoComment REDACTED:73— # TODO: this is an ugly patch due to the fact that BART tokenizer splits after ':' — source code is not a task system: move the work to your tracker and leave a reference instead (e.g. `# REF: #123`), so the task is planned where tasks live and the ticket links back to the code.
TodoComment REDACTED:238— # TODO: this is an ugly patch due to the fact that BART tokenizer splits after ':' — source code is not a task system: move the work to your tracker and leave a reference instead (e.g. `# REF: #123`), so the task is planned where tasks live and the ticket links back to the code.
TodoComment REDACTED:104— # TODO: this is an ugly patch due to the fact that BART tokenizer splits after ':' — source code is not a task system: move the work to your tracker and leave a reference instead (e.g. `# REF: #123`), so the task is planned where tasks live and the ticket links back to the code.
Duplicated block (8 lines × 2) hanlp/components/amr/amrbart/bart_amr_parser.py:184— hanlp/components/amr/amrbart/bart_amr_parser.py:184-191 | hanlp/components/amr/seq2seq/seq2seq_amr_parser.py:236-243 — the copies span different directories, so extracting a shared function means choosing where it lives: put it somewhere both call sites can already reach — a location they all depend on today, or a new shared one if there is none — and call it from each site; until then, every change has to be made twice.
Duplicated block (8 lines × 2) REDACTED:475— REDACTED:475-482 | REDACTED:597-604 — before extracting anything, compare `REDACTED` and `REDACTED` as WHOLE FILES: this scan already matched 15 separate duplicated blocks between them, totalling at least 404 lines, which is the signature of one file having been copied from the other rather than of a helper waiting to be extracted. The two sit in different directories, so one cannot simply be deleted in favour of the other while both are reached separately: hoist the shared part into a location both already depend on and have each file call it, and retire whichever file turns out to have no caller of its own left. Extracting one helper per block leaves the fork in place.
Duplicated block (8 lines × 2) hanlp/components/mtl/tasks/dep.py:154— hanlp/components/mtl/tasks/dep.py:154-161 | hanlp/components/parsers/biaffine/biaffine_dep.py:566-573 — the copies span different directories, so extracting a shared function means choosing where it lives: put it somewhere both call sites can already reach — a location they all depend on today, or a new shared one if there is none — and call it from each site; until then, every change has to be made twice. Note first that the copies are not typed on the same thing: the declarations holding them bind `batch` to `Dict[str, Any]` in one and `dict` in another, and the duplicated lines use it. The extracted unit therefore needs a parameter type that fits BOTH — their common supertype where they have one, or a new abstraction over them where they do not — and settling that is the step that comes BEFORE the extraction above. Where the two types are deliberately unrelated, the duplication is the price of that separation and the honest resolution is to record the decision rather than to extract.
Duplicated block (8 lines × 2) hanlp/components/ner/biaffine_ner/biaffine_ner.py:175— hanlp/components/ner/biaffine_ner/biaffine_ner.py:175-182 | hanlp/components/srl/span_rank/span_rank.py:180-187 — the copies span different directories, so extracting a shared function means choosing where it lives: put it somewhere both call sites can already reach — a location they all depend on today, or a new shared one if there is none — and call it from each site; until then, every change has to be made twice.
Duplicated block (8 lines × 2) hanlp/components/parsers/alg_tf.py:58— hanlp/components/parsers/alg_tf.py:58-65 | hanlp/components/parsers/biaffine_tf/alg.py:58-65 — before extracting anything, compare `hanlp/components/parsers/alg_tf.py` and `hanlp/components/parsers/biaffine_tf/alg.py` as WHOLE FILES: this scan already matched 9 separate duplicated blocks between them, totalling at least 185 lines, which is the signature of one file having been copied from the other rather than of a helper waiting to be extracted. The two sit in different directories, so one cannot simply be deleted in favour of the other while both are reached separately: hoist the shared part into a location both already depend on and have each file call it, and retire whichever file turns out to have no caller of its own left. Extracting one helper per block leaves the fork in place.
Duplicated block (8 lines × 2) hanlp/components/parsers/constituency/crf_constituency_parser.py:93— hanlp/components/parsers/constituency/crf_constituency_parser.py:93-100 | hanlp/components/parsers/ud/ud_parser.py:260-267 — the copies span different directories, so extracting a shared function means choosing where it lives: put it somewhere both call sites can already reach — a location they all depend on today, or a new shared one if there is none — and call it from each site; until then, every change has to be made twice. ★ These copies have DRIFTED, and that is worth reading before extracting anything: just after the matched lines, `hanlp/components/parsers/constituency/crf_constituency_parser.py:101` calls `compute_loss` and `hanlp/components/parsers/ud/ud_parser.py:268` does not — after which the two agree again for 4 more lines. One of those two behaviours is the intended one and the other is what a copy-paste left behind, so decide which BEFORE unifying them: extracting the shared part will silently settle it, and if the copy that skips the call is the wrong one, that bug is already live.
Duplicated block (8 lines × 2) hanlp/components/parsers/ud/ud_parser.py:59— hanlp/components/parsers/ud/ud_parser.py:59-66 | hanlp/components/srl/span_bio/span_bio.py:258-265 — the copies span different directories, so extracting a shared function means choosing where it lives: put it somewhere both call sites can already reach — a location they all depend on today, or a new shared one if there is none — and call it from each site; until then, every change has to be made twice.
Duplicated block (8 lines × 2) hanlp/components/amr/amrbart/bart_amr_generation.py:67— hanlp/components/amr/amrbart/bart_amr_generation.py:67-74 | hanlp/components/amr/amrbart/bart_amr_parser.py:68-75 — before extracting anything, compare `hanlp/components/amr/amrbart/bart_amr_generation.py` and `hanlp/components/amr/amrbart/bart_amr_parser.py` as WHOLE FILES: this scan already matched 4 separate duplicated blocks between them, totalling at least 35 lines, which is the signature of one file having been copied from the other rather than of a helper waiting to be extracted. If that is what happened, the fix is to keep one copy and have the other call it (or delete it), which resolves this row and its siblings together — extracting one helper per block leaves the fork in place. The `return` at the foot of the matched lines is the enclosing body's own terminal exit, not an early one: it moves with them unchanged, and each site calls the extracted unit from the position that `return` occupied — no decision has to be handed back and re-acted on.
Swallowed exception (empty catch) plugins/hanlp_restful_java/src/main/java/com/hankcs/hanlp/restful/HanLPClient.java:611— An empty catch block silently discards the error — failures vanish with no log and no rethrow. Log it, handle it, or don't catch it.
Swallowed exception (empty catch) hanlp/utils/component_util.py:186— An empty catch block silently discards the error — failures vanish with no log and no rethrow. Log it, handle it, or don't catch it.
Swallowed exception (empty catch) hanlp/components/amr/seq2seq/seq2seq_amr_parser.py:250— An empty catch block silently discards the error — failures vanish with no log and no rethrow. Log it, handle it, or don't catch it.
Swallowed exception (empty catch) REDACTED:580— An empty catch block silently discards the error — failures vanish with no log and no rethrow. Log it, handle it, or don't catch it.
Swallowed exception (empty catch) hanlp/components/amr/amrbart/bart_amr_parser.py:198— An empty catch block silently discards the error — failures vanish with no log and no rethrow. Log it, handle it, or don't catch it.
Swallowed exception (empty catch) REDACTED:458— An empty catch block silently discards the error — failures vanish with no log and no rethrow. Log it, handle it, or don't catch it.
Duplicated block (11 lines × 2) hanlp/common/vocab.py:55— hanlp/common/vocab.py:55-65 | hanlp/common/vocab_tf.py:40-50 — the copies sit in sibling files of one directory, so a shared home is within easy reach: extract the block into a single shared function the call sites can all reach — a file they already depend on, or a new one alongside them — and call it from both call sites, so a change lands once. The matched lines also transfer control out of the body holding them, which cannot survive a move into a called unit unchanged: have the extracted unit return that decision and let each site act on it. ★ These copies have DRIFTED, and that is worth reading before extracting anything: just before the matched lines, `hanlp/common/vocab.py:50` calls `token` and `hanlp/common/vocab_tf.py:39` does not — after which the two agree again for 2 more lines. One of those two behaviours is the intended one and the other is what a copy-paste left behind, so decide which BEFORE unifying them: extracting the shared part will silently settle it, and if the copy that skips the call is the wrong one, that bug is already live.
Duplicated block (11 lines × 2) hanlp/components/amr/seq2seq/dataset/tokenization_bart.py:283— hanlp/components/amr/seq2seq/dataset/tokenization_bart.py:283-293 | hanlp/components/amr/seq2seq/dataset/tokenization_t5.py:276-286 — before extracting anything, compare `hanlp/components/amr/seq2seq/dataset/tokenization_bart.py` and `hanlp/components/amr/seq2seq/dataset/tokenization_t5.py` as WHOLE FILES: this scan already matched 21 separate duplicated blocks between them, totalling at least 476 lines, which is the signature of one file having been copied from the other rather than of a helper waiting to be extracted. If that is what happened, the fix is to keep one copy and have the other call it (or delete it), which resolves this row and its siblings together — extracting one helper per block leaves the fork in place.
Duplicated block (11 lines × 2) hanlp/components/ner/biaffine_ner/biaffine_ner.py:133— hanlp/components/ner/biaffine_ner/biaffine_ner.py:133-143 | hanlp/components/sts/transformer_sts.py:112-122 — the copies span different directories, so extracting a shared function means choosing where it lives: put it somewhere both call sites can already reach — a location they all depend on today, or a new shared one if there is none — and call it from each site; until then, every change has to be made twice. ★ These copies have DRIFTED, and that is worth reading before extracting anything: just after the matched lines, `hanlp/components/sts/transformer_sts.py:123` calls `zero_grad` and `hanlp/components/ner/biaffine_ner/biaffine_ner.py:144` does not — after which the two agree again for 4 more lines. One of those two behaviours is the intended one and the other is what a copy-paste left behind, so decide which BEFORE unifying them: extracting the shared part will silently settle it, and if the copy that skips the call is the wrong one, that bug is already live.
Duplicated block (11 lines × 2) hanlp/components/srl/span_rank/highway_variational_lstm.py:24— hanlp/components/srl/span_rank/highway_variational_lstm.py:24-34 | hanlp/components/srl/span_rank/highway_variational_lstm.py:126-136 — both copies are in the same file, so extract the block into one function there and call it from each site — the copies drift apart the first time only one of them is edited.
Duplicated block (11 lines × 2) hanlp/datasets/coref/loaders/conll12coref.py:33— hanlp/datasets/coref/loaders/conll12coref.py:33-43 | hanlp/datasets/srl/loaders/ontonotes_loader.py:216-226 — the copies span different directories, so extracting a shared function means choosing where it lives: put it somewhere both call sites can already reach — a location they all depend on today, or a new shared one if there is none — and call it from each site; until then, every change has to be made twice.
TooManyMethods: KerasComponent hanlp/common/keras_component.py:33— TooManyMethods — 37 methods. The bar is 30 methods; this is 7 over it, 1.23× the bar. To reduce it, group the members that share the same data into a smaller type of their own and delegate to it, so no single type carries every responsibility.
TooManyMethods: MultiTaskLearning hanlp/components/mtl/multi_task_learning.py:93— TooManyMethods — 34 methods. The bar is 30 methods; this is 4 over it, 1.13× the bar. To reduce it, group the members that share the same data into a smaller type of their own and delegate to it, so no single type carries every responsibility.
TooManyMethods: BiaffineDependencyParser hanlp/components/parsers/biaffine/biaffine_dep.py:33— TooManyMethods — 34 methods. The bar is 30 methods; this is 4 over it, 1.13× the bar. To reduce it, group the members that share the same data into a smaller type of their own and delegate to it, so no single type carries every responsibility.
TooManyMethods: HanLPClient plugins/hanlp_restful_java/src/main/java/com/hankcs/hanlp/restful/HanLPClient.java:32— TooManyMethods — 32 methods. The bar is 30 methods; this is 2 over it, 1.07× the bar. To reduce it, group the members that share the same data into a smaller type of their own and delegate to it, so no single type carries every responsibility.
Duplicated block (13 lines × 2) hanlp/components/amr/amrbart/bart_amr_parser.py:193— hanlp/components/amr/amrbart/bart_amr_parser.py:193-205 | hanlp/components/amr/seq2seq/seq2seq_amr_parser.py:245-257 — the copies span different directories, so extracting a shared function means choosing where it lives: put it somewhere both call sites can already reach — a location they all depend on today, or a new shared one if there is none — and call it from each site; until then, every change has to be made twice. The `return` at the foot of the matched lines is the enclosing body's own terminal exit, not an early one: it moves with them unchanged, and each site calls the extracted unit from the position that `return` occupied — no decision has to be handed back and re-acted on.
Duplicated block (13 lines × 2) REDACTED:40— REDACTED:40-52 | REDACTED:16-28 — before extracting anything, compare `REDACTED` and `REDACTED` as WHOLE FILES: this scan already matched 15 separate duplicated blocks between them, totalling at least 404 lines, which is the signature of one file having been copied from the other rather than of a helper waiting to be extracted. The two sit in different directories, so one cannot simply be deleted in favour of the other while both are reached separately: hoist the shared part into a location both already depend on and have each file call it, and retire whichever file turns out to have no caller of its own left. Extracting one helper per block leaves the fork in place. The matched lines also transfer control out of the body holding them, which cannot survive a move into a called unit unchanged: have the extracted unit return that decision and let each site act on it.
Duplicated block (13 lines × 2) hanlp/components/amr/amrbart/model_interface/modeling_bart.py:1357— hanlp/components/amr/amrbart/model_interface/modeling_bart.py:1357-1369 | hanlp/components/amr/amrbart/model_interface/modeling_bart.py:1840-1854 — both copies are in the same file, so extract the block into one function there and call it from each site — the copies drift apart the first time only one of them is edited. Read the line range as the matched WINDOW rather than a finished unit: at `hanlp/components/amr/amrbart/model_interface/modeling_bart.py:1357` it begins part-way through the construct above it, so those exact lines cannot be lifted as they stand — widen the region to the smallest complete statement or declaration that contains it, and extract that. Note that the copies do not run to the end of the range shown: their LAST lines are different code, not the same code under different names — the matched region ends inside that line. Extract the lines above it, and read the last line of each site separately.
Duplicated block (13 lines × 2) hanlp/components/ner/biaffine_ner/biaffine_ner.py:52— hanlp/components/ner/biaffine_ner/biaffine_ner.py:52-64 | hanlp/components/srl/span_rank/span_rank.py:54-66 — the copies span different directories, so extracting a shared function means choosing where it lives: put it somewhere both call sites can already reach — a location they all depend on today, or a new shared one if there is none — and call it from each site; until then, every change has to be made twice. Read the line range as the matched WINDOW rather than a finished unit: at `hanlp/components/ner/biaffine_ner/biaffine_ner.py:52` it begins part-way through the construct above it, so those exact lines cannot be lifted as they stand — widen the region to the smallest complete statement or declaration that contains it, and extract that. The `return` at the foot of the matched lines is the enclosing body's own terminal exit, not an early one: it moves with them unchanged, and each site calls the extracted unit from the position that `return` occupied — no decision has to be handed back and re-acted on. ★ These copies have DRIFTED, and that is worth reading before extracting anything: just after the matched lines, `hanlp/components/ner/biaffine_ner/biaffine_ner.py:67` calls `use_transformer` and `hanlp/components/srl/span_rank/span_rank.py:68` does not — after which the two agree again for 2 more lines. One of those two behaviours is the intended one and the other is what a copy-paste left behind, so decide which BEFORE unifying them: extracting the shared part will silently settle it, and if the copy that skips the call is the wrong one, that bug is already live.
Duplicated block (10 lines × 2) hanlp/common/vocab.py:245— hanlp/common/vocab.py:245-254 | hanlp/common/vocab.py:486-495 — both copies are in the same file, so extract the block into one function there and call it from each site — the copies drift apart the first time only one of them is edited. The matched lines also transfer control out of the body holding them, which cannot survive a move into a called unit unchanged: have the extracted unit return that decision and let each site act on it.
Duplicated block (10 lines × 2) hanlp/components/mtl/tasks/lem.py:86— hanlp/components/mtl/tasks/lem.py:86-95 | hanlp/components/mtl/tasks/pos.py:123-132 — the copies sit in sibling files of one directory, so a shared home is within easy reach: extract the block into a single shared function the call sites can all reach — a file they already depend on, or a new one alongside them — and call it from both call sites, so a change lands once. Read the line range as the matched WINDOW rather than a finished unit: at `hanlp/components/mtl/tasks/lem.py:86` it begins part-way through the construct above it, so those exact lines cannot be lifted as they stand — widen the region to the smallest complete statement or declaration that contains it, and extract that. The matched lines also transfer control out of the body holding them, which cannot survive a move into a called unit unchanged: have the extracted unit return that decision and let each site act on it.
Duplicated block (10 lines × 2) hanlp/components/srl/span_rank/highway_variational_lstm.py:67— hanlp/components/srl/span_rank/highway_variational_lstm.py:67-76 | hanlp/components/srl/span_rank/highway_variational_lstm.py:195-204 — both copies are in the same file, so extract the block into one function there and call it from each site — the copies drift apart the first time only one of them is edited. The `return` at the foot of the matched lines is the enclosing body's own terminal exit, not an early one: it moves with them unchanged, and each site calls the extracted unit from the position that `return` occupied — no decision has to be handed back and re-acted on.
Duplicated block (10 lines × 2) hanlp/transform/conll_tf.py:257— hanlp/transform/conll_tf.py:257-266 | hanlp/transform/conll_tf.py:679-688 — both copies are in the same file, so extract the block into one function there and call it from each site — the copies drift apart the first time only one of them is edited.
Duplicated block (17 lines × 3) REDACTED:279— REDACTED:279-295 | hanlp/components/amr/seq2seq/dataset/linearization.py:145-161 | REDACTED:407-423 — before extracting anything, compare `REDACTED` and `REDACTED` as WHOLE FILES: this scan already matched 15 separate duplicated blocks between them, totalling at least 404 lines, which is the signature of one file having been copied from the other rather than of a helper waiting to be extracted. The two sit in different directories, so one cannot simply be deleted in favour of the other while both are reached separately: hoist the shared part into a location both already depend on and have each file call it, and retire whichever file turns out to have no caller of its own left. Extracting one helper per block leaves the fork in place.
Duplicated block (17 lines × 3) hanlp/components/parsers/alg_tf.py:221— hanlp/components/parsers/alg_tf.py:221-237 | hanlp/components/parsers/biaffine_tf/alg.py:221-237 | hanlp/components/parsers/parse_alg.py:139-155 — before extracting anything, compare `hanlp/components/parsers/alg_tf.py` and `hanlp/components/parsers/biaffine_tf/alg.py` as WHOLE FILES: this scan already matched 9 separate duplicated blocks between them, totalling at least 185 lines, which is the signature of one file having been copied from the other rather than of a helper waiting to be extracted. The two sit in different directories, so one cannot simply be deleted in favour of the other while both are reached separately: hoist the shared part into a location both already depend on and have each file call it, and retire whichever file turns out to have no caller of its own left. Extracting one helper per block leaves the fork in place.
Duplicated block (17 lines × 3) hanlp/components/parsers/alg_tf.py:240— hanlp/components/parsers/alg_tf.py:240-256 | hanlp/components/parsers/biaffine_tf/alg.py:240-256 | hanlp/components/parsers/parse_alg.py:158-174 — before extracting anything, compare `hanlp/components/parsers/alg_tf.py` and `hanlp/components/parsers/biaffine_tf/alg.py` as WHOLE FILES: this scan already matched 9 separate duplicated blocks between them, totalling at least 185 lines, which is the signature of one file having been copied from the other rather than of a helper waiting to be extracted. The two sit in different directories, so one cannot simply be deleted in favour of the other while both are reached separately: hoist the shared part into a location both already depend on and have each file call it, and retire whichever file turns out to have no caller of its own left. Extracting one helper per block leaves the fork in place. The matched lines also transfer control out of the body holding them, which cannot survive a move into a called unit unchanged: have the extracted unit return that decision and let each site act on it.
Duplicated block (12 lines × 2) hanlp/components/amr/amrbart/model_interface/modeling_bart.py:785— hanlp/components/amr/amrbart/model_interface/modeling_bart.py:785-796 | hanlp/components/amr/amrbart/model_interface/modeling_bart.py:1010-1021 — both copies are in the same file, so extract the block into one function there and call it from each site — the copies drift apart the first time only one of them is edited.
Duplicated block (12 lines × 2) hanlp/components/classifiers/transformer_classifier_hf.py:25— hanlp/components/classifiers/transformer_classifier_hf.py:25-36 | hanlp/components/classifiers/transformer_regression_hf.py:25-36 — before extracting anything, compare `hanlp/components/classifiers/transformer_classifier_hf.py` and `hanlp/components/classifiers/transformer_regression_hf.py` as WHOLE FILES: 91% of the shorter file's lines also appear in the other, so this reads as one file having been copied from the other rather than as a helper waiting to be extracted. The 1 duplicated block(s) this scan matched between them are fragments of that copy, not the extent of it — treat the file pair as the unit. If that is what happened, the fix is to keep one copy and have the other call it (or delete it), which resolves this row and its siblings together — extracting one helper per block leaves the fork in place. The `return` at the foot of the matched lines is the enclosing body's own terminal exit, not an early one: it moves with them unchanged, and each site calls the extracted unit from the position that `return` occupied — no decision has to be handed back and re-acted on.
Duplicated block (12 lines × 2) hanlp/components/ner/biaffine_ner/biaffine_ner.py:101— hanlp/components/ner/biaffine_ner/biaffine_ner.py:101-112 | hanlp/components/srl/span_rank/span_rank.py:99-110 — the copies span different directories, so extracting a shared function means choosing where it lives: put it somewhere both call sites can already reach — a location they all depend on today, or a new shared one if there is none — and call it from each site; until then, every change has to be made twice. Read the line range as the matched WINDOW rather than a finished unit: at `hanlp/components/ner/biaffine_ner/biaffine_ner.py:101` it begins part-way through the construct above it, so those exact lines cannot be lifted as they stand — widen the region to the smallest complete statement or declaration that contains it, and extract that.
Low cohesion: Transform (LCOM4 4) hanlp/common/transform_tf.py:16— Transform's methods fall into 4 groups that share no field and call none of each other, against a bar of more than 3 for this run (LCOM4, configurable — your repository's bar is the one quoted here). Each group is a set of methods reachable from one another through shared fields or direct calls, so 4 groups means the type has that many internally-connected clusters with nothing tying them together. Types whose shape makes a high count expected — and which would otherwise dominate this list — are excluded before this row is raised, so this is a genuine split candidate rather than a metric reading. It is still a shape, not a defect: confirm the groups match responsibilities you can name before splitting.
Low cohesion: Word2VecEmbeddingComponent (LCOM4 4) hanlp/layers/embeddings/word2vec.py:175— Word2VecEmbeddingComponent's methods fall into 4 groups that share no field and call none of each other, against a bar of more than 3 for this run (LCOM4, configurable — your repository's bar is the one quoted here). Each group is a set of methods reachable from one another through shared fields or direct calls, so 4 groups means the type has that many internally-connected clusters with nothing tying them together. Types whose shape makes a high count expected — and which would otherwise dominate this list — are excluded before this row is raised, so this is a genuine split candidate rather than a metric reading. It is still a shape, not a defect: confirm the groups match responsibilities you can name before splitting.
Low cohesion: CoNLLTransform (LCOM4 4) hanlp/transform/conll_tf.py:25— CoNLLTransform's methods fall into 4 groups that share no field and call none of each other, against a bar of more than 3 for this run (LCOM4, configurable — your repository's bar is the one quoted here). Each group is a set of methods reachable from one another through shared fields or direct calls, so 4 groups means the type has that many internally-connected clusters with nothing tying them together. Types whose shape makes a high count expected — and which would otherwise dominate this list — are excluded before this row is raised, so this is a genuine split candidate rather than a metric reading. It is still a shape, not a defect: confirm the groups match responsibilities you can name before splitting.
postprocessing._reconstruct_graph_from_nodes (cyclomatic 40) REDACTED:314— postprocessing._reconstruct_graph_from_nodes has cyclomatic complexity 40 (threshold 15). To reduce it, separate the cases: extract each independent branch into its own named function, and where the body has guards that only reject input, fold those into early returns at the top.
postprocessing._reconstruct_graph_from_nodes (cyclomatic 40) REDACTED:442— postprocessing._reconstruct_graph_from_nodes has cyclomatic complexity 40 (threshold 15). To reduce it, separate the cases: extract each independent branch into its own named function, and where the body has guards that only reject input, fold those into early returns at the top.
postprocessing._reconstruct_graph_from_nodes (cognitive 64) REDACTED:314— postprocessing._reconstruct_graph_from_nodes has cognitive complexity 64 (threshold 15). Drivers by points: if/else 17 (50 pts), boolean chains 11, loops 2 (3 pts) (nesting depth added 34). To reduce it, split the body into named stages: move each independent step or branch into its own named function so the body reads as a short sequence of named calls rather than one long body.
postprocessing._reconstruct_graph_from_nodes (cognitive 64) REDACTED:442— postprocessing._reconstruct_graph_from_nodes has cognitive complexity 64 (threshold 15). Drivers by points: if/else 17 (50 pts), boolean chains 11, loops 2 (3 pts) (nesting depth added 34). To reduce it, split the body into named stages: move each independent step or branch into its own named function so the body reads as a short sequence of named calls rather than one long body.
Change coupling: REDACTED ↔ REDACTED REDACTED:1— `REDACTED` and `plugins/hanlp_restful/REDACTED` change together 100% of the time (16 of the 16 commits that touched whichever of the two files changed less often, counting a file under its earlier names as well — a repo-wide or module-wide sweep is evidence about the sweep rather than about any pair inside it and is left out of BOTH sides of this ratio, while a dependency bump, a formatter/rename sweep, or a commit whose edit to one of the two files was a tool directive such as //go:generate or whitespace only is left out of the shared count ONLY, so the two sides are not taken over identical commit sets) with no explicit dependency — the edge is real but nothing declares it. Read the pair before acting: if one registers itself into the other through a hook or an initialiser, the missing dependency is DELIBERATE — the registration is the link, and it is meant not to be an import — and the thing to add is a comment on each side naming the other, not a merge; if they simply belong together, co-locate them; if neither holds, the coupling is hidden and worth breaking. You can check this without leaving the row: of the 16 shared commits counted here, the most recent 3 are `007f29f9` Support coarse tokenization of Japanese; `bf77f509` Release `sentiment_analysis` APIs; `9d0f8804` Release `abstractive_summarization` APIs — run `git show` on any of them.
Change coupling: document.py ↔ REDACTED plugins/hanlp_common/hanlp_common/document.py:1— `plugins/hanlp_common/hanlp_common/document.py` and `plugins/hanlp_common/REDACTED` change together 71% of the time (15 of the 21 commits that touched whichever of the two files changed less often, counting a file under its earlier names as well — a repo-wide or module-wide sweep is evidence about the sweep rather than about any pair inside it and is left out of BOTH sides of this ratio, while a dependency bump, a formatter/rename sweep, or a commit whose edit to one of the two files was a tool directive such as //go:generate or whitespace only is left out of the shared count ONLY, so the two sides are not taken over identical commit sets) with no explicit dependency — the edge is real but nothing declares it. Read the pair before acting: if one registers itself into the other through a hook or an initialiser, the missing dependency is DELIBERATE — the registration is the link, and it is meant not to be an import — and the thing to add is a comment on each side naming the other, not a merge; if they simply belong together, co-locate them; if neither holds, the coupling is hidden and worth breaking. You can check this without leaving the row: of the 15 shared commits counted here, the most recent 3 are `7aeb0fc5` Support XPOS for CoNLL-U; `c01f4f64` Fix `dep` key of `to_pretty`; `7f7d7b62` Fix printing dummy root constituent — run `git show` on any of them.
D4 · Code Duplication· Members sharing a duplicated core (6 members, 50+ identical tokens) · ×2
Members sharing a duplicated core (6 members, 50+ identical tokens) hanlp/components/amr/amrbart/model_interface/tokenization_bart.py:67— hanlp/components/amr/amrbart/model_interface/tokenization_bart.py:67-81 | hanlp/components/amr/amrbart/model_interface/tokenization_bart.py:84-95 | hanlp/components/amr/seq2seq/dataset/tokenization_bart.py:88-102 | hanlp/components/amr/seq2seq/dataset/tokenization_bart.py:107-118 | hanlp/components/amr/seq2seq/dataset/tokenization_t5.py:80-94 | hanlp/components/amr/seq2seq/dataset/tokenization_t5.py:99-107 — These 6 members share a duplicated core: a run of at least 50 identical tokens appears in every one of them. That run is NOT broken out as duplicated-block rows below — it is what admitted this row, and the blocks below cover only the part of it that clears the block floor, so they understate the correspondence. Read the members as one construct written 6 times. The repair is at the members' grain — factor the shared implementation out once and have all of them call it with their differences as parameters or as an injected step, or, where the difference is systematic, generate them from one template. Extracting the individual blocks below is not the same fix: it leaves every body in place and the next edit still has to be made 6 times.
Members sharing a duplicated core (6 members, 50+ identical tokens) hanlp/components/amr/seq2seq/seq2seq_amr_parser.py:141— hanlp/components/amr/seq2seq/seq2seq_amr_parser.py:141-175 | hanlp/components/parsers/constituency/crf_constituency_parser.py:49-78 | hanlp/components/parsers/ud/ud_parser.py:216-245 | hanlp/components/srl/span_bio/span_bio.py:91-120 | hanlp/components/sts/transformer_sts.py:82-99 | hanlp/components/taggers/tagger.py:71-99 — These 6 members share a duplicated core: a run of at least 50 identical tokens appears in every one of them. That run is NOT broken out as duplicated-block rows below — it is what admitted this row, and the blocks below cover only the part of it that clears the block floor, so they understate the correspondence. Read the members as one construct written 6 times. The repair is at the members' grain — factor the shared implementation out once and have all of them call it with their differences as parameters or as an injected step, or, where the difference is systematic, generate them from one template. Extracting the individual blocks below is not the same fix: it leaves every body in place and the next edit still has to be made 6 times.
D4 · Code Duplication· Members sharing a duplicated core (4 members, 50+ identical tokens) · ×2
Members sharing a duplicated core (4 members, 50+ identical tokens) hanlp/components/amr/amrbart/model_interface/modeling_bart.py:1334— hanlp/components/amr/amrbart/model_interface/modeling_bart.py:1334-1391 | hanlp/components/amr/amrbart/model_interface/modeling_bart.py:1482-1559 | hanlp/components/amr/amrbart/model_interface/modeling_bart.py:1609-1682 | hanlp/components/amr/amrbart/model_interface/modeling_bart.py:1747-1872 — These 4 members share a duplicated core: a run of at least 50 identical tokens appears in every one of them. That run is NOT broken out as duplicated-block rows below — it is what admitted this row, and the blocks below cover only the part of it that clears the block floor, so they understate the correspondence. Read the members as one construct written 4 times. The repair is at the members' grain — factor the shared implementation out once and have all of them call it with their differences as parameters or as an injected step, or, where the difference is systematic, generate them from one template. Extracting the individual blocks below is not the same fix: it leaves every body in place and the next edit still has to be made 4 times.
Members sharing a duplicated core (4 members, 50+ identical tokens) hanlp/components/amr/seq2seq/dataset/tokenization_bart.py:224— hanlp/components/amr/seq2seq/dataset/tokenization_bart.py:224-254 | hanlp/components/amr/seq2seq/dataset/tokenization_bart.py:636-671 | hanlp/components/amr/seq2seq/dataset/tokenization_t5.py:214-245 | hanlp/components/amr/seq2seq/dataset/tokenization_t5.py:631-669 — These 4 members share a duplicated core: a run of at least 50 identical tokens appears in every one of them. That run is NOT broken out as duplicated-block rows below — it is what admitted this row, and the blocks below cover only the part of it that clears the block floor, so they understate the correspondence. Read the members as one construct written 4 times. The repair is at the members' grain — factor the shared implementation out once and have all of them call it with their differences as parameters or as an injected step, or, where the difference is systematic, generate them from one template. Extracting the individual blocks below is not the same fix: it leaves every body in place and the next edit still has to be made 4 times.
Duplicated block (23 lines × 2) hanlp/components/amr/seq2seq/dataset/tokenization_bart.py:154— hanlp/components/amr/seq2seq/dataset/tokenization_bart.py:154-176 | hanlp/components/amr/seq2seq/dataset/tokenization_t5.py:144-166 — before extracting anything, compare `hanlp/components/amr/seq2seq/dataset/tokenization_bart.py` and `hanlp/components/amr/seq2seq/dataset/tokenization_t5.py` as WHOLE FILES: this scan already matched 21 separate duplicated blocks between them, totalling at least 476 lines, which is the signature of one file having been copied from the other rather than of a helper waiting to be extracted. If that is what happened, the fix is to keep one copy and have the other call it (or delete it), which resolves this row and its siblings together — extracting one helper per block leaves the fork in place. Read the line range as the matched WINDOW rather than a finished unit: at `hanlp/components/amr/seq2seq/dataset/tokenization_bart.py:154` it begins part-way through the construct above it, so those exact lines cannot be lifted as they stand — widen the region to the smallest complete statement or declaration that contains it, and extract that.
Duplicated block (23 lines × 2) hanlp/components/srl/span_rank/srl_eval_utils.py:105— hanlp/components/srl/span_rank/srl_eval_utils.py:105-127 | hanlp/components/srl/span_rank/srl_eval_utils.py:131-153 — both copies are in the same file, so extract the block into one function there and call it from each site — the copies drift apart the first time only one of them is edited. The `return` at the foot of the matched lines is the enclosing body's own terminal exit, not an early one: it moves with them unchanged, and each site calls the extracted unit from the position that `return` occupied — no decision has to be handed back and re-acted on.
Duplicated block (17 lines × 2) hanlp/components/parsers/constituency/crf_constituency_parser.py:57— hanlp/components/parsers/constituency/crf_constituency_parser.py:57-73 | hanlp/components/taggers/tagger.py:77-93 — the copies span different directories, so extracting a shared function means choosing where it lives: put it somewhere both call sites can already reach — a location they all depend on today, or a new shared one if there is none — and call it from each site; until then, every change has to be made twice. Read the line range as the matched WINDOW rather than a finished unit: at `hanlp/components/parsers/constituency/crf_constituency_parser.py:57` it begins part-way through the construct above it, so those exact lines cannot be lifted as they stand — widen the region to the smallest complete statement or declaration that contains it, and extract that. The matched lines also transfer control out of the body holding them, which cannot survive a move into a called unit unchanged: have the extracted unit return that decision and let each site act on it.
Duplicated block (17 lines × 2) hanlp/transform/conll_tf.py:456— hanlp/transform/conll_tf.py:456-472 | hanlp/transform/conll_tf.py:556-572 — both copies are in the same file, so extract the block into one function there and call it from each site — the copies drift apart the first time only one of them is edited. ★ These copies have DRIFTED, and that is worth reading before extracting anything: just before the matched lines, `hanlp/transform/conll_tf.py:555` calls `adjust_tokens_for_transformers` and `hanlp/transform/conll_tf.py:455` does not — after which the two agree again for 2 more lines. One of those two behaviours is the intended one and the other is what a copy-paste left behind, so decide which BEFORE unifying them: extracting the shared part will silently settle it, and if the copy that skips the call is the wrong one, that bug is already live.
Duplicated block (16 lines × 2) hanlp/components/amr/amrbart/bart_amr_generation.py:88— hanlp/components/amr/amrbart/bart_amr_generation.py:88-103 | hanlp/components/amr/amrbart/bart_amr_parser.py:89-104 — before extracting anything, compare `hanlp/components/amr/amrbart/bart_amr_generation.py` and `hanlp/components/amr/amrbart/bart_amr_parser.py` as WHOLE FILES: this scan already matched 4 separate duplicated blocks between them, totalling at least 35 lines, which is the signature of one file having been copied from the other rather than of a helper waiting to be extracted. If that is what happened, the fix is to keep one copy and have the other call it (or delete it), which resolves this row and its siblings together — extracting one helper per block leaves the fork in place. Read the line range as the matched WINDOW rather than a finished unit: at `hanlp/components/amr/amrbart/bart_amr_generation.py:88` it begins part-way through the construct above it, so those exact lines cannot be lifted as they stand — widen the region to the smallest complete statement or declaration that contains it, and extract that. The `return` at the foot of the matched lines is the enclosing body's own terminal exit, not an early one: it moves with them unchanged, and each site calls the extracted unit from the position that `return` occupied — no decision has to be handed back and re-acted on.
Duplicated block (16 lines × 2) hanlp/components/amr/amrbart/model_interface/modeling_bart.py:1351— hanlp/components/amr/amrbart/model_interface/modeling_bart.py:1351-1366 | hanlp/components/amr/amrbart/model_interface/modeling_bart.py:1494-1509 — both copies are in the same file, so extract the block into one function there and call it from each site — the copies drift apart the first time only one of them is edited. Read the line range as the matched WINDOW rather than a finished unit: at `hanlp/components/amr/amrbart/model_interface/modeling_bart.py:1351` it begins part-way through the construct above it, so those exact lines cannot be lifted as they stand — widen the region to the smallest complete statement or declaration that contains it, and extract that.
Duplicated block (15 lines × 2) hanlp/components/amr/seq2seq/dataset/tokenization_bart.py:52— hanlp/components/amr/seq2seq/dataset/tokenization_bart.py:52-66 | hanlp/components/amr/seq2seq/dataset/tokenization_t5.py:54-68 — before extracting anything, compare `hanlp/components/amr/seq2seq/dataset/tokenization_bart.py` and `hanlp/components/amr/seq2seq/dataset/tokenization_t5.py` as WHOLE FILES: this scan already matched 21 separate duplicated blocks between them, totalling at least 476 lines, which is the signature of one file having been copied from the other rather than of a helper waiting to be extracted. If that is what happened, the fix is to keep one copy and have the other call it (or delete it), which resolves this row and its siblings together — extracting one helper per block leaves the fork in place.
Duplicated block (15 lines × 2) hanlp/components/amr/seq2seq/dataset/tokenization_bart.py:137— hanlp/components/amr/seq2seq/dataset/tokenization_bart.py:137-151 | hanlp/components/amr/seq2seq/dataset/tokenization_t5.py:127-141 — before extracting anything, compare `hanlp/components/amr/seq2seq/dataset/tokenization_bart.py` and `hanlp/components/amr/seq2seq/dataset/tokenization_t5.py` as WHOLE FILES: this scan already matched 21 separate duplicated blocks between them, totalling at least 476 lines, which is the signature of one file having been copied from the other rather than of a helper waiting to be extracted. If that is what happened, the fix is to keep one copy and have the other call it (or delete it), which resolves this row and its siblings together — extracting one helper per block leaves the fork in place.
Duplicated block (11 lines × 3) REDACTED:88— REDACTED:88-98 | REDACTED:57-67 | REDACTED:222-232 — before extracting anything, compare `REDACTED` and `REDACTED` as WHOLE FILES: this scan already matched 15 separate duplicated blocks between them, totalling at least 404 lines, which is the signature of one file having been copied from the other rather than of a helper waiting to be extracted. The two sit in different directories, so one cannot simply be deleted in favour of the other while both are reached separately: hoist the shared part into a location both already depend on and have each file call it, and retire whichever file turns out to have no caller of its own left. Extracting one helper per block leaves the fork in place.
Duplicated block (11 lines × 3) hanlp/datasets/parsing/ud/ud210.py:22— hanlp/datasets/parsing/ud/ud210.py:22-32 | hanlp/datasets/parsing/ud/ud23.py:15-25 | hanlp/datasets/parsing/ud/ud27.py:22-32 — the copies sit in sibling files of one directory, so a shared home is within easy reach: extract the block into a single shared function the call sites can all reach — a file they already depend on, or a new one alongside them — and call it from all 3 call sites, so a change lands once. Read the line range as the matched WINDOW rather than a finished unit: at `hanlp/datasets/parsing/ud/ud210.py:22` it does not close everything it opens, so those exact lines cannot be lifted as they stand — widen the region to the smallest complete statement or declaration that contains it, and extract that.
Duplicated block (6 lines × 3) hanlp/components/amr/amrbart/model_interface/tokenization_bart.py:102— hanlp/components/amr/amrbart/model_interface/tokenization_bart.py:102-107 | hanlp/components/amr/seq2seq/dataset/tokenization_bart.py:135-140 | hanlp/components/amr/seq2seq/dataset/tokenization_t5.py:125-130 — before extracting anything, compare `hanlp/components/amr/amrbart/model_interface/tokenization_bart.py` and `hanlp/components/amr/seq2seq/dataset/tokenization_bart.py` as WHOLE FILES: this scan already matched 10 separate duplicated blocks between them, totalling at least 329 lines, which is the signature of one file having been copied from the other rather than of a helper waiting to be extracted. The two sit in different directories, so one cannot simply be deleted in favour of the other while both are reached separately: hoist the shared part into a location both already depend on and have each file call it, and retire whichever file turns out to have no caller of its own left. Extracting one helper per block leaves the fork in place.
Duplicated block (6 lines × 3) hanlp/components/mtl/tasks/dep.py:156— hanlp/components/mtl/tasks/dep.py:156-161 | hanlp/components/mtl/tasks/sdp.py:156-161 | hanlp/components/parsers/biaffine/biaffine_dep.py:568-573 — the copies span different directories, so extracting a shared function means choosing where it lives: put it somewhere all 3 call sites can already reach — a location they all depend on today, or a new shared one if there is none — and call it from each site; until then, every change has to be made 3 times. Note first that the copies are not typed on the same thing: the declarations holding them bind `batch` to `Dict[str, Any]` in one and `dict` in another, and the duplicated lines use it. The extracted unit therefore needs a parameter type that fits BOTH — their common supertype where they have one, or a new abstraction over them where they do not — and settling that is the step that comes BEFORE the extraction above. Where the two types are deliberately unrelated, the duplication is the price of that separation and the honest resolution is to record the decision rather than to extract.
PENMANT5Tokenizer._fix_and_make_graph (cyclomatic 75) hanlp/components/amr/seq2seq/dataset/tokenization_t5.py:347— PENMANT5Tokenizer._fix_and_make_graph has cyclomatic complexity 75 (threshold 15). To reduce it, separate the cases: extract each independent branch into its own named function, and where the body has guards that only reject input, fold those into early returns at the top.
AMRBartTokenizer._fix_and_make_graph (cyclomatic 74) hanlp/components/amr/amrbart/model_interface/tokenization_bart.py:174— AMRBartTokenizer._fix_and_make_graph has cyclomatic complexity 74 (threshold 15). To reduce it, separate the cases: extract each independent branch into its own named function, and where the body has guards that only reject input, fold those into early returns at the top.
PENMANBartTokenizer._fix_and_make_graph (cyclomatic 74) hanlp/components/amr/seq2seq/dataset/tokenization_bart.py:354— PENMANBartTokenizer._fix_and_make_graph has cyclomatic complexity 74 (threshold 15). To reduce it, separate the cases: extract each independent branch into its own named function, and where the body has guards that only reject input, fold those into early returns at the top.
Document.to_pretty (cyclomatic 57) plugins/hanlp_common/hanlp_common/document.py:188— Document.to_pretty has cyclomatic complexity 57 (threshold 15). To reduce it, separate the cases: extract each independent branch into its own named function, and where the body has guards that only reject input, fold those into early returns at the top.
TransformerSequenceTokenizer.__call__ (cyclomatic 55) hanlp/transform/transformer_tokenizer.py:246— TransformerSequenceTokenizer.__call__ has cyclomatic complexity 55 (threshold 15). To reduce it, separate the cases: extract each independent branch into its own named function, and where the body has guards that only reject input, fold those into early returns at the top.
BartDecoder.forward (cyclomatic 41) hanlp/components/amr/amrbart/model_interface/modeling_bart.py:925— BartDecoder.forward has cyclomatic complexity 41 (threshold 15). To reduce it, separate the cases: extract each independent branch into its own named function, and where the body has guards that only reject input, fold those into early returns at the top.
postprocessing.decode_into_node_and_backreferences_without_space (cyclomatic 40) REDACTED:191— postprocessing.decode_into_node_and_backreferences_without_space has cyclomatic complexity 40 (threshold 15). To reduce it, separate the cases: extract each independent branch into its own named function, and where the body has guards that only reject input, fold those into early returns at the top.
postprocessing.decode_into_node_and_backreferences (cyclomatic 36) REDACTED:31— postprocessing.decode_into_node_and_backreferences has cyclomatic complexity 36 (threshold 15). To reduce it, separate the cases: extract each independent branch into its own named function, and where the body has guards that only reject input, fold those into early returns at the top.
CoNLL_Transformer_Transform.batched_inputs_to_batches (cyclomatic 35) hanlp/transform/conll_tf.py:392— CoNLL_Transformer_Transform.batched_inputs_to_batches has cyclomatic complexity 35 (threshold 15). To reduce it, separate the cases: extract each independent branch into its own named function, and where the body has guards that only reject input, fold those into early returns at the top.
component_util.load_from_meta_file (cyclomatic 35) hanlp/utils/component_util.py:15— component_util.load_from_meta_file has cyclomatic complexity 35 (threshold 15). To reduce it, separate the cases: extract each independent branch into its own named function, and where the body has guards that only reject input, fold those into early returns at the top.
postprocessing.decode_into_node_and_backreferences (cyclomatic 33) REDACTED:55— postprocessing.decode_into_node_and_backreferences has cyclomatic complexity 33 (threshold 15). To reduce it, separate the cases: extract each independent branch into its own named function, and where the body has guards that only reject input, fold those into early returns at the top.
txt_tf.extract_ngram_features (cyclomatic 33) hanlp/transform/txt_tf.py:64— txt_tf.extract_ngram_features has cyclomatic complexity 33 (threshold 15). To reduce it, separate the cases: extract each independent branch into its own named function, and where the body has guards that only reject input, fold those into early returns at the top.
TorchComponent.fit (cyclomatic 30) hanlp/common/torch_component.py:189— TorchComponent.fit has cyclomatic complexity 30 (threshold 15). To reduce it, separate the cases: extract each independent branch into its own named function, and where the body has guards that only reject input, fold those into early returns at the top.
BartEncoder.forward (cyclomatic 26) hanlp/components/amr/amrbart/model_interface/modeling_bart.py:735— BartEncoder.forward has cyclomatic complexity 26 (threshold 15). To reduce it, separate the cases: extract each independent branch into its own named function, and where the body has guards that only reject input, fold those into early returns at the top.
chu_liu_edmonds.chu_liu_edmonds (cyclomatic 26) hanlp/components/parsers/chu_liu_edmonds.py:100— chu_liu_edmonds.chu_liu_edmonds has cyclomatic complexity 26 (threshold 15). To reduce it, separate the cases: extract each independent branch into its own named function, and where the body has guards that only reject input, fold those into early returns at the top.
visualization.pretty_tree_horizontal (cyclomatic 26) plugins/hanlp_common/hanlp_common/visualization.py:26— visualization.pretty_tree_horizontal has cyclomatic complexity 26 (threshold 15). To reduce it, separate the cases: extract each independent branch into its own named function, and where the body has guards that only reject input, fold those into early returns at the top.
MultiTaskLearning.predict (cyclomatic 25) hanlp/components/mtl/multi_task_learning.py:459— MultiTaskLearning.predict has cyclomatic complexity 25 (threshold 15). To reduce it, separate the cases: extract each independent branch into its own named function, and where the body has guards that only reject input, fold those into early returns at the top.
english_tokenizer.tokenize_english (cyclomatic 25) hanlp/utils/lang/en/english_tokenizer.py:223— english_tokenizer.tokenize_english has cyclomatic complexity 25 (threshold 15). To reduce it, separate the cases: extract each independent branch into its own named function, and where the body has guards that only reject input, fold those into early returns at the top.
TorchComponent.to (cyclomatic 24) hanlp/common/torch_component.py:513— TorchComponent.to has cyclomatic complexity 24 (threshold 15). To reduce it, separate the cases: extract each independent branch into its own named function, and where the body has guards that only reject input, fold those into early returns at the top.
AMRLinearizer._linearize (cyclomatic 24) hanlp/components/amr/seq2seq/dataset/linearization.py:179— AMRLinearizer._linearize has cyclomatic complexity 24 (threshold 15). To reduce it, separate the cases: extract each independent branch into its own named function, and where the body has guards that only reject input, fold those into early returns at the top.
hotpotqa.build_graph (cyclomatic 23) hanlp/datasets/qa/hotpotqa.py:147— hotpotqa.build_graph has cyclomatic complexity 23 (threshold 15). To reduce it, separate the cases: extract each independent branch into its own named function, and where the body has guards that only reject input, fold those into early returns at the top.
io_util.get_resource (cyclomatic 23) REDACTED:284— io_util.get_resource has cyclomatic complexity 23 (threshold 15). To reduce it, separate the cases: extract each independent branch into its own named function, and where the body has guards that only reject input, fold those into early returns at the top.
GraphRepair.remove_redundant_edges (cyclomatic 23) plugins/hanlp_common/hanlp_common/amr.py:904— GraphRepair.remove_redundant_edges has cyclomatic complexity 23 (threshold 15). To reduce it, separate the cases: extract each independent branch into its own named function, and where the body has guards that only reject input, fold those into early returns at the top.
io_util.uncompress (cyclomatic 22) REDACTED:204— io_util.uncompress has cyclomatic complexity 22 (threshold 15). To reduce it, separate the cases: extract each independent branch into its own named function, and where the body has guards that only reject input, fold those into early returns at the top.
span_util.is_transition_allowed (cyclomatic 22) hanlp/utils/span_util.py:147— span_util.is_transition_allowed has cyclomatic complexity 22 (threshold 15). To reduce it, separate the cases: extract each independent branch into its own named function, and where the body has guards that only reject input, fold those into early returns at the top.
TransformerNamedEntityRecognizer.tag_to_span (cyclomatic 20) hanlp/components/ner/transformer_ner.py:51— TransformerNamedEntityRecognizer.tag_to_span has cyclomatic complexity 20 (threshold 15). To reduce it, separate the cases: extract each independent branch into its own named function, and where the body has guards that only reject input, fold those into early returns at the top.
conll.read_conll (cyclomatic 20) hanlp/components/parsers/conll.py:27— conll.read_conll has cyclomatic complexity 20 (threshold 15). To reduce it, separate the cases: extract each independent branch into its own named function, and where the body has guards that only reject input, fold those into early returns at the top.
amr.batchify (cyclomatic 20) hanlp/datasets/parsing/amr.py:83— amr.batchify has cyclomatic complexity 20 (threshold 15). To reduce it, separate the cases: extract each independent branch into its own named function, and where the body has guards that only reject input, fold those into early returns at the top.
TransformerSequenceTokenizer.__init__ (cyclomatic 20) hanlp/transform/transformer_tokenizer.py:84— TransformerSequenceTokenizer.__init__ has cyclomatic complexity 20 (threshold 15). To reduce it, separate the cases: extract each independent branch into its own named function, and where the body has guards that only reject input, fold those into early returns at the top.
AMRGraph.get_list_data (cyclomatic 20) plugins/hanlp_common/hanlp_common/amr.py:549— AMRGraph.get_list_data has cyclomatic complexity 20 (threshold 15). To reduce it, separate the cases: extract each independent branch into its own named function, and where the body has guards that only reject input, fold those into early returns at the top.
ontonotes_loader.make_coref_instance (cyclomatic 19) hanlp/datasets/srl/loaders/ontonotes_loader.py:463— ontonotes_loader.make_coref_instance has cyclomatic complexity 19 (threshold 15). To reduce it, separate the cases: extract each independent branch into its own named function, and where the body has guards that only reject input, fold those into early returns at the top.
AMRGraph.from_prediction (cyclomatic 19) plugins/hanlp_common/hanlp_common/amr.py:724— AMRGraph.from_prediction has cyclomatic complexity 19 (threshold 15). To reduce it, separate the cases: extract each independent branch into its own named function, and where the body has guards that only reject input, fold those into early returns at the top.
KerasComponent.fit (cyclomatic 18) hanlp/common/keras_component.py:300— KerasComponent.fit has cyclomatic complexity 18 (threshold 15). To reduce it, separate the cases: extract each independent branch into its own named function, and where the body has guards that only reject input, fold those into early returns at the top.
BartForSequenceClassification.forward (cyclomatic 18) hanlp/components/amr/amrbart/model_interface/modeling_bart.py:1464— BartForSequenceClassification.forward has cyclomatic complexity 18 (threshold 15). To reduce it, separate the cases: extract each independent branch into its own named function, and where the body has guards that only reject input, fold those into early returns at the top.
TransformerEncoder.__init__ (cyclomatic 18) hanlp/layers/transformers/encoder.py:19— TransformerEncoder.__init__ has cyclomatic complexity 18 (threshold 15). To reduce it, separate the cases: extract each independent branch into its own named function, and where the body has guards that only reject input, fold those into early returns at the top.
BiaffineDependencyParser.build_vocabs (cyclomatic 17) hanlp/components/parsers/biaffine/biaffine_dep.py:357— BiaffineDependencyParser.build_vocabs has cyclomatic complexity 17 (threshold 15). To reduce it, split the body: these branches sit side by side rather than nested inside one another, so extracting each one on its own would leave a function per branch. Group the statements between the checks into named steps and move each step into its own function, so the body reads as a short sequence of named stages.
lemma_edit.apply_lemma_rule (cyclomatic 17) hanlp/components/parsers/ud/lemma_edit.py:79— lemma_edit.apply_lemma_rule has cyclomatic complexity 17 (threshold 15). To reduce it, separate the cases: extract each independent branch into its own named function, and where the body has guards that only reject input, fold those into early returns at the top.
Word2VecEmbeddingTF.__init__ (cyclomatic 17) hanlp/layers/embeddings/word2vec_tf.py:114— Word2VecEmbeddingTF.__init__ has cyclomatic complexity 17 (threshold 15). To reduce it, separate the cases: extract each independent branch into its own named function, and where the body has guards that only reject input, fold those into early returns at the top.
sequence_labeling.end_of_chunk (cyclomatic 17) hanlp/metrics/chunking/sequence_labeling.py:83— sequence_labeling.end_of_chunk has cyclomatic complexity 17 (threshold 15). To reduce it, split the body: these branches sit side by side rather than nested inside one another, so extracting each one on its own would leave a function per branch. Group the statements between the checks into named steps and move each step into its own function, so the body reads as a short sequence of named stages. This shape REPEATS in the file: one other method here (sequence_labeling.start_of_chunk) has the same decision points, in the same order, at the same nesting depths — so this is one pattern written twice rather than two separate problems. Splitting this body alone leaves the other exactly as it is. Where these are variations on one operation, the change that clears both is the shared one: lift the common shape into a single routine the variants call, parameterised by whatever genuinely differs between them, and keep in each method only the part that is not shared.
sequence_labeling.start_of_chunk (cyclomatic 17) hanlp/metrics/chunking/sequence_labeling.py:114— sequence_labeling.start_of_chunk has cyclomatic complexity 17 (threshold 15). To reduce it, split the body: these branches sit side by side rather than nested inside one another, so extracting each one on its own would leave a function per branch. Group the statements between the checks into named steps and move each step into its own function, so the body reads as a short sequence of named stages. This shape REPEATS in the file: one other method here (sequence_labeling.end_of_chunk) has the same decision points, in the same order, at the same nesting depths — so this is one pattern written twice rather than two separate problems. Splitting this body alone leaves the other exactly as it is. Where these are variations on one operation, the change that clears both is the shared one: lift the common shape into a single routine the variants call, parameterised by whatever genuinely differs between them, and keep in each method only the part that is not shared.
KerasComponent.evaluate (cyclomatic 16) hanlp/common/keras_component.py:49— KerasComponent.evaluate has cyclomatic complexity 16 (threshold 15). To reduce it, separate the cases: extract each independent branch into its own named function, and where the body has guards that only reject input, fold those into early returns at the top.
TorchComponent.evaluate (cyclomatic 16) hanlp/common/torch_component.py:445— TorchComponent.evaluate has cyclomatic complexity 16 (threshold 15). To reduce it, split the body: these branches sit side by side rather than nested inside one another, so extracting each one on its own would leave a function per branch. Group the statements between the checks into named steps and move each step into its own function, so the body reads as a short sequence of named stages.
AMRBartTokenizer.tokenize_amr (cyclomatic 16) hanlp/components/amr/seq2seq/dataset/tokenization_bart.py:125— AMRBartTokenizer.tokenize_amr has cyclomatic complexity 16 (threshold 15). To reduce it, separate the cases: extract each independent branch into its own named function, and where the body has guards that only reject input, fold those into early returns at the top.
AMRT5Tokenizer.tokenize_amr (cyclomatic 16) hanlp/components/amr/seq2seq/dataset/tokenization_t5.py:114— AMRT5Tokenizer.tokenize_amr has cyclomatic complexity 16 (threshold 15). To reduce it, separate the cases: extract each independent branch into its own named function, and where the body has guards that only reject input, fold those into early returns at the top.
NgramSentenceBoundaryDetector.predict (cyclomatic 16) hanlp/components/eos/ngram.py:191— NgramSentenceBoundaryDetector.predict has cyclomatic complexity 16 (threshold 15). To reduce it, separate the cases: extract each independent branch into its own named function, and where the body has guards that only reject input, fold those into early returns at the top.
inference_utils._dp_decode_non_overlapping_spans (cyclomatic 16) hanlp/components/srl/span_rank/inference_utils.py:147— inference_utils._dp_decode_non_overlapping_spans has cyclomatic complexity 16 (threshold 15). To reduce it, separate the cases: extract each independent branch into its own named function, and where the body has guards that only reject input, fold those into early returns at the top.
TransformerTaggingTokenizer.tag_to_span (cyclomatic 16) hanlp/components/tokenizers/transformer.py:115— TransformerTaggingTokenizer.tag_to_span has cyclomatic complexity 16 (threshold 15). To reduce it, separate the cases: extract each independent branch into its own named function, and where the body has guards that only reject input, fold those into early returns at the top.
util.index_word2vec_with_vocab (cyclomatic 16) hanlp/layers/embeddings/util.py:14— util.index_word2vec_with_vocab has cyclomatic complexity 16 (threshold 15). To reduce it, separate the cases: extract each independent branch into its own named function, and where the body has guards that only reject input, fold those into early returns at the top.
CoNLL_SDP_Transform.file_to_inputs (cyclomatic 16) hanlp/transform/conll_tf.py:628— CoNLL_SDP_Transform.file_to_inputs has cyclomatic complexity 16 (threshold 15). To reduce it, separate the cases: extract each independent branch into its own named function, and where the body has guards that only reject input, fold those into early returns at the top.
io_util.split_file (cyclomatic 16) REDACTED:485— io_util.split_file has cyclomatic complexity 16 (threshold 15). To reduce it, separate the cases: extract each independent branch into its own named function, and where the body has guards that only reject input, fold those into early returns at the top.
FixmeComment hanlp/components/srl/span_rank/srl_eval_utils.py:79— # FIXME: scalar index error — source code is not a task system: move the work to your tracker and leave a reference instead (e.g. `# REF: #123`), so the task is planned where tasks live and the ticket links back to the code.
Document.to_pretty (cognitive 201) plugins/hanlp_common/hanlp_common/document.py:188— Document.to_pretty has cognitive complexity 201 (threshold 15). Drivers by points: if/else 35 (130 pts), loops 18 (64 pts), boolean chains 4, ternaries 1 (3 pts) (nesting depth added 143). To reduce it, flatten the nesting: this score is depth rather than breadth — most of its points come from checks stacked inside one another, so the work sits several levels in. Invert each enclosing check into an early exit (a return, or the language's equivalent) so the happy path stays at one level, and where a level cannot be exited early, lift the block it encloses into its own named function.
TransformerSequenceTokenizer.__call__ (cognitive 162) hanlp/transform/transformer_tokenizer.py:246— TransformerSequenceTokenizer.__call__ has cognitive complexity 162 (threshold 15). Drivers by points: if/else 52 (124 pts), loops 8 (29 pts), ternaries 2 (5 pts), boolean chains 4 (nesting depth added 96). To reduce it, flatten the nesting: this score is depth rather than breadth — most of its points come from checks stacked inside one another, so the work sits several levels in. Invert each enclosing check into an early exit (a return, or the language's equivalent) so the happy path stays at one level, and where a level cannot be exited early, lift the block it encloses into its own named function.
PENMANT5Tokenizer._fix_and_make_graph (cognitive 143) hanlp/components/amr/seq2seq/dataset/tokenization_t5.py:347— PENMANT5Tokenizer._fix_and_make_graph has cognitive complexity 143 (threshold 15). Drivers by points: if/else 47 (118 pts), loops 11 (16 pts), boolean chains 9 (nesting depth added 76). To reduce it, flatten the nesting: this score is depth rather than breadth — most of its points come from checks stacked inside one another, so the work sits several levels in. Invert each enclosing check into an early exit (a return, or the language's equivalent) so the happy path stays at one level, and where a level cannot be exited early, lift the block it encloses into its own named function.
AMRBartTokenizer._fix_and_make_graph (cognitive 142) hanlp/components/amr/amrbart/model_interface/tokenization_bart.py:174— AMRBartTokenizer._fix_and_make_graph has cognitive complexity 142 (threshold 15). Drivers by points: if/else 46 (117 pts), loops 11 (16 pts), boolean chains 9 (nesting depth added 76). To reduce it, flatten the nesting: this score is depth rather than breadth — most of its points come from checks stacked inside one another, so the work sits several levels in. Invert each enclosing check into an early exit (a return, or the language's equivalent) so the happy path stays at one level, and where a level cannot be exited early, lift the block it encloses into its own named function.
PENMANBartTokenizer._fix_and_make_graph (cognitive 142) hanlp/components/amr/seq2seq/dataset/tokenization_bart.py:354— PENMANBartTokenizer._fix_and_make_graph has cognitive complexity 142 (threshold 15). Drivers by points: if/else 46 (117 pts), loops 11 (16 pts), boolean chains 9 (nesting depth added 76). To reduce it, flatten the nesting: this score is depth rather than breadth — most of its points come from checks stacked inside one another, so the work sits several levels in. Invert each enclosing check into an early exit (a return, or the language's equivalent) so the happy path stays at one level, and where a level cannot be exited early, lift the block it encloses into its own named function.
CoNLL_Transformer_Transform.batched_inputs_to_batches (cognitive 114) hanlp/transform/conll_tf.py:392— CoNLL_Transformer_Transform.batched_inputs_to_batches has cognitive complexity 114 (threshold 15). Drivers by points: loops 14 (50 pts), if/else 18 (38 pts), ternaries 7 (26 pts) (nesting depth added 75). To reduce it, split the body into named stages: move each independent step or branch into its own named function so the body reads as a short sequence of named calls rather than one long body.
MultiTaskLearning.predict (cognitive 66) hanlp/components/mtl/multi_task_learning.py:459— MultiTaskLearning.predict has cognitive complexity 66 (threshold 15). Drivers by points: if/else 13 (42 pts), loops 10 (22 pts), boolean chains 2 (nesting depth added 41). To reduce it, split the body into named stages: move each independent step or branch into its own named function so the body reads as a short sequence of named calls rather than one long body.
TransformerNamedEntityRecognizer.tag_to_span (cognitive 66) hanlp/components/ner/transformer_ner.py:51— TransformerNamedEntityRecognizer.tag_to_span has cognitive complexity 66 (threshold 15). Drivers by points: if/else 11 (35 pts), loops 7 (28 pts), boolean chains 3 (nesting depth added 45). To reduce it, flatten the nesting: this score is depth rather than breadth — most of its points come from checks stacked inside one another, so the work sits several levels in. Invert each enclosing check into an early exit (a return, or the language's equivalent) so the happy path stays at one level, and where a level cannot be exited early, lift the block it encloses into its own named function.
GraphRepair.remove_redundant_edges (cognitive 66) plugins/hanlp_common/hanlp_common/amr.py:904— GraphRepair.remove_redundant_edges has cognitive complexity 66 (threshold 15). Drivers by points: if/else 11 (40 pts), loops 8 (23 pts), boolean chains 3 (nesting depth added 44). To reduce it, split the body into named stages: move each independent step or branch into its own named function so the body reads as a short sequence of named calls rather than one long body.
BartDecoder.forward (cognitive 64) hanlp/components/amr/amrbart/model_interface/modeling_bart.py:925— BartDecoder.forward has cognitive complexity 64 (threshold 15). Drivers by points: if/else 17 (30 pts), ternaries 16 (27 pts), boolean chains 5, loops 2 (nesting depth added 24). To reduce it, split the body into named stages: move each independent step or branch into its own named function so the body reads as a short sequence of named calls rather than one long body.
component_util.load_from_meta_file (cognitive 64) hanlp/utils/component_util.py:15— component_util.load_from_meta_file has cognitive complexity 64 (threshold 15). Drivers by points: if/else 23 (39 pts), error handling 8 (16 pts), ternaries 1 (4 pts), loops 1 (3 pts), boolean chains 2 (nesting depth added 29). To reduce it, split the body into named stages: move each independent step or branch into its own named function so the body reads as a short sequence of named calls rather than one long body.
AMRLinearizer._linearize (cognitive 62) hanlp/components/amr/seq2seq/dataset/linearization.py:179— AMRLinearizer._linearize has cognitive complexity 62 (threshold 15). Drivers by points: if/else 18 (44 pts), loops 8 (16 pts), boolean chains 2 (nesting depth added 34). To reduce it, split the body into named stages: move each independent step or branch into its own named function so the body reads as a short sequence of named calls rather than one long body.
TorchComponent.to (cognitive 59) hanlp/common/torch_component.py:513— TorchComponent.to has cognitive complexity 59 (threshold 15). Drivers by points: if/else 14 (45 pts), loops 2 (7 pts), error handling 1 (5 pts), boolean chains 2 (nesting depth added 40). To reduce it, flatten the nesting: this score is depth rather than breadth — most of its points come from checks stacked inside one another, so the work sits several levels in. Invert each enclosing check into an early exit (a return, or the language's equivalent) so the happy path stays at one level, and where a level cannot be exited early, lift the block it encloses into its own named function.
postprocessing.decode_into_node_and_backreferences_without_space (cognitive 59) REDACTED:191— postprocessing.decode_into_node_and_backreferences_without_space has cognitive complexity 59 (threshold 15). Drivers by points: if/else 11 (20 pts), loops 7 (17 pts), boolean chains 12, ternaries 5 (8 pts), error handling 1 (2 pts) (nesting depth added 23). To reduce it, split the body into named stages: move each independent step or branch into its own named function so the body reads as a short sequence of named calls rather than one long body.
txt_tf.extract_ngram_features (cognitive 57) hanlp/transform/txt_tf.py:64— txt_tf.extract_ngram_features has cognitive complexity 57 (threshold 15). Drivers by points: ternaries 13 (35 pts), boolean chains 13, if/else 5 (9 pts) (nesting depth added 26). To reduce it, split the body into named stages: move each independent step or branch into its own named function so the body reads as a short sequence of named calls rather than one long body.
postprocessing.decode_into_node_and_backreferences (cognitive 55) REDACTED:31— postprocessing.decode_into_node_and_backreferences has cognitive complexity 55 (threshold 15). Drivers by points: if/else 11 (20 pts), loops 7 (17 pts), boolean chains 8, ternaries 5 (8 pts), error handling 1 (2 pts) (nesting depth added 23). To reduce it, split the body into named stages: move each independent step or branch into its own named function so the body reads as a short sequence of named calls rather than one long body.
english_tokenizer.tokenize_english (cognitive 53) hanlp/utils/lang/en/english_tokenizer.py:223— english_tokenizer.tokenize_english has cognitive complexity 53 (threshold 15). Drivers by points: if/else 12 (28 pts), loops 8 (20 pts), boolean chains 4, ternaries 1 (nesting depth added 28). To reduce it, split the body into named stages: move each independent step or branch into its own named function so the body reads as a short sequence of named calls rather than one long body.
io_util.uncompress (cognitive 52) REDACTED:204— io_util.uncompress has cognitive complexity 52 (threshold 15). Drivers by points: if/else 15 (34 pts), ternaries 4 (13 pts), error handling 2 (4 pts), boolean chains 1 (nesting depth added 30). To reduce it, flatten the nesting: this score is depth rather than breadth — most of its points come from checks stacked inside one another, so the work sits several levels in. Invert each enclosing check into an early exit (a return, or the language's equivalent) so the happy path stays at one level, and where a level cannot be exited early, lift the block it encloses into its own named function.
postprocessing.decode_into_node_and_backreferences (cognitive 51) REDACTED:55— postprocessing.decode_into_node_and_backreferences has cognitive complexity 51 (threshold 15). Drivers by points: if/else 9 (18 pts), loops 7 (17 pts), ternaries 5 (8 pts), boolean chains 6, error handling 1 (2 pts) (nesting depth added 23). To reduce it, split the body into named stages: move each independent step or branch into its own named function so the body reads as a short sequence of named calls rather than one long body.
chu_liu_edmonds.chu_liu_edmonds (cognitive 51) hanlp/components/parsers/chu_liu_edmonds.py:100— chu_liu_edmonds.chu_liu_edmonds has cognitive complexity 51 (threshold 15). Drivers by points: if/else 12 (32 pts), loops 11 (17 pts), boolean chains 2 (nesting depth added 26). To reduce it, split the body into named stages: move each independent step or branch into its own named function so the body reads as a short sequence of named calls rather than one long body.
conll.read_conll (cognitive 49) hanlp/components/parsers/conll.py:27— conll.read_conll has cognitive complexity 49 (threshold 15). Drivers by points: if/else 15 (35 pts), error handling 1 (5 pts), loops 2 (5 pts), boolean chains 4 (nesting depth added 27). To reduce it, split the body into named stages: move each independent step or branch into its own named function so the body reads as a short sequence of named calls rather than one long body.
CoNLL_SDP_Transform.file_to_inputs (cognitive 48) hanlp/transform/conll_tf.py:628— CoNLL_SDP_Transform.file_to_inputs has cognitive complexity 48 (threshold 15). Drivers by points: if/else 7 (24 pts), ternaries 4 (16 pts), loops 3 (7 pts), boolean chains 1 (nesting depth added 33). To reduce it, split the body into named stages: move each independent step or branch into its own named function so the body reads as a short sequence of named calls rather than one long body.
visualization.pretty_tree_horizontal (cognitive 48) plugins/hanlp_common/hanlp_common/visualization.py:26— visualization.pretty_tree_horizontal has cognitive complexity 48 (threshold 15). Drivers by points: if/else 10 (24 pts), loops 8 (15 pts), ternaries 3 (5 pts), boolean chains 4 (nesting depth added 23). To reduce it, split the body into named stages: move each independent step or branch into its own named function so the body reads as a short sequence of named calls rather than one long body.
hotpotqa.build_graph (cognitive 47) hanlp/datasets/qa/hotpotqa.py:147— hotpotqa.build_graph has cognitive complexity 47 (threshold 15). Drivers by points: if/else 10 (25 pts), loops 7 (10 pts), ternaries 3 (9 pts), boolean chains 3 (nesting depth added 24). To reduce it, split the body into named stages: move each independent step or branch into its own named function so the body reads as a short sequence of named calls rather than one long body.
TransformerTaggingTokenizer.tag_to_span (cognitive 46) hanlp/components/tokenizers/transformer.py:115— TransformerTaggingTokenizer.tag_to_span has cognitive complexity 46 (threshold 15). Drivers by points: if/else 10 (28 pts), loops 6 (18 pts) (nesting depth added 30). To reduce it, split the body into named stages: move each independent step or branch into its own named function so the body reads as a short sequence of named calls rather than one long body.
inference_utils._dp_decode_non_overlapping_spans (cognitive 44) hanlp/components/srl/span_rank/inference_utils.py:147— inference_utils._dp_decode_non_overlapping_spans has cognitive complexity 44 (threshold 15). Drivers by points: if/else 9 (25 pts), loops 6 (17 pts), boolean chains 2 (nesting depth added 27). To reduce it, split the body into named stages: move each independent step or branch into its own named function so the body reads as a short sequence of named calls rather than one long body.
ontonotes_loader.make_coref_instance (cognitive 44) hanlp/datasets/srl/loaders/ontonotes_loader.py:463— ontonotes_loader.make_coref_instance has cognitive complexity 44 (threshold 15). Drivers by points: if/else 11 (27 pts), loops 6 (15 pts), boolean chains 1, ternaries 1 (nesting depth added 25). To reduce it, split the body into named stages: move each independent step or branch into its own named function so the body reads as a short sequence of named calls rather than one long body.
semdep_eval.sdp_eval (cognitive 42) hanlp/metrics/parsing/semdep_eval.py:28— semdep_eval.sdp_eval has cognitive complexity 42 (threshold 15). Drivers by points: if/else 12 (31 pts), loops 4 (10 pts), boolean chains 1 (nesting depth added 25). To reduce it, split the body into named stages: move each independent step or branch into its own named function so the body reads as a short sequence of named calls rather than one long body.
BartEncoder.forward (cognitive 39) hanlp/components/amr/amrbart/model_interface/modeling_bart.py:735— BartEncoder.forward has cognitive complexity 39 (threshold 15). Drivers by points: if/else 15 (22 pts), ternaries 7 (13 pts), boolean chains 3, loops 1 (nesting depth added 13). To reduce it, split the body into named stages: move each independent step or branch into its own named function so the body reads as a short sequence of named calls rather than one long body.
amr.batchify (cognitive 39) hanlp/datasets/parsing/amr.py:83— amr.batchify has cognitive complexity 39 (threshold 15). Drivers by points: if/else 12 (21 pts), loops 4 (10 pts), ternaries 5 (6 pts), boolean chains 2 (nesting depth added 16). To reduce it, split the body into named stages: move each independent step or branch into its own named function so the body reads as a short sequence of named calls rather than one long body.
TorchComponent.fit (cognitive 38) hanlp/common/torch_component.py:189— TorchComponent.fit has cognitive complexity 38 (threshold 15). Drivers by points: if/else 18 (23 pts), boolean chains 8, ternaries 4 (7 pts) (nesting depth added 8). To reduce it, split the body into named stages: move each independent step or branch into its own named function so the body reads as a short sequence of named calls rather than one long body.
srl_eval_utils.compute_srl_f1 (cognitive 38) hanlp/components/srl/span_rank/srl_eval_utils.py:162— srl_eval_utils.compute_srl_f1 has cognitive complexity 38 (threshold 15). Drivers by points: if/else 8 (24 pts), loops 5 (12 pts), boolean chains 2 (nesting depth added 23). To reduce it, split the body into named stages: move each independent step or branch into its own named function so the body reads as a short sequence of named calls rather than one long body.
io_util.get_resource (cognitive 38) REDACTED:284— io_util.get_resource has cognitive complexity 38 (threshold 15). Drivers by points: if/else 15 (33 pts), boolean chains 5 (nesting depth added 18). To reduce it, flatten the nesting: this score is depth rather than breadth — most of its points come from checks stacked inside one another, so the work sits several levels in. Invert each enclosing check into an early exit (a return, or the language's equivalent) so the happy path stays at one level, and where a level cannot be exited early, lift the block it encloses into its own named function.
PadSequenceDataLoader.pad_data (cognitive 37) hanlp/common/dataset.py:468— PadSequenceDataLoader.pad_data has cognitive complexity 37 (threshold 15). Drivers by points: if/else 9 (28 pts), loops 3 (9 pts) (nesting depth added 25). To reduce it, flatten the nesting: this score is depth rather than breadth — most of its points come from checks stacked inside one another, so the work sits several levels in. Invert each enclosing check into an early exit (a return, or the language's equivalent) so the happy path stays at one level, and where a level cannot be exited early, lift the block it encloses into its own named function.
lemma_edit.apply_lemma_rule (cognitive 37) hanlp/components/parsers/ud/lemma_edit.py:79— lemma_edit.apply_lemma_rule has cognitive complexity 37 (threshold 15). Drivers by points: if/else 9 (18 pts), loops 5 (11 pts), ternaries 2 (5 pts), error handling 1 (2 pts), boolean chains 1 (nesting depth added 19). To reduce it, split the body into named stages: move each independent step or branch into its own named function so the body reads as a short sequence of named calls rather than one long body.
io_util.download (cognitive 35) REDACTED:144— io_util.download has cognitive complexity 35 (threshold 15). Drivers by points: if/else 14 (29 pts), error handling 2 (5 pts), boolean chains 1 (nesting depth added 18). To reduce it, split the body into named stages: move each independent step or branch into its own named function so the body reads as a short sequence of named calls rather than one long body.
RAdam.step (cognitive 34) hanlp/components/amr/seq2seq/optim.py:32— RAdam.step has cognitive complexity 34 (threshold 15). Drivers by points: if/else 12 (31 pts), loops 2 (3 pts) (nesting depth added 20). To reduce it, split the body into named stages: move each independent step or branch into its own named function so the body reads as a short sequence of named calls rather than one long body.
AMRGraph.from_prediction (cognitive 34) plugins/hanlp_common/hanlp_common/amr.py:724— AMRGraph.from_prediction has cognitive complexity 34 (threshold 15). Drivers by points: if/else 11 (24 pts), loops 5 (7 pts), boolean chains 2, error handling 1 (nesting depth added 15). To reduce it, split the body into named stages: move each independent step or branch into its own named function so the body reads as a short sequence of named calls rather than one long body.
Document.__init__ (cognitive 32) plugins/hanlp_common/hanlp_common/document.py:18— Document.__init__ has cognitive complexity 32 (threshold 15). Drivers by points: if/else 9 (26 pts), loops 2 (4 pts), boolean chains 2 (nesting depth added 19). To reduce it, split the body into named stages: move each independent step or branch into its own named function so the body reads as a short sequence of named calls rather than one long body.
MultiTaskLearning.fit_dataloader (cognitive 31) hanlp/components/mtl/multi_task_learning.py:323— MultiTaskLearning.fit_dataloader has cognitive complexity 31 (threshold 15). Drivers by points: if/else 10 (27 pts), ternaries 1 (2 pts), boolean chains 1, loops 1 (nesting depth added 18). To reduce it, split the body into named stages: move each independent step or branch into its own named function so the body reads as a short sequence of named calls rather than one long body.
AMRGraph.get_list_data (cognitive 31) plugins/hanlp_common/hanlp_common/amr.py:549— AMRGraph.get_list_data has cognitive complexity 31 (threshold 15). Drivers by points: if/else 11 (19 pts), loops 6 (10 pts), boolean chains 1, ternaries 1 (nesting depth added 12). To reduce it, split the body into named stages: move each independent step or branch into its own named function so the body reads as a short sequence of named calls rather than one long body.
Ontonotes._conll_rows_to_sentence (cognitive 29) hanlp/datasets/srl/loaders/ontonotes_loader.py:240— Ontonotes._conll_rows_to_sentence has cognitive complexity 29 (threshold 15). Drivers by points: if/else 10 (19 pts), ternaries 4 (8 pts), boolean chains 1, loops 1 (nesting depth added 13). To reduce it, split the body into named stages: move each independent step or branch into its own named function so the body reads as a short sequence of named calls rather than one long body.
span_util.to_bioul (cognitive 29) hanlp/utils/span_util.py:567— span_util.to_bioul has cognitive complexity 29 (threshold 15). Drivers by points: if/else 11 (26 pts), boolean chains 2, loops 1 (nesting depth added 15). To reduce it, flatten the nesting: this score is depth rather than breadth — most of its points come from checks stacked inside one another, so the work sits several levels in. Invert each enclosing check into an early exit (a return, or the language's equivalent) so the happy path stays at one level, and where a level cannot be exited early, lift the block it encloses into its own named function.
NgramSentenceBoundaryDetector.predict (cognitive 28) hanlp/components/eos/ngram.py:191— NgramSentenceBoundaryDetector.predict has cognitive complexity 28 (threshold 15). Drivers by points: if/else 9 (17 pts), loops 6 (11 pts) (nesting depth added 13). To reduce it, split the body into named stages: move each independent step or branch into its own named function so the body reads as a short sequence of named calls rather than one long body.
_ctb_utils.make_ctb_tasks (cognitive 28) hanlp/datasets/parsing/loaders/_ctb_utils.py:160— _ctb_utils.make_ctb_tasks has cognitive complexity 28 (threshold 15). Drivers by points: loops 7 (13 pts), if/else 3 (11 pts), error handling 1 (3 pts), boolean chains 1 (nesting depth added 16). To reduce it, break up the iteration: give each loop body a named function, and split a multi-phase loop into one function per phase so no single body carries the whole pipeline.
_utils.convert_jsonlines_to_IOBES (cognitive 28) hanlp/datasets/srl/ontonotes5/_utils.py:432— _utils.convert_jsonlines_to_IOBES has cognitive complexity 28 (threshold 15). Drivers by points: if/else 6 (14 pts), loops 5 (14 pts) (nesting depth added 17). To reduce it, split the body into named stages: move each independent step or branch into its own named function so the body reads as a short sequence of named calls rather than one long body.
BiaffineNamedEntityRecognizer.get_pred_ner (cognitive 27) hanlp/components/ner/biaffine_ner/biaffine_ner.py:368— BiaffineNamedEntityRecognizer.get_pred_ner has cognitive complexity 27 (threshold 15). Drivers by points: loops 8 (14 pts), if/else 3 (10 pts), boolean chains 3 (nesting depth added 13). To reduce it, break up the iteration: give each loop body a named function, and split a multi-phase loop into one function per phase so no single body carries the whole pipeline.
JsonNERDataset.load_file (cognitive 27) hanlp/datasets/ner/loaders/json_ner.py:34— JsonNERDataset.load_file has cognitive complexity 27 (threshold 15). Drivers by points: if/else 7 (21 pts), loops 3 (6 pts) (nesting depth added 17). To reduce it, split the body into named stages: move each independent step or branch into its own named function so the body reads as a short sequence of named calls rather than one long body.
TransformerEncoder.__init__ (cognitive 27) hanlp/layers/transformers/encoder.py:19— TransformerEncoder.__init__ has cognitive complexity 27 (threshold 15). Drivers by points: if/else 14 (22 pts), boolean chains 3, loops 1 (2 pts) (nesting depth added 9). To reduce it, split the body into named stages: move each independent step or branch into its own named function so the body reads as a short sequence of named calls rather than one long body.
TransformerSequenceTokenizer.__init__ (cognitive 27) hanlp/transform/transformer_tokenizer.py:84— TransformerSequenceTokenizer.__init__ has cognitive complexity 27 (threshold 15). Drivers by points: if/else 14 (19 pts), ternaries 3 (5 pts), boolean chains 3 (nesting depth added 7). To reduce it, split the body into named stages: move each independent step or branch into its own named function so the body reads as a short sequence of named calls rather than one long body.
io_util.generate_words_tags_from_tsv (cognitive 27) REDACTED:455— io_util.generate_words_tags_from_tsv has cognitive complexity 27 (threshold 15). Drivers by points: if/else 8 (19 pts), error handling 1 (4 pts), loops 2 (4 pts) (nesting depth added 16). To reduce it, split the body into named stages: move each independent step or branch into its own named function so the body reads as a short sequence of named calls rather than one long body.
span_util.is_transition_allowed (cognitive 27) hanlp/utils/span_util.py:147— span_util.is_transition_allowed has cognitive complexity 27 (threshold 15). Drivers by points: if/else 11 (19 pts), boolean chains 8 (nesting depth added 8). To reduce it, split the body into named stages: move each independent step or branch into its own named function so the body reads as a short sequence of named calls rather than one long body.
BiaffineDependencyParser.build_vocabs (cognitive 26) hanlp/components/parsers/biaffine/biaffine_dep.py:357— BiaffineDependencyParser.build_vocabs has cognitive complexity 26 (threshold 15). Drivers by points: if/else 12 (15 pts), loops 4 (9 pts), boolean chains 2 (nesting depth added 8). To reduce it, split the body: most of this score is breadth rather than depth — checks laid out side by side rather than stacked — so group the statements between the checks into named steps and move each step into its own function. Some of it IS depth: where a check sits inside another whose only job is to reach it, merge the two into one condition, and where an else follows a branch that already returns, drop the trailing else and let the rest of the body continue at one level.
_utils.handle_line (cognitive 26) hanlp/datasets/srl/ontonotes5/_utils.py:269— _utils.handle_line has cognitive complexity 26 (threshold 15). Drivers by points: if/else 10 (21 pts), loops 2 (5 pts) (nesting depth added 14). To reduce it, flatten the nesting: this score is depth rather than breadth — most of its points come from checks stacked inside one another, so the work sits several levels in. Invert each enclosing check into an early exit (a return, or the language's equivalent) so the happy path stays at one level, and where a level cannot be exited early, lift the block it encloses into its own named function.
CoNLL_SDP_Transform.batched_inputs_to_batches (cognitive 26) hanlp/transform/conll_tf.py:734— CoNLL_SDP_Transform.batched_inputs_to_batches has cognitive complexity 26 (threshold 15). Drivers by points: loops 7 (17 pts), if/else 3 (8 pts), ternaries 1 (nesting depth added 15). To reduce it, break up the iteration: give each loop body a named function, and split a multi-phase loop into one function per phase so no single body carries the whole pipeline.
io_util.split_file (cognitive 26) REDACTED:485— io_util.split_file has cognitive complexity 26 (threshold 15). Drivers by points: if/else 11 (17 pts), loops 6 (9 pts) (nesting depth added 9). To reduce it, split the body into named stages: move each independent step or branch into its own named function so the body reads as a short sequence of named calls rather than one long body.
util.collapse_json (cognitive 26) plugins/hanlp_common/hanlp_common/util.py:110— util.collapse_json has cognitive complexity 26 (threshold 15). Drivers by points: if/else 9 (25 pts), loops 1 (nesting depth added 16). To reduce it, split the body into named stages: move each independent step or branch into its own named function so the body reads as a short sequence of named calls rather than one long body.
PadSequenceDataLoader.__init__ (cognitive 25) hanlp/common/dataset.py:356— PadSequenceDataLoader.__init__ has cognitive complexity 25 (threshold 15). Drivers by points: if/else 12 (22 pts), loops 1 (2 pts), boolean chains 1 (nesting depth added 11). To reduce it, split the body into named stages: move each independent step or branch into its own named function so the body reads as a short sequence of named calls rather than one long body.
EncoderWithContextualLayer.__init__ (cognitive 25) hanlp/components/parsers/biaffine/biaffine_model.py:19— EncoderWithContextualLayer.__init__ has cognitive complexity 25 (threshold 15). Drivers by points: if/else 15 (25 pts) (nesting depth added 10). To reduce it, split the body into named stages: move each independent step or branch into its own named function so the body reads as a short sequence of named calls rather than one long body.
SpanRankingSemanticRoleLabeler.decode_output (cognitive 25) hanlp/components/srl/span_rank/span_rank.py:385— SpanRankingSemanticRoleLabeler.decode_output has cognitive complexity 25 (threshold 15). Drivers by points: if/else 4 (13 pts), loops 3 (9 pts), ternaries 1 (2 pts), boolean chains 1 (nesting depth added 16). To reduce it, split the body into named stages: move each independent step or branch into its own named function so the body reads as a short sequence of named calls rather than one long body.
srl_eval_utils.print_to_conll (cognitive 25) hanlp/components/srl/span_rank/srl_eval_utils.py:262— srl_eval_utils.print_to_conll has cognitive complexity 25 (threshold 15). Drivers by points: if/else 5 (13 pts), loops 4 (11 pts), boolean chains 1 (nesting depth added 15). To reduce it, split the body into named stages: move each independent step or branch into its own named function so the body reads as a short sequence of named calls rather than one long body.
TransformerTaggingTokenizer.spans_to_tokens (cognitive 25) hanlp/components/tokenizers/transformer.py:174— TransformerTaggingTokenizer.spans_to_tokens has cognitive complexity 25 (threshold 15). Drivers by points: if/else 8 (21 pts), loops 2 (4 pts) (nesting depth added 15). To reduce it, split the body into named stages: move each independent step or branch into its own named function so the body reads as a short sequence of named calls rather than one long body.
SentenceBoundaryDetectionDataset.load_file (cognitive 25) hanlp/datasets/eos/eos.py:48— SentenceBoundaryDetectionDataset.load_file has cognitive complexity 25 (threshold 15). Drivers by points: if/else 9 (16 pts), loops 3 (4 pts), ternaries 1 (3 pts), boolean chains 2 (nesting depth added 10). To reduce it, split the body into named stages: move each independent step or branch into its own named function so the body reads as a short sequence of named calls rather than one long body.
CoNLL2012SRLDataset.load_file (cognitive 25) hanlp/datasets/srl/loaders/conll2012.py:91— CoNLL2012SRLDataset.load_file has cognitive complexity 25 (threshold 15). Drivers by points: if/else 5 (16 pts), loops 4 (9 pts) (nesting depth added 16). To reduce it, split the body into named stages: move each independent step or branch into its own named function so the body reads as a short sequence of named calls rather than one long body.
KerasComponent.fit (cognitive 24) hanlp/common/keras_component.py:300— KerasComponent.fit has cognitive complexity 24 (threshold 15). Drivers by points: if/else 10 (14 pts), ternaries 4 (6 pts), boolean chains 3, error handling 1 (nesting depth added 6). To reduce it, split the body into named stages: move each independent step or branch into its own named function so the body reads as a short sequence of named calls rather than one long body.
dataset.dfs_linearize_constituency (cognitive 24) hanlp/components/amr/seq2seq/dataset/dataset.py:127— dataset.dfs_linearize_constituency has cognitive complexity 24 (threshold 15). Drivers by points: if/else 8 (17 pts), loops 3 (6 pts), boolean chains 1 (nesting depth added 12). To reduce it, split the body into named stages: move each independent step or branch into its own named function so the body reads as a short sequence of named calls rather than one long body.
dataset.dfs_linearize_tokenize_with_linguistic_structures (cognitive 24) hanlp/components/amr/seq2seq/dataset/dataset.py:171— dataset.dfs_linearize_tokenize_with_linguistic_structures has cognitive complexity 24 (threshold 15). Drivers by points: loops 5 (15 pts), if/else 5 (9 pts) (nesting depth added 14). To reduce it, break up the iteration: give each loop body a named function, and split a multi-phase loop into one function per phase so no single body carries the whole pipeline.
BiaffineDependencyParserTF.train_loop (cognitive 24) hanlp/components/parsers/biaffine_parser_tf.py:94— BiaffineDependencyParserTF.train_loop has cognitive complexity 24 (threshold 15). Drivers by points: loops 9 (17 pts), if/else 3 (7 pts) (nesting depth added 12). To reduce it, break up the iteration: give each loop body a named function, and split a multi-phase loop into one function per phase so no single body carries the whole pipeline.
transformer_tokenizer.convert_examples_to_features (cognitive 24) hanlp/transform/transformer_tokenizer.py:558— transformer_tokenizer.convert_examples_to_features has cognitive complexity 24 (threshold 15). Drivers by points: if/else 10 (12 pts), ternaries 5 (10 pts), boolean chains 1, loops 1 (nesting depth added 7). To reduce it, split the body into named stages: move each independent step or branch into its own named function so the body reads as a short sequence of named calls rather than one long body.
KerasComponent.evaluate (cognitive 23) hanlp/common/keras_component.py:49— KerasComponent.evaluate has cognitive complexity 23 (threshold 15). Drivers by points: if/else 10 (14 pts), loops 2 (4 pts), ternaries 2 (4 pts), boolean chains 1 (nesting depth added 8). To reduce it, split the body into named stages: move each independent step or branch into its own named function so the body reads as a short sequence of named calls rather than one long body.
BartForSequenceClassification.forward (cognitive 23) hanlp/components/amr/amrbart/model_interface/modeling_bart.py:1464— BartForSequenceClassification.forward has cognitive complexity 23 (threshold 15). Drivers by points: if/else 11 (17 pts), boolean chains 3, ternaries 2 (3 pts) (nesting depth added 7). To reduce it, split the body into named stages: move each independent step or branch into its own named function so the body reads as a short sequence of named calls rather than one long body.
lemma_edit.min_edit_script (cognitive 23) hanlp/components/parsers/ud/lemma_edit.py:9— lemma_edit.min_edit_script has cognitive complexity 23 (threshold 15). Drivers by points: if/else 5 (16 pts), boolean chains 4, loops 2 (3 pts) (nesting depth added 12). To reduce it, split the body into named stages: move each independent step or branch into its own named function so the body reads as a short sequence of named calls rather than one long body.
lemma_edit.gen_lemma_rule (cognitive 23) hanlp/components/parsers/ud/lemma_edit.py:35— lemma_edit.gen_lemma_rule has cognitive complexity 23 (threshold 15). Drivers by points: ternaries 3 (8 pts), if/else 4 (7 pts), loops 4 (7 pts), boolean chains 1 (nesting depth added 11). To reduce it, split the body into named stages: move each independent step or branch into its own named function so the body reads as a short sequence of named calls rather than one long body.
SpanRankingSemanticRoleLabeler.evaluate_dataloader (cognitive 23) hanlp/components/srl/span_rank/span_rank.py:156— SpanRankingSemanticRoleLabeler.evaluate_dataloader has cognitive complexity 23 (threshold 15). Drivers by points: loops 4 (13 pts), if/else 6 (10 pts) (nesting depth added 13). To reduce it, break up the iteration: give each loop body a named function, and split a multi-phase loop into one function per phase so no single body carries the whole pipeline.
util.index_word2vec_with_vocab (cognitive 23) hanlp/layers/embeddings/util.py:14— util.index_word2vec_with_vocab has cognitive complexity 23 (threshold 15). Drivers by points: if/else 10 (15 pts), loops 2 (3 pts), ternaries 1 (3 pts), boolean chains 2 (nesting depth added 8). To reduce it, split the body into named stages: move each independent step or branch into its own named function so the body reads as a short sequence of named calls rather than one long body.
utils_tf.convert_examples_to_features (cognitive 23) hanlp/layers/transformers/utils_tf.py:13— utils_tf.convert_examples_to_features has cognitive complexity 23 (threshold 15). Drivers by points: if/else 10 (12 pts), ternaries 5 (10 pts), loops 1 (nesting depth added 7). To reduce it, split the body into named stages: move each independent step or branch into its own named function so the body reads as a short sequence of named calls rather than one long body.
Document.to_dict (cognitive 23) plugins/hanlp_common/hanlp_common/document.py:104— Document.to_dict has cognitive complexity 23 (threshold 15). Drivers by points: if/else 6 (17 pts), loops 2 (4 pts), boolean chains 2 (nesting depth added 13). To reduce it, split the body into named stages: move each independent step or branch into its own named function so the body reads as a short sequence of named calls rather than one long body.
Seq2seq_AMR_Parser._init_new_embeddings (cognitive 22) hanlp/components/amr/seq2seq/seq2seq_amr_parser.py:325— Seq2seq_AMR_Parser._init_new_embeddings has cognitive complexity 22 (threshold 15). Drivers by points: if/else 8 (16 pts), loops 3 (5 pts), boolean chains 1 (nesting depth added 10). To reduce it, split the body into named stages: move each independent step or branch into its own named function so the body reads as a short sequence of named calls rather than one long body.
MultiTaskLearning.build_dataloader (cognitive 22) hanlp/components/mtl/multi_task_learning.py:114— MultiTaskLearning.build_dataloader has cognitive complexity 22 (threshold 15). Drivers by points: if/else 8 (13 pts), boolean chains 3, loops 2 (3 pts), ternaries 1 (3 pts) (nesting depth added 8). To reduce it, split the body into named stages: move each independent step or branch into its own named function so the body reads as a short sequence of named calls rather than one long body.
EncoderWithContextualLayer.forward (cognitive 22) hanlp/components/parsers/biaffine/biaffine_model.py:89— EncoderWithContextualLayer.forward has cognitive complexity 22 (threshold 15). Drivers by points: if/else 15 (22 pts) (nesting depth added 7). To reduce it, split the body into named stages: move each independent step or branch into its own named function so the body reads as a short sequence of named calls rather than one long body.
BiaffineSemanticDependencyParser.cache_dataset (cognitive 22) hanlp/components/parsers/biaffine/biaffine_sdp.py:79— BiaffineSemanticDependencyParser.cache_dataset has cognitive complexity 22 (threshold 15). Drivers by points: if/else 8 (16 pts), loops 2 (5 pts), boolean chains 1 (nesting depth added 11). To reduce it, split the body into named stages: move each independent step or branch into its own named function so the body reads as a short sequence of named calls rather than one long body.
semeval15.unpack_deps_to_head_deprel (cognitive 22) hanlp/datasets/parsing/semeval15.py:16— semeval15.unpack_deps_to_head_deprel has cognitive complexity 22 (threshold 15). Drivers by points: if/else 4 (11 pts), loops 3 (8 pts), ternaries 1 (3 pts) (nesting depth added 14). To reduce it, split the body into named stages: move each independent step or branch into its own named function so the body reads as a short sequence of named calls rather than one long body.
utils.transformer_encode (cognitive 22) hanlp/layers/transformers/utils.py:14— utils.transformer_encode has cognitive complexity 22 (threshold 15). Drivers by points: if/else 14 (21 pts), boolean chains 1 (nesting depth added 7). To reduce it, split the body into named stages: move each independent step or branch into its own named function so the body reads as a short sequence of named calls rather than one long body.
CoNLL_SDP_Transform.fit (cognitive 22) hanlp/transform/conll_tf.py:678— CoNLL_SDP_Transform.fit has cognitive complexity 22 (threshold 15). Drivers by points: if/else 9 (18 pts), loops 3 (4 pts) (nesting depth added 10). To reduce it, split the body into named stages: move each independent step or branch into its own named function so the body reads as a short sequence of named calls rather than one long body.
english_tokenizer.split_contractions (cognitive 22) hanlp/utils/lang/en/english_tokenizer.py:134— english_tokenizer.split_contractions has cognitive complexity 22 (threshold 15). Drivers by points: if/else 4 (16 pts), loops 2 (5 pts), boolean chains 1 (nesting depth added 15). To reduce it, flatten the nesting: this score is depth rather than breadth — most of its points come from checks stacked inside one another, so the work sits several levels in. Invert each enclosing check into an early exit (a return, or the language's equivalent) so the happy path stays at one level, and where a level cannot be exited early, lift the block it encloses into its own named function.
TorchComponent.evaluate (cognitive 21) hanlp/common/torch_component.py:445— TorchComponent.evaluate has cognitive complexity 21 (threshold 15). Drivers by points: if/else 11 (13 pts), ternaries 3 (5 pts), boolean chains 2, loops 1 (nesting depth added 4). To reduce it, split the body: most of this score is breadth rather than depth — checks laid out side by side rather than stacked — so group the statements between the checks into named steps and move each step into its own function. Some of it IS depth: where a check sits inside another whose only job is to reach it, merge the two into one condition, and where an else follows a branch that already returns, drop the trailing else and let the rest of the body continue at one level.
read_and_process.dfs_linearize (cognitive 21) hanlp/components/amr/amrbart/preprocess/read_and_process.py:50— read_and_process.dfs_linearize has cognitive complexity 21 (threshold 15). Drivers by points: if/else 6 (16 pts), loops 2 (4 pts), boolean chains 1 (nesting depth added 12). To reduce it, split the body into named stages: move each independent step or branch into its own named function so the body reads as a short sequence of named calls rather than one long body.
AMRBartTokenizer.tokenize_amr (cognitive 21) hanlp/components/amr/seq2seq/dataset/tokenization_bart.py:125— AMRBartTokenizer.tokenize_amr has cognitive complexity 21 (threshold 15). Drivers by points: if/else 8 (15 pts), boolean chains 5, loops 1 (nesting depth added 7). To reduce it, split the body into named stages: move each independent step or branch into its own named function so the body reads as a short sequence of named calls rather than one long body.
PENMANBartTokenizer._get_nodes_and_backreferences (cognitive 21) hanlp/components/amr/seq2seq/dataset/tokenization_bart.py:295— PENMANBartTokenizer._get_nodes_and_backreferences has cognitive complexity 21 (threshold 15). Drivers by points: if/else 6 (16 pts), loops 2 (4 pts), boolean chains 1 (nesting depth added 12). To reduce it, split the body into named stages: move each independent step or branch into its own named function so the body reads as a short sequence of named calls rather than one long body.
AMRT5Tokenizer.tokenize_amr (cognitive 21) hanlp/components/amr/seq2seq/dataset/tokenization_t5.py:114— AMRT5Tokenizer.tokenize_amr has cognitive complexity 21 (threshold 15). Drivers by points: if/else 8 (15 pts), boolean chains 5, loops 1 (nesting depth added 7). To reduce it, split the body into named stages: move each independent step or branch into its own named function so the body reads as a short sequence of named calls rather than one long body.
PENMANT5Tokenizer._get_nodes_and_backreferences (cognitive 21) hanlp/components/amr/seq2seq/dataset/tokenization_t5.py:288— PENMANT5Tokenizer._get_nodes_and_backreferences has cognitive complexity 21 (threshold 15). Drivers by points: if/else 6 (16 pts), loops 2 (4 pts), boolean chains 1 (nesting depth added 12). To reduce it, split the body into named stages: move each independent step or branch into its own named function so the body reads as a short sequence of named calls rather than one long body.
Pipe.predict (cognitive 21) hanlp/components/pipeline.py:33— Pipe.predict has cognitive complexity 21 (threshold 15). Drivers by points: if/else 13 (18 pts), loops 1 (3 pts) (nesting depth added 7). To reduce it, split the body into named stages: move each independent step or branch into its own named function so the body reads as a short sequence of named calls rather than one long body.
Word2VecEmbeddingTF.__init__ (cognitive 21) hanlp/layers/embeddings/word2vec_tf.py:114— Word2VecEmbeddingTF.__init__ has cognitive complexity 21 (threshold 15). Drivers by points: if/else 9 (11 pts), ternaries 2 (4 pts), boolean chains 3, loops 2 (3 pts) (nesting depth added 5). To reduce it, split the body into named stages: move each independent step or branch into its own named function so the body reads as a short sequence of named calls rather than one long body.
utils.transformer_sliding_window (cognitive 21) hanlp/layers/transformers/utils.py:101— utils.transformer_sliding_window has cognitive complexity 21 (threshold 15). Drivers by points: if/else 14 (19 pts), ternaries 1 (2 pts) (nesting depth added 6). To reduce it, split the body into named stages: move each independent step or branch into its own named function so the body reads as a short sequence of named calls rather than one long body.
TrieDict.merge_batch (cognitive 21) plugins/hanlp_trie/hanlp_trie/dictionary.py:106— TrieDict.merge_batch has cognitive complexity 21 (threshold 15). Drivers by points: if/else 5 (11 pts), loops 5 (10 pts) (nesting depth added 11). To reduce it, split the body into named stages: move each independent step or branch into its own named function so the body reads as a short sequence of named calls rather than one long body.
TransformerClassifier.build_dataloader (cognitive 20) hanlp/components/classifiers/transformer_classifier.py:259— TransformerClassifier.build_dataloader has cognitive complexity 20 (threshold 15). Drivers by points: if/else 9 (15 pts), loops 1 (3 pts), boolean chains 2 (nesting depth added 8). To reduce it, split the body into named stages: move each independent step or branch into its own named function so the body reads as a short sequence of named calls rather than one long body.
inference_utils.greedy_decode (cognitive 20) hanlp/components/srl/span_rank/inference_utils.py:34— inference_utils.greedy_decode has cognitive complexity 20 (threshold 15). Drivers by points: if/else 5 (10 pts), loops 4 (9 pts), boolean chains 1 (nesting depth added 10). To reduce it, split the body into named stages: move each independent step or branch into its own named function so the body reads as a short sequence of named calls rather than one long body.
LanguageModelDataset.load_file (cognitive 20) hanlp/datasets/lm/loaders/lm_dataset.py:66— LanguageModelDataset.load_file has cognitive complexity 20 (threshold 15). Drivers by points: if/else 6 (13 pts), loops 2 (5 pts), boolean chains 2 (nesting depth added 10). To reduce it, split the body into named stages: move each independent step or branch into its own named function so the body reads as a short sequence of named calls rather than one long body.
_utils.filter_data (cognitive 20) hanlp/datasets/srl/ontonotes5/_utils.py:147— _utils.filter_data has cognitive complexity 20 (threshold 15). Drivers by points: loops 5 (10 pts), if/else 5 (7 pts), boolean chains 3 (nesting depth added 7). To reduce it, break up the iteration: give each loop body a named function, and split a multi-phase loop into one function per phase so no single body carries the whole pipeline.
string_util.split_long_sentence_into (cognitive 20) hanlp/utils/string_util.py:17— string_util.split_long_sentence_into has cognitive complexity 20 (threshold 15). Drivers by points: if/else 9 (13 pts), boolean chains 4, loops 2 (3 pts) (nesting depth added 5). To reduce it, split the body into named stages: move each independent step or branch into its own named function so the body reads as a short sequence of named calls rather than one long body.
Document.translate (cognitive 20) plugins/hanlp_common/hanlp_common/document.py:436— Document.translate has cognitive complexity 20 (threshold 15). Drivers by points: if/else 7 (13 pts), loops 3 (6 pts), boolean chains 1 (nesting depth added 9). To reduce it, split the body into named stages: move each independent step or branch into its own named function so the body reads as a short sequence of named calls rather than one long body.
PadSequenceDataLoader.tensorize (cognitive 19) hanlp/common/dataset.py:434— PadSequenceDataLoader.tensorize has cognitive complexity 19 (threshold 15). Drivers by points: if/else 5 (9 pts), ternaries 2 (4 pts), boolean chains 3, loops 2 (3 pts) (nesting depth added 7). To reduce it, split the body into named stages: move each independent step or branch into its own named function so the body reads as a short sequence of named calls rather than one long body.
BART_AMR_Parser.evaluate_dataloader (cognitive 19) hanlp/components/amr/amrbart/bart_amr_parser.py:170— BART_AMR_Parser.evaluate_dataloader has cognitive complexity 19 (threshold 15). Drivers by points: if/else 5 (10 pts), error handling 1 (3 pts), loops 2 (3 pts), ternaries 1 (3 pts) (nesting depth added 10). To reduce it, split the body into named stages: move each independent step or branch into its own named function so the body reads as a short sequence of named calls rather than one long body.
Seq2seq_AMR_Parser.evaluate_dataloader (cognitive 19) hanlp/components/amr/seq2seq/seq2seq_amr_parser.py:221— Seq2seq_AMR_Parser.evaluate_dataloader has cognitive complexity 19 (threshold 15). Drivers by points: if/else 5 (10 pts), error handling 1 (3 pts), loops 2 (3 pts), ternaries 1 (3 pts) (nesting depth added 10). To reduce it, split the body into named stages: move each independent step or branch into its own named function so the body reads as a short sequence of named calls rather than one long body.
TransformerClassifierHF.predict (cognitive 19) hanlp/components/classifiers/transformer_classifier_hf.py:66— TransformerClassifierHF.predict has cognitive complexity 19 (threshold 15). Drivers by points: if/else 9 (16 pts), loops 2 (3 pts) (nesting depth added 8). To reduce it, split the body into named stages: move each independent step or branch into its own named function so the body reads as a short sequence of named calls rather than one long body.
NgramSentenceBoundaryDetector.build_vocabs (cognitive 19) hanlp/components/eos/ngram.py:273— NgramSentenceBoundaryDetector.build_vocabs has cognitive complexity 19 (threshold 15). Drivers by points: loops 5 (12 pts), if/else 4 (7 pts) (nesting depth added 10). To reduce it, break up the iteration: give each loop body a named function, and split a multi-phase loop into one function per phase so no single body carries the whole pipeline.
BiaffineDependencyParserTF.evaluate_dataset (cognitive 19) hanlp/components/parsers/biaffine_parser_tf.py:291— BiaffineDependencyParserTF.evaluate_dataset has cognitive complexity 19 (threshold 15). Drivers by points: loops 6 (10 pts), if/else 7 (9 pts) (nesting depth added 6). To reduce it, break up the iteration: give each loop body a named function, and split a multi-phase loop into one function per phase so no single body carries the whole pipeline.
SpanRankingSemanticRoleLabeler.format_dict_to_results (cognitive 19) hanlp/components/srl/span_rank/span_rank.py:291— SpanRankingSemanticRoleLabeler.format_dict_to_results has cognitive complexity 19 (threshold 15). Drivers by points: if/else 5 (16 pts), loops 2 (3 pts) (nesting depth added 12). To reduce it, split the body into named stages: move each independent step or branch into its own named function so the body reads as a short sequence of named calls rather than one long body.
srl_eval_utils.evaluate_retrieval (cognitive 19) hanlp/components/srl/span_rank/srl_eval_utils.py:43— srl_eval_utils.evaluate_retrieval has cognitive complexity 19 (threshold 15). Drivers by points: if/else 9 (18 pts), loops 1 (nesting depth added 9). To reduce it, split the body into named stages: move each independent step or branch into its own named function so the body reads as a short sequence of named calls rather than one long body.
util_tf.build_embedding (cognitive 19) hanlp/layers/embeddings/util_tf.py:23— util_tf.build_embedding has cognitive complexity 19 (threshold 15). Drivers by points: if/else 8 (15 pts), loops 1 (3 pts), boolean chains 1 (nesting depth added 9). To reduce it, split the body into named stages: move each independent step or branch into its own named function so the body reads as a short sequence of named calls rather than one long body.
conlleval.evaluate_iobes (cognitive 19) hanlp/metrics/chunking/conlleval.py:151— conlleval.evaluate_iobes has cognitive complexity 19 (threshold 15). Drivers by points: if/else 7 (14 pts), boolean chains 4, loops 1 (nesting depth added 7). To reduce it, split the body into named stages: move each independent step or branch into its own named function so the body reads as a short sequence of named calls rather than one long body.
TableTransform.file_to_inputs (cognitive 19) REDACTED:26— TableTransform.file_to_inputs has cognitive complexity 19 (threshold 15). Drivers by points: if/else 9 (18 pts), loops 1 (nesting depth added 9). To reduce it, split the body into named stages: move each independent step or branch into its own named function so the body reads as a short sequence of named calls rather than one long body.
span_util.bio_tags_to_spans (cognitive 19) hanlp/utils/span_util.py:314— span_util.bio_tags_to_spans has cognitive complexity 19 (threshold 15). Drivers by points: if/else 7 (15 pts), boolean chains 3, loops 1 (nesting depth added 8). To reduce it, split the body into named stages: move each independent step or branch into its own named function so the body reads as a short sequence of named calls rather than one long body.
CoNLLSentence.from_str (cognitive 19) plugins/hanlp_common/hanlp_common/conll.py:197— CoNLLSentence.from_str has cognitive complexity 19 (threshold 15). Drivers by points: if/else 6 (11 pts), ternaries 2 (5 pts), loops 2 (3 pts) (nesting depth added 9). To reduce it, split the body into named stages: move each independent step or branch into its own named function so the body reads as a short sequence of named calls rather than one long body.
Vocab3D.__call__ (cognitive 18) hanlp/common/vocab.py:475— Vocab3D.__call__ has cognitive complexity 18 (threshold 15). Drivers by points: loops 3 (9 pts), if/else 5 (8 pts), boolean chains 1 (nesting depth added 9). To reduce it, break up the iteration: give each loop body a named function, and split a multi-phase loop into one function per phase so no single body carries the whole pipeline.
AMRBartTokenizer.tokenize_amr (cognitive 18) hanlp/components/amr/amrbart/model_interface/tokenization_bart.py:97— AMRBartTokenizer.tokenize_amr has cognitive complexity 18 (threshold 15). Drivers by points: if/else 6 (12 pts), boolean chains 5, loops 1 (nesting depth added 6). To reduce it, split the body into named stages: move each independent step or branch into its own named function so the body reads as a short sequence of named calls rather than one long body.
alg_tf.tarjan (cognitive 18) hanlp/components/parsers/alg_tf.py:174— alg_tf.tarjan has cognitive complexity 18 (threshold 15). Drivers by points: if/else 5 (12 pts), loops 2 (6 pts) (nesting depth added 11). This file's own header attributes it to another copyright holder, so it is code this repository carries rather than code it wrote: restructuring the body in place forks it from upstream and turns every future re-sync into a manual merge. The performable moves are to leave the body as close to its upstream form as possible and keep it behind a narrow interface of your own, and to re-sync it when upstream changes — or, if it has already diverged far enough that you maintain it here, adopt it deliberately and then split the body into named stages.
alg.tarjan (cognitive 18) hanlp/components/parsers/biaffine_tf/alg.py:174— alg.tarjan has cognitive complexity 18 (threshold 15). Drivers by points: if/else 5 (12 pts), loops 2 (6 pts) (nesting depth added 11). This file's own header attributes it to another copyright holder, so it is code this repository carries rather than code it wrote: restructuring the body in place forks it from upstream and turns every future re-sync into a manual merge. The performable moves are to leave the body as close to its upstream form as possible and keep it behind a narrow interface of your own, and to re-sync it when upstream changes — or, if it has already diverged far enough that you maintain it here, adopt it deliberately and then split the body into named stages.
amr.make_batch_for_squeeze (cognitive 18) hanlp/datasets/parsing/amr.py:210— amr.make_batch_for_squeeze has cognitive complexity 18 (threshold 15). Drivers by points: loops 4 (12 pts), if/else 3 (6 pts) (nesting depth added 11). To reduce it, break up the iteration: give each loop body a named function, and split a multi-phase loop into one function per phase so no single body carries the whole pipeline.
amr.linearize (cognitive 18) hanlp/datasets/parsing/amr.py:259— amr.linearize has cognitive complexity 18 (threshold 15). Drivers by points: if/else 3 (10 pts), loops 6 (8 pts) (nesting depth added 9). To reduce it, split the body into named stages: move each independent step or branch into its own named function so the body reads as a short sequence of named calls rather than one long body.
KerasComponent.build (cognitive 17) hanlp/common/keras_component.py:223— KerasComponent.build has cognitive complexity 17 (threshold 15). Drivers by points: if/else 7 (11 pts), loops 1 (4 pts), boolean chains 1, ternaries 1 (nesting depth added 7). To reduce it, split the body into named stages: move each independent step or branch into its own named function so the body reads as a short sequence of named calls rather than one long body.
BiaffineDependencyParser.predictions_to_human (cognitive 17) hanlp/components/parsers/biaffine/biaffine_dep.py:108— BiaffineDependencyParser.predictions_to_human has cognitive complexity 17 (threshold 15). Drivers by points: loops 4 (10 pts), if/else 4 (7 pts) (nesting depth added 9). To reduce it, break up the iteration: give each loop body a named function, and split a multi-phase loop into one function per phase so no single body carries the whole pipeline.
UniversalDependenciesDecoder.forward (cognitive 17) hanlp/components/parsers/ud/ud_model.py:89— UniversalDependenciesDecoder.forward has cognitive complexity 17 (threshold 15). Drivers by points: if/else 9 (16 pts), loops 1 (nesting depth added 7). To reduce it, split the body into named stages: move each independent step or branch into its own named function so the body reads as a short sequence of named calls rather than one long body.
conll_dataset.get_sibs (cognitive 17) hanlp/datasets/parsing/loaders/conll_dataset.py:90— conll_dataset.get_sibs has cognitive complexity 17 (threshold 15). Drivers by points: if/else 4 (11 pts), loops 2 (5 pts), boolean chains 1 (nesting depth added 10). To reduce it, split the body into named stages: move each independent step or branch into its own named function so the body reads as a short sequence of named calls rather than one long body.
CoNLL_DEP_Transform.inputs_to_samples (cognitive 17) hanlp/transform/conll_tf.py:210— CoNLL_DEP_Transform.inputs_to_samples has cognitive complexity 17 (threshold 15). Drivers by points: if/else 4 (12 pts), loops 2 (3 pts), ternaries 1 (2 pts) (nesting depth added 10). To reduce it, split the body into named stages: move each independent step or branch into its own named function so the body reads as a short sequence of named calls rather than one long body.
CoNLL_DEP_Transform.XY_to_inputs_outputs (cognitive 17) hanlp/transform/conll_tf.py:232— CoNLL_DEP_Transform.XY_to_inputs_outputs has cognitive complexity 17 (threshold 15). Drivers by points: if/else 5 (9 pts), ternaries 1 (4 pts), loops 2 (3 pts), boolean chains 1 (nesting depth added 8). To reduce it, split the body into named stages: move each independent step or branch into its own named function so the body reads as a short sequence of named calls rather than one long body.
CoNLL_DEP_Transform.fit (cognitive 17) hanlp/transform/conll_tf.py:255— CoNLL_DEP_Transform.fit has cognitive complexity 17 (threshold 15). Drivers by points: if/else 7 (13 pts), loops 3 (4 pts) (nesting depth added 7). To reduce it, split the body into named stages: move each independent step or branch into its own named function so the body reads as a short sequence of named calls rather than one long body.
span_util.bioul_tags_to_spans (cognitive 17) hanlp/utils/span_util.py:465— span_util.bioul_tags_to_spans has cognitive complexity 17 (threshold 15). Drivers by points: if/else 4 (11 pts), loops 2 (4 pts), boolean chains 2 (nesting depth added 9). To reduce it, split the body into named stages: move each independent step or branch into its own named function so the body reads as a short sequence of named calls rather than one long body. This shape REPEATS in the file: one other method here (span_util.iobes_tags_to_spans) has the same decision points, in the same order, at the same nesting depths — so this is one pattern written twice rather than two separate problems. Splitting this body alone leaves the other exactly as it is. Where these are variations on one operation, the change that clears both is the shared one: lift the common shape into a single routine the variants call, parameterised by whatever genuinely differs between them, and keep in each method only the part that is not shared.
span_util.iobes_tags_to_spans (cognitive 17) hanlp/utils/span_util.py:512— span_util.iobes_tags_to_spans has cognitive complexity 17 (threshold 15). Drivers by points: if/else 4 (11 pts), loops 2 (4 pts), boolean chains 2 (nesting depth added 9). To reduce it, split the body into named stages: move each independent step or branch into its own named function so the body reads as a short sequence of named calls rather than one long body. This shape REPEATS in the file: one other method here (span_util.bioul_tags_to_spans) has the same decision points, in the same order, at the same nesting depths — so this is one pattern written twice rather than two separate problems. Splitting this body alone leaves the other exactly as it is. Where these are variations on one operation, the change that clears both is the shared one: lift the common shape into a single routine the variants call, parameterised by whatever genuinely differs between them, and keep in each method only the part that is not shared.
conll.isprojective (cognitive 17) plugins/hanlp_common/hanlp_common/conll.py:359— conll.isprojective has cognitive complexity 17 (threshold 15). Drivers by points: if/else 4 (10 pts), boolean chains 4, loops 2 (3 pts) (nesting depth added 7). To reduce it, split the body into named stages: move each independent step or branch into its own named function so the body reads as a short sequence of named calls rather than one long body.
TupleTrieDict.parse_longest (cognitive 17) plugins/hanlp_trie/hanlp_trie/dictionary.py:184— TupleTrieDict.parse_longest has cognitive complexity 17 (threshold 15). Drivers by points: if/else 4 (13 pts), loops 2 (4 pts) (nesting depth added 11). To reduce it, split the body into named stages: move each independent step or branch into its own named function so the body reads as a short sequence of named calls rather than one long body.
Trie.parse_longest (cognitive 17) plugins/hanlp_trie/hanlp_trie/trie.py:123— Trie.parse_longest has cognitive complexity 17 (threshold 15). Drivers by points: if/else 4 (13 pts), loops 2 (4 pts) (nesting depth added 11). To reduce it, split the body into named stages: move each independent step or branch into its own named function so the body reads as a short sequence of named calls rather than one long body.
TransformerTextTransform.inputs_to_samples (cognitive 16) hanlp/components/classifiers/transformer_classifier_tf.py:24— TransformerTextTransform.inputs_to_samples has cognitive complexity 16 (threshold 15). Drivers by points: if/else 5 (10 pts), ternaries 3 (5 pts), loops 1 (nesting depth added 7). To reduce it, split the body into named stages: move each independent step or branch into its own named function so the body reads as a short sequence of named calls rather than one long body.
MultiTaskLearning.build_optimizer (cognitive 16) hanlp/components/mtl/multi_task_learning.py:208— MultiTaskLearning.build_optimizer has cognitive complexity 16 (threshold 15). Drivers by points: if/else 5 (9 pts), loops 5 (6 pts), boolean chains 1 (nesting depth added 5). To reduce it, split the body into named stages: move each independent step or branch into its own named function so the body reads as a short sequence of named calls rather than one long body.
chu_liu_edmonds._find_cycle (cognitive 16) hanlp/components/parsers/chu_liu_edmonds.py:274— chu_liu_edmonds._find_cycle has cognitive complexity 16 (threshold 15). Drivers by points: if/else 4 (9 pts), loops 3 (6 pts), boolean chains 1 (nesting depth added 8). To reduce it, split the body into named stages: move each independent step or branch into its own named function so the body reads as a short sequence of named calls rather than one long body.
parse_alg.tarjan (cognitive 16) hanlp/components/parsers/parse_alg.py:123— parse_alg.tarjan has cognitive complexity 16 (threshold 15). Drivers by points: if/else 4 (10 pts), loops 2 (6 pts) (nesting depth added 10). To reduce it, split the body into named stages: move each independent step or branch into its own named function so the body reads as a short sequence of named calls rather than one long body.
parse_alg.add_secondary_arcs_by_preds (cognitive 16) hanlp/components/parsers/parse_alg.py:275— parse_alg.add_secondary_arcs_by_preds has cognitive complexity 16 (threshold 15). Drivers by points: if/else 3 (7 pts), loops 4 (5 pts), boolean chains 2, error handling 1 (2 pts) (nesting depth added 6). To reduce it, split the body into named stages: move each independent step or branch into its own named function so the body reads as a short sequence of named calls rather than one long body.
udify_util.get_ud_treebank_files (cognitive 16) hanlp/components/parsers/ud/udify_util.py:31— udify_util.get_ud_treebank_files has cognitive complexity 16 (threshold 15). Drivers by points: if/else 4 (8 pts), ternaries 4 (7 pts), loops 1 (nesting depth added 7). To reduce it, split the body into named stages: move each independent step or branch into its own named function so the body reads as a short sequence of named calls rather than one long body.
srl_eval_utils.compute_span_f1 (cognitive 16) hanlp/components/srl/span_rank/srl_eval_utils.py:104— srl_eval_utils.compute_span_f1 has cognitive complexity 16 (threshold 15). Drivers by points: if/else 2 (9 pts), loops 3 (6 pts), boolean chains 1 (nesting depth added 10). To reduce it, split the body into named stages: move each independent step or branch into its own named function so the body reads as a short sequence of named calls rather than one long body. This shape REPEATS in the file: one other method here (srl_eval_utils.compute_unlabeled_span_f1) has the same decision points, in the same order, at the same nesting depths — so this is one pattern written twice rather than two separate problems. Splitting this body alone leaves the other exactly as it is. Where these are variations on one operation, the change that clears both is the shared one: lift the common shape into a single routine the variants call, parameterised by whatever genuinely differs between them, and keep in each method only the part that is not shared.
srl_eval_utils.compute_unlabeled_span_f1 (cognitive 16) hanlp/components/srl/span_rank/srl_eval_utils.py:130— srl_eval_utils.compute_unlabeled_span_f1 has cognitive complexity 16 (threshold 15). Drivers by points: if/else 2 (9 pts), loops 3 (6 pts), boolean chains 1 (nesting depth added 10). To reduce it, split the body into named stages: move each independent step or branch into its own named function so the body reads as a short sequence of named calls rather than one long body. This shape REPEATS in the file: one other method here (srl_eval_utils.compute_span_f1) has the same decision points, in the same order, at the same nesting depths — so this is one pattern written twice rather than two separate problems. Splitting this body alone leaves the other exactly as it is. Where these are variations on one operation, the change that clears both is the shared one: lift the common shape into a single routine the variants call, parameterised by whatever genuinely differs between them, and keep in each method only the part that is not shared.
json_ner.prune_ner_tagset (cognitive 16) hanlp/datasets/ner/loaders/json_ner.py:141— json_ner.prune_ner_tagset has cognitive complexity 16 (threshold 15). Drivers by points: if/else 5 (14 pts), loops 1 (2 pts) (nesting depth added 10). To reduce it, flatten the nesting: this score is depth rather than breadth — most of its points come from checks stacked inside one another, so the work sits several levels in. Invert each enclosing check into an early exit (a return, or the language's equivalent) so the happy path stays at one level, and where a level cannot be exited early, lift the block it encloses into its own named function.
ontonotes_loader._canonicalize_clusters (cognitive 16) hanlp/datasets/srl/loaders/ontonotes_loader.py:586— ontonotes_loader._canonicalize_clusters has cognitive complexity 16 (threshold 15). Drivers by points: if/else 4 (10 pts), loops 3 (6 pts) (nesting depth added 9). To reduce it, split the body into named stages: move each independent step or branch into its own named function so the body reads as a short sequence of named calls rather than one long body.
DocumentState.finalize (cognitive 16) hanlp/datasets/srl/ontonotes5/_utils.py:112— DocumentState.finalize has cognitive complexity 16 (threshold 15). Drivers by points: if/else 4 (10 pts), loops 3 (6 pts) (nesting depth added 9). To reduce it, split the body into named stages: move each independent step or branch into its own named function so the body reads as a short sequence of named calls rather than one long body.
ChunkingDataset._generate_chars_tags (cognitive 16) hanlp/datasets/tokenization/loaders/chunking_dataset.py:32— ChunkingDataset._generate_chars_tags has cognitive complexity 16 (threshold 15). Drivers by points: if/else 4 (10 pts), loops 2 (4 pts), boolean chains 2 (nesting depth added 8). To reduce it, split the body into named stages: move each independent step or branch into its own named function so the body reads as a short sequence of named calls rather than one long body.
sequence_labeling.end_of_chunk (cognitive 16) hanlp/metrics/chunking/sequence_labeling.py:83— sequence_labeling.end_of_chunk has cognitive complexity 16 (threshold 15). Drivers by points: if/else 9, boolean chains 7. To reduce it, split the body: this score is breadth rather than depth — many checks laid out side by side rather than nested inside one another, so inverting conditions into early returns has nothing left to flatten. Group the statements between the checks into named steps and move each step into its own function, so the body reads as a short sequence of named stages. This shape REPEATS in the file: one other method here (sequence_labeling.start_of_chunk) has the same decision points, in the same order, at the same nesting depths — so this is one pattern written twice rather than two separate problems. Splitting this body alone leaves the other exactly as it is. Where these are variations on one operation, the change that clears both is the shared one: lift the common shape into a single routine the variants call, parameterised by whatever genuinely differs between them, and keep in each method only the part that is not shared.
sequence_labeling.start_of_chunk (cognitive 16) hanlp/metrics/chunking/sequence_labeling.py:114— sequence_labeling.start_of_chunk has cognitive complexity 16 (threshold 15). Drivers by points: if/else 9, boolean chains 7. To reduce it, split the body: this score is breadth rather than depth — many checks laid out side by side rather than nested inside one another, so inverting conditions into early returns has nothing left to flatten. Group the statements between the checks into named steps and move each step into its own function, so the body reads as a short sequence of named stages. This shape REPEATS in the file: one other method here (sequence_labeling.end_of_chunk) has the same decision points, in the same order, at the same nesting depths — so this is one pattern written twice rather than two separate problems. Splitting this body alone leaves the other exactly as it is. Where these are variations on one operation, the change that clears both is the shared one: lift the common shape into a single routine the variants call, parameterised by whatever genuinely differs between them, and keep in each method only the part that is not shared.
GraphRepair.remove_unknown_nodes (cognitive 16) plugins/hanlp_common/hanlp_common/amr.py:890— GraphRepair.remove_unknown_nodes has cognitive complexity 16 (threshold 15). Drivers by points: if/else 3 (8 pts), loops 3 (7 pts), boolean chains 1 (nesting depth added 9). To reduce it, split the body into named stages: move each independent step or branch into its own named function so the body reads as a short sequence of named calls rather than one long body.
util.infer_space_after (cognitive 16) plugins/hanlp_common/hanlp_common/util.py:82— util.infer_space_after has cognitive complexity 16 (threshold 15). Drivers by points: if/else 6 (15 pts), loops 1 (nesting depth added 9). To reduce it, split the body into named stages: move each independent step or branch into its own named function so the body reads as a short sequence of named calls rather than one long body.
Change-coupling hub: version.py → transformer.py, transformer_tokenizer.py, REDACTED hanlp/version.py— `hanlp/version.py` changes together with 3 other files — `hanlp/components/tokenizers/transformer.py`, `hanlp/transform/transformer_tokenizer.py`, `REDACTED` — none of which declares a dependency on it: one file is the hub of 3 separate couplings, not 3 unrelated pairs. Read the hub first: if the others each duplicate a part of what it does, the shared concern belongs in ONE unit and extracting it clears every edge at once; if the hub is a registry, dispatcher or barrel that must name each of them, the coupling is structural and the question is whether that list can be discovered instead of enumerated. Fixing the hub is one change; breaking the couplings one pair at a time is 3.
D4 · Code Duplication· Near-duplicate member family (3 members, 268 shared lines) · ×1
Near-duplicate member family (3 members, 268 shared lines) hanlp/components/amr/amrbart/model_interface/tokenization_bart.py:176— hanlp/components/amr/amrbart/model_interface/tokenization_bart.py:176-448 | hanlp/components/amr/seq2seq/dataset/tokenization_bart.py:356-633 | hanlp/components/amr/seq2seq/dataset/tokenization_t5.py:349-628 — These 3 members are variants of one another: a block of 268 lines reported below appears in every one of them, and the pairwise near-duplicate rows they would otherwise produce are collapsed into this row. Read them as one construct written 3 times. The repair is at the members' grain — factor the shared implementation out once and have all of them call it with their differences as parameters or as an injected step, or, where the difference is systematic, generate them from one template. Extracting the individual blocks below is not the same fix: it leaves every body in place and the next edit still has to be made 3 times.
D4 · Code Duplication· Near-duplicate member family (3 members, 17 shared lines) · ×1
Near-duplicate member family (3 members, 17 shared lines) REDACTED:279— REDACTED:279-311 | hanlp/components/amr/seq2seq/dataset/linearization.py:145-168 | REDACTED:407-439 — These 3 members are variants of one another: a block of 17 lines reported below appears in every one of them, and the pairwise near-duplicate rows they would otherwise produce are collapsed into this row. Read them as one construct written 3 times. The repair is at the members' grain — factor the shared implementation out once and have all of them call it with their differences as parameters or as an injected step, or, where the difference is systematic, generate them from one template. Extracting the individual blocks below is not the same fix: it leaves every body in place and the next edit still has to be made 3 times.
Near-duplicate member pair (113 shared lines) REDACTED:32— REDACTED:32-188 | REDACTED:192-357 — These two members are variants of one another: 113 of their lines are already reported as duplicated blocks below, spread through both bodies rather than gathered into one. Read them as a single construct written twice. The repair is at the members' grain — factor the shared pipeline into one implementation the two call with their differences as parameters or as an injected step, or, where the difference is systematic (sync against async, one transport against another), generate one from the other. Extracting the individual blocks below is not the same fix: it leaves the two bodies in place and the next edit still has to be made twice.
Near-duplicate member pair (112 shared lines) REDACTED:315— REDACTED:315-437 | REDACTED:443-560 — These two members are variants of one another: 112 of their lines are already reported as duplicated blocks below, spread through both bodies rather than gathered into one. Read them as a single construct written twice. The repair is at the members' grain — factor the shared pipeline into one implementation the two call with their differences as parameters or as an injected step, or, where the difference is systematic (sync against async, one transport against another), generate one from the other. Extracting the individual blocks below is not the same fix: it leaves the two bodies in place and the next edit still has to be made twice.
Near-duplicate member pair (64 shared lines) hanlp/components/parsers/alg_tf.py:175— hanlp/components/parsers/alg_tf.py:175-256 | hanlp/components/parsers/biaffine_tf/alg.py:175-256 — These two members are variants of one another: 64 of their lines are already reported as duplicated blocks below, spread through both bodies rather than gathered into one. Read them as a single construct written twice. The repair is at the members' grain — factor the shared pipeline into one implementation the two call with their differences as parameters or as an injected step, or, where the difference is systematic (sync against async, one transport against another), generate one from the other. Extracting the individual blocks below is not the same fix: it leaves the two bodies in place and the next edit still has to be made twice.
Near-duplicate member pair (36 shared lines) hanlp/components/parsers/alg_tf.py:34— hanlp/components/parsers/alg_tf.py:34-87 | hanlp/components/parsers/biaffine_tf/alg.py:34-87 — These two members are variants of one another: 36 of their lines are already reported as duplicated blocks below, spread through both bodies rather than gathered into one. Read them as a single construct written twice. The repair is at the members' grain — factor the shared pipeline into one implementation the two call with their differences as parameters or as an injected step, or, where the difference is systematic (sync against async, one transport against another), generate one from the other. Extracting the individual blocks below is not the same fix: it leaves the two bodies in place and the next edit still has to be made twice.
D4 · Code Duplication· Edited copy of a member (14 corresponding lines) · ×1
Edited copy of a member (14 corresponding lines) hanlp/components/amr/seq2seq/dataset/tokenization_bart.py:266— hanlp/components/amr/seq2seq/dataset/tokenization_bart.py:266-280 | hanlp/components/amr/seq2seq/dataset/tokenization_t5.py:257-273 — These two members are one piece of code written twice and then edited apart: 14 consecutive lines correspond almost exactly, broken only by small local edits. Most of that correspondence is NOT reported as duplicated blocks below — the edits cut it into fragments and only the largest of them clear the block floor, so the rows below understate it. The repair is at the members' grain — factor the shared implementation into one the two call with their differences as parameters or as an injected step, or, where the difference is systematic (an extra return value, one transport against another), generate one from the other. Left alone, the next edit has to be made twice and the two will drift further apart.
Duplicated block (209 lines × 3) hanlp/components/amr/amrbart/model_interface/tokenization_bart.py:211— hanlp/components/amr/amrbart/model_interface/tokenization_bart.py:211-419 | hanlp/components/amr/seq2seq/dataset/tokenization_bart.py:392-600 | hanlp/components/amr/seq2seq/dataset/tokenization_t5.py:387-595 — before extracting anything, compare `hanlp/components/amr/amrbart/model_interface/tokenization_bart.py` and `hanlp/components/amr/seq2seq/dataset/tokenization_bart.py` as WHOLE FILES: this scan already matched 10 separate duplicated blocks between them, totalling at least 329 lines, which is the signature of one file having been copied from the other rather than of a helper waiting to be extracted. The two sit in different directories, so one cannot simply be deleted in favour of the other while both are reached separately: hoist the shared part into a location both already depend on and have each file call it, and retire whichever file turns out to have no caller of its own left. Extracting one helper per block leaves the fork in place. Read the line range as the matched WINDOW rather than a finished unit: at `hanlp/components/amr/amrbart/model_interface/tokenization_bart.py:211` it begins part-way through the construct above it, so those exact lines cannot be lifted as they stand — widen the region to the smallest complete statement or declaration that contains it, and extract that.
Duplicated block (83–90 lines × 3) REDACTED:130— REDACTED:130-219 | REDACTED:95-177 | REDACTED:264-346 — before extracting anything, compare `REDACTED` and `REDACTED` as WHOLE FILES: this scan already matched 15 separate duplicated blocks between them, totalling at least 404 lines, which is the signature of one file having been copied from the other rather than of a helper waiting to be extracted. The two sit in different directories, so one cannot simply be deleted in favour of the other while both are reached separately: hoist the shared part into a location both already depend on and have each file call it, and retire whichever file turns out to have no caller of its own left. Extracting one helper per block leaves the fork in place. The matched lines also transfer control out of the body holding them, which cannot survive a move into a called unit unchanged: have the extracted unit return that decision and let each site act on it.
Duplicated block (82–84 lines × 2) REDACTED:315— REDACTED:315-396 | REDACTED:443-526 — before extracting anything, compare `REDACTED` and `REDACTED` as WHOLE FILES: this scan already matched 15 separate duplicated blocks between them, totalling at least 404 lines, which is the signature of one file having been copied from the other rather than of a helper waiting to be extracted. The two sit in different directories, so one cannot simply be deleted in favour of the other while both are reached separately: hoist the shared part into a location both already depend on and have each file call it, and retire whichever file turns out to have no caller of its own left. Extracting one helper per block leaves the fork in place. Read the line range as the matched WINDOW rather than a finished unit: at `REDACTED:315` it does not close everything it opens, so those exact lines cannot be lifted as they stand — widen the region to the smallest complete statement or declaration that contains it, and extract that. The matched lines also transfer control out of the body holding them, which cannot survive a move into a called unit unchanged: have the extracted unit return that decision and let each site act on it. Note that the copies do not run to the end of the range shown: their LAST lines are different code, not the same code under different names — the matched region ends inside that line. Extract the lines above it, and read the last line of each site separately.
Duplicated block (69 lines × 2) hanlp/layers/transformers/utils_tf.py:82— hanlp/layers/transformers/utils_tf.py:82-150 | hanlp/transform/transformer_tokenizer.py:627-695 — the copies span different directories, so extracting a shared function means choosing where it lives: put it somewhere both call sites can already reach — a location they all depend on today, or a new shared one if there is none — and call it from each site; until then, every change has to be made twice. Read the line range as the matched WINDOW rather than a finished unit: at `hanlp/layers/transformers/utils_tf.py:82` it begins part-way through the construct above it, so those exact lines cannot be lifted as they stand — widen the region to the smallest complete statement or declaration that contains it, and extract that. The `return` at the foot of the matched lines is the enclosing body's own terminal exit, not an early one: it moves with them unchanged, and each site calls the extracted unit from the position that `return` occupied — no decision has to be handed back and re-acted on.
Duplicated block (47 lines × 2) hanlp/layers/transformers/utils_tf.py:33— hanlp/layers/transformers/utils_tf.py:33-79 | hanlp/transform/transformer_tokenizer.py:578-624 — the copies span different directories, so extracting a shared function means choosing where it lives: put it somewhere both call sites can already reach — a location they all depend on today, or a new shared one if there is none — and call it from each site; until then, every change has to be made twice.
Duplicated block (42 lines × 2) hanlp/utils/span_util.py:468— hanlp/utils/span_util.py:468-509 | hanlp/utils/span_util.py:515-556 — both copies are in the same file, so extract the block into one function there and call it from each site — the copies drift apart the first time only one of them is edited.
Duplicated block (35 lines × 2) REDACTED:78— REDACTED:78-112 | REDACTED:47-81 — before extracting anything, compare `REDACTED` and `REDACTED` as WHOLE FILES: this scan already matched 15 separate duplicated blocks between them, totalling at least 404 lines, which is the signature of one file having been copied from the other rather than of a helper waiting to be extracted. The two sit in different directories, so one cannot simply be deleted in favour of the other while both are reached separately: hoist the shared part into a location both already depend on and have each file call it, and retire whichever file turns out to have no caller of its own left. Extracting one helper per block leaves the fork in place.
Duplicated block (33 lines × 3) hanlp/components/parsers/alg_tf.py:122— hanlp/components/parsers/alg_tf.py:122-154 | hanlp/components/parsers/biaffine_tf/alg.py:122-154 | hanlp/components/parsers/parse_alg.py:42-74 — before extracting anything, compare `hanlp/components/parsers/alg_tf.py` and `hanlp/components/parsers/biaffine_tf/alg.py` as WHOLE FILES: this scan already matched 9 separate duplicated blocks between them, totalling at least 185 lines, which is the signature of one file having been copied from the other rather than of a helper waiting to be extracted. The two sit in different directories, so one cannot simply be deleted in favour of the other while both are reached separately: hoist the shared part into a location both already depend on and have each file call it, and retire whichever file turns out to have no caller of its own left. Extracting one helper per block leaves the fork in place. The `return` at the foot of the matched lines is the enclosing body's own terminal exit, not an early one: it moves with them unchanged, and each site calls the extracted unit from the position that `return` occupied — no decision has to be handed back and re-acted on.
Duplicated block (27–31 lines × 3) hanlp/components/amr/amrbart/model_interface/tokenization_bart.py:422— hanlp/components/amr/amrbart/model_interface/tokenization_bart.py:422-448 | hanlp/components/amr/seq2seq/dataset/tokenization_bart.py:603-633 | hanlp/components/amr/seq2seq/dataset/tokenization_t5.py:598-628 — before extracting anything, compare `hanlp/components/amr/amrbart/model_interface/tokenization_bart.py` and `hanlp/components/amr/seq2seq/dataset/tokenization_bart.py` as WHOLE FILES: this scan already matched 10 separate duplicated blocks between them, totalling at least 329 lines, which is the signature of one file having been copied from the other rather than of a helper waiting to be extracted. The two sit in different directories, so one cannot simply be deleted in favour of the other while both are reached separately: hoist the shared part into a location both already depend on and have each file call it, and retire whichever file turns out to have no caller of its own left. Extracting one helper per block leaves the fork in place. Read the line range as the matched WINDOW rather than a finished unit: at `hanlp/components/amr/amrbart/model_interface/tokenization_bart.py:422` it begins part-way through the construct above it, so those exact lines cannot be lifted as they stand — widen the region to the smallest complete statement or declaration that contains it, and extract that.
Duplicated block (30 lines × 3) hanlp/components/parsers/alg_tf.py:190— hanlp/components/parsers/alg_tf.py:190-219 | hanlp/components/parsers/biaffine_tf/alg.py:190-219 | hanlp/components/parsers/parse_alg.py:188-217 — before extracting anything, compare `hanlp/components/parsers/alg_tf.py` and `hanlp/components/parsers/biaffine_tf/alg.py` as WHOLE FILES: this scan already matched 9 separate duplicated blocks between them, totalling at least 185 lines, which is the signature of one file having been copied from the other rather than of a helper waiting to be extracted. The two sit in different directories, so one cannot simply be deleted in favour of the other while both are reached separately: hoist the shared part into a location both already depend on and have each file call it, and retire whichever file turns out to have no caller of its own left. Extracting one helper per block leaves the fork in place.
Duplicated block (30 lines × 2) hanlp/components/parsers/alg_tf.py:260— hanlp/components/parsers/alg_tf.py:260-289 | hanlp/components/parsers/biaffine_tf/alg.py:260-289 — before extracting anything, compare `hanlp/components/parsers/alg_tf.py` and `hanlp/components/parsers/biaffine_tf/alg.py` as WHOLE FILES: this scan already matched 9 separate duplicated blocks between them, totalling at least 185 lines, which is the signature of one file having been copied from the other rather than of a helper waiting to be extracted. The two sit in different directories, so one cannot simply be deleted in favour of the other while both are reached separately: hoist the shared part into a location both already depend on and have each file call it, and retire whichever file turns out to have no caller of its own left. Extracting one helper per block leaves the fork in place. The matched lines also transfer control out of the body holding them, which cannot survive a move into a called unit unchanged: have the extracted unit return that decision and let each site act on it.
Duplicated block (28 lines × 2) hanlp/metrics/chunking/sequence_labeling.py:84— hanlp/metrics/chunking/sequence_labeling.py:84-111 | hanlp/metrics/chunking/sequence_labeling.py:115-142 — both copies are in the same file, so extract the block into one function there and call it from each site — the copies drift apart the first time only one of them is edited. The `return` at the foot of the matched lines is the enclosing body's own terminal exit, not an early one: it moves with them unchanged, and each site calls the extracted unit from the position that `return` occupied — no decision has to be handed back and re-acted on.
Duplicated block (26 lines × 2) REDACTED:489— REDACTED:489-514 | REDACTED:611-636 — before extracting anything, compare `REDACTED` and `REDACTED` as WHOLE FILES: this scan already matched 15 separate duplicated blocks between them, totalling at least 404 lines, which is the signature of one file having been copied from the other rather than of a helper waiting to be extracted. The two sit in different directories, so one cannot simply be deleted in favour of the other while both are reached separately: hoist the shared part into a location both already depend on and have each file call it, and retire whichever file turns out to have no caller of its own left. Extracting one helper per block leaves the fork in place. The `return` at the foot of the matched lines is the enclosing body's own terminal exit, not an early one: it moves with them unchanged, and each site calls the extracted unit from the position that `return` occupied — no decision has to be handed back and re-acted on.
Duplicated block (25 lines × 3) hanlp/components/amr/amrbart/model_interface/tokenization_bart.py:451— hanlp/components/amr/amrbart/model_interface/tokenization_bart.py:451-475 | hanlp/components/amr/seq2seq/dataset/tokenization_bart.py:328-352 | hanlp/components/amr/seq2seq/dataset/tokenization_t5.py:321-345 — before extracting anything, compare `hanlp/components/amr/amrbart/model_interface/tokenization_bart.py` and `hanlp/components/amr/seq2seq/dataset/tokenization_bart.py` as WHOLE FILES: this scan already matched 10 separate duplicated blocks between them, totalling at least 329 lines, which is the signature of one file having been copied from the other rather than of a helper waiting to be extracted. The two sit in different directories, so one cannot simply be deleted in favour of the other while both are reached separately: hoist the shared part into a location both already depend on and have each file call it, and retire whichever file turns out to have no caller of its own left. Extracting one helper per block leaves the fork in place. The matched lines also transfer control out of the body holding them, which cannot survive a move into a called unit unchanged: have the extracted unit return that decision and let each site act on it.
Duplicated block (24–25 lines × 2) REDACTED:413— REDACTED:413-437 | REDACTED:537-560 — before extracting anything, compare `REDACTED` and `REDACTED` as WHOLE FILES: this scan already matched 15 separate duplicated blocks between them, totalling at least 404 lines, which is the signature of one file having been copied from the other rather than of a helper waiting to be extracted. The two sit in different directories, so one cannot simply be deleted in favour of the other while both are reached separately: hoist the shared part into a location both already depend on and have each file call it, and retire whichever file turns out to have no caller of its own left. Extracting one helper per block leaves the fork in place. Read the line range as the matched WINDOW rather than a finished unit: at `REDACTED:413` it begins part-way through the construct above it, so those exact lines cannot be lifted as they stand — widen the region to the smallest complete statement or declaration that contains it, and extract that. The `return` at the foot of the matched lines is the enclosing body's own terminal exit, not an early one: it moves with them unchanged, and each site calls the extracted unit from the position that `return` occupied — no decision has to be handed back and re-acted on.
Duplicated block (24 lines × 3) hanlp/components/parsers/alg_tf.py:95— hanlp/components/parsers/alg_tf.py:95-118 | hanlp/components/parsers/biaffine_tf/alg.py:95-118 | hanlp/components/parsers/parse_alg.py:15-38 — before extracting anything, compare `hanlp/components/parsers/alg_tf.py` and `hanlp/components/parsers/biaffine_tf/alg.py` as WHOLE FILES: this scan already matched 9 separate duplicated blocks between them, totalling at least 185 lines, which is the signature of one file having been copied from the other rather than of a helper waiting to be extracted. The two sit in different directories, so one cannot simply be deleted in favour of the other while both are reached separately: hoist the shared part into a location both already depend on and have each file call it, and retire whichever file turns out to have no caller of its own left. Extracting one helper per block leaves the fork in place.
Duplicated block (21–24 lines × 2) REDACTED:254— REDACTED:254-274 | REDACTED:379-402 — before extracting anything, compare `REDACTED` and `REDACTED` as WHOLE FILES: this scan already matched 15 separate duplicated blocks between them, totalling at least 404 lines, which is the signature of one file having been copied from the other rather than of a helper waiting to be extracted. The two sit in different directories, so one cannot simply be deleted in favour of the other while both are reached separately: hoist the shared part into a location both already depend on and have each file call it, and retire whichever file turns out to have no caller of its own left. Extracting one helper per block leaves the fork in place.
Duplicated block (22 lines × 3) hanlp/components/amr/amrbart/model_interface/tokenization_bart.py:188— hanlp/components/amr/amrbart/model_interface/tokenization_bart.py:188-209 | hanlp/components/amr/seq2seq/dataset/tokenization_bart.py:369-390 | hanlp/components/amr/seq2seq/dataset/tokenization_t5.py:364-385 — before extracting anything, compare `hanlp/components/amr/amrbart/model_interface/tokenization_bart.py` and `hanlp/components/amr/seq2seq/dataset/tokenization_bart.py` as WHOLE FILES: this scan already matched 10 separate duplicated blocks between them, totalling at least 329 lines, which is the signature of one file having been copied from the other rather than of a helper waiting to be extracted. The two sit in different directories, so one cannot simply be deleted in favour of the other while both are reached separately: hoist the shared part into a location both already depend on and have each file call it, and retire whichever file turns out to have no caller of its own left. Extracting one helper per block leaves the fork in place. ★ These copies have DRIFTED, and that is worth reading before extracting anything: just before the matched lines, `hanlp/components/amr/seq2seq/dataset/tokenization_t5.py:360` calls `Graph` and `hanlp/components/amr/seq2seq/dataset/tokenization_bart.py:365` does not — after which the two agree again for 2 more lines. One of those two behaviours is the intended one and the other is what a copy-paste left behind, so decide which BEFORE unifying them: extracting the shared part will silently settle it, and if the copy that skips the call is the wrong one, that bug is already live.
Duplicated block (21 lines × 2) REDACTED:291— REDACTED:291-311 | REDACTED:419-439 — before extracting anything, compare `REDACTED` and `REDACTED` as WHOLE FILES: this scan already matched 15 separate duplicated blocks between them, totalling at least 404 lines, which is the signature of one file having been copied from the other rather than of a helper waiting to be extracted. The two sit in different directories, so one cannot simply be deleted in favour of the other while both are reached separately: hoist the shared part into a location both already depend on and have each file call it, and retire whichever file turns out to have no caller of its own left. Extracting one helper per block leaves the fork in place.
Duplicated block (2–21 lines × 4) hanlp/components/parsers/alg_tf.py:34— hanlp/components/parsers/alg_tf.py:34-54 | hanlp/components/parsers/alg_tf.py:76-77 | hanlp/components/parsers/biaffine_tf/alg.py:34-54 | hanlp/components/parsers/biaffine_tf/alg.py:76-77 — before extracting anything, compare `hanlp/components/parsers/alg_tf.py` and `hanlp/components/parsers/biaffine_tf/alg.py` as WHOLE FILES: this scan already matched 9 separate duplicated blocks between them, totalling at least 185 lines, which is the signature of one file having been copied from the other rather than of a helper waiting to be extracted. The two sit in different directories, so one cannot simply be deleted in favour of the other while both are reached separately: hoist the shared part into a location both already depend on and have each file call it, and retire whichever file turns out to have no caller of its own left. Extracting one helper per block leaves the fork in place. Note that the copies do not run to the end of the range shown: their LAST lines are different code, not the same code under different names — the matched region ends inside that line. Extract the lines above it, and read the last line of each site separately. ★ These copies have DRIFTED, and that is worth reading before extracting anything: just after the matched lines, `hanlp/components/parsers/alg_tf.py:54` calls `abs`, `expand_dims` and `hanlp/components/parsers/alg_tf.py:77` does not — after which the two agree again for 2 more lines. One of those two behaviours is the intended one and the other is what a copy-paste left behind, so decide which BEFORE unifying them: extracting the shared part will silently settle it, and if the copy that skips the call is the wrong one, that bug is already live.
Duplicated block (20 lines × 3) hanlp/components/amr/amrbart/preprocess/read_and_process.py:57— hanlp/components/amr/amrbart/preprocess/read_and_process.py:57-76 | hanlp/components/amr/seq2seq/dataset/tokenization_bart.py:302-321 | hanlp/components/amr/seq2seq/dataset/tokenization_t5.py:295-314 — before extracting anything, compare `hanlp/components/amr/seq2seq/dataset/tokenization_bart.py` and `hanlp/components/amr/seq2seq/dataset/tokenization_t5.py` as WHOLE FILES: this scan already matched 21 separate duplicated blocks between them, totalling at least 476 lines, which is the signature of one file having been copied from the other rather than of a helper waiting to be extracted. If that is what happened, the fix is to keep one copy and have the other call it (or delete it), which resolves this row and its siblings together — extracting one helper per block leaves the fork in place.
Duplicated block (16–20 lines × 2) hanlp/components/parsers/ud/udify_util.py:77— hanlp/components/parsers/ud/udify_util.py:77-92 | hanlp/components/parsers/ud/udify_util.py:237-256 — both copies are in the same file, so extract the block into one function there and call it from each site — the copies drift apart the first time only one of them is edited. Note that the copies do not run to the end of the range shown: their LAST lines are different code, not the same code under different names — the matched region ends inside that line. Extract the lines above it, and read the last line of each site separately.
Duplicated block (19–20 lines × 2) hanlp/transform/conll_tf.py:147— hanlp/transform/conll_tf.py:147-165 | hanlp/transform/conll_tf.py:581-600 — both copies are in the same file, so extract the block into one function there and call it from each site — the copies drift apart the first time only one of them is edited.
Duplicated block (19 lines × 2) REDACTED:32— REDACTED:32-50 | REDACTED:192-210 — both copies are in the same file, so extract the block into one function there and call it from each site — the copies drift apart the first time only one of them is edited.
Duplicated block (18 lines × 2) REDACTED:457— REDACTED:457-474 | REDACTED:579-596 — before extracting anything, compare `REDACTED` and `REDACTED` as WHOLE FILES: this scan already matched 15 separate duplicated blocks between them, totalling at least 404 lines, which is the signature of one file having been copied from the other rather than of a helper waiting to be extracted. The two sit in different directories, so one cannot simply be deleted in favour of the other while both are reached separately: hoist the shared part into a location both already depend on and have each file call it, and retire whichever file turns out to have no caller of its own left. Extracting one helper per block leaves the fork in place. The matched lines also transfer control out of the body holding them, which cannot survive a move into a called unit unchanged: have the extracted unit return that decision and let each site act on it. Note that the copies do not run to the end of the range shown: their LAST lines are different code, not the same code under different names — the matched region ends inside that line. Extract the lines above it, and read the last line of each site separately.
Duplicated block (15–17 lines × 2) REDACTED:234— REDACTED:234-250 | REDACTED:361-375 — before extracting anything, compare `REDACTED` and `REDACTED` as WHOLE FILES: this scan already matched 15 separate duplicated blocks between them, totalling at least 404 lines, which is the signature of one file having been copied from the other rather than of a helper waiting to be extracted. The two sit in different directories, so one cannot simply be deleted in favour of the other while both are reached separately: hoist the shared part into a location both already depend on and have each file call it, and retire whichever file turns out to have no caller of its own left. Extracting one helper per block leaves the fork in place.
Duplicated block (16–17 lines × 2) hanlp/components/amr/amrbart/model_interface/modeling_bart.py:1497— hanlp/components/amr/amrbart/model_interface/modeling_bart.py:1497-1512 | hanlp/components/amr/amrbart/model_interface/modeling_bart.py:1624-1640 — both copies are in the same file, so extract the block into one function there and call it from each site — the copies drift apart the first time only one of them is edited. Read the line range as the matched WINDOW rather than a finished unit: at `hanlp/components/amr/amrbart/model_interface/modeling_bart.py:1497` it begins part-way through the construct above it, so those exact lines cannot be lifted as they stand — widen the region to the smallest complete statement or declaration that contains it, and extract that.
Duplicated block (15 lines × 3) hanlp/components/amr/amrbart/model_interface/modeling_bart.py:1353— hanlp/components/amr/amrbart/model_interface/modeling_bart.py:1353-1367 | hanlp/components/amr/amrbart/model_interface/modeling_bart.py:1496-1510 | hanlp/components/amr/amrbart/model_interface/modeling_bart.py:1623-1637 — all 3 copies are in the same file, so extract the block into one function there and call it from every one of those sites — resolving only two of them leaves the rest to drift apart the first time one is edited. Read the line range as the matched WINDOW rather than a finished unit: at `hanlp/components/amr/amrbart/model_interface/modeling_bart.py:1353` it does not close everything it opens, so those exact lines cannot be lifted as they stand — widen the region to the smallest complete statement or declaration that contains it, and extract that.
Duplicated block (14 lines × 4) hanlp/components/amr/seq2seq/dataset/tokenization_bart.py:236— hanlp/components/amr/seq2seq/dataset/tokenization_bart.py:236-249 | hanlp/components/amr/seq2seq/dataset/tokenization_bart.py:653-666 | hanlp/components/amr/seq2seq/dataset/tokenization_t5.py:227-240 | hanlp/components/amr/seq2seq/dataset/tokenization_t5.py:650-663 — before extracting anything, compare `hanlp/components/amr/seq2seq/dataset/tokenization_bart.py` and `hanlp/components/amr/seq2seq/dataset/tokenization_t5.py` as WHOLE FILES: this scan already matched 21 separate duplicated blocks between them, totalling at least 476 lines, which is the signature of one file having been copied from the other rather than of a helper waiting to be extracted. If that is what happened, the fix is to keep one copy and have the other call it (or delete it), which resolves this row and its siblings together — extracting one helper per block leaves the fork in place. The matched lines also transfer control out of the body holding them, which cannot survive a move into a called unit unchanged: have the extracted unit return that decision and let each site act on it.
Duplicated block (9–14 lines × 2) REDACTED:51— REDACTED:51-64 | REDACTED:27-35 — the copies span different directories, so extracting a shared function means choosing where it lives: put it somewhere both call sites can already reach — a location they all depend on today, or a new shared one if there is none — and call it from each site; until then, every change has to be made twice. The `return` at the foot of the matched lines is the enclosing body's own terminal exit, not an early one: it moves with them unchanged, and each site calls the extracted unit from the position that `return` occupied — no decision has to be handed back and re-acted on.
Duplicated block (14 lines × 2) hanlp/components/amr/seq2seq/dataset/tokenization_bart.py:206— hanlp/components/amr/seq2seq/dataset/tokenization_bart.py:206-219 | hanlp/components/amr/seq2seq/dataset/tokenization_t5.py:196-209 — before extracting anything, compare `hanlp/components/amr/seq2seq/dataset/tokenization_bart.py` and `hanlp/components/amr/seq2seq/dataset/tokenization_t5.py` as WHOLE FILES: this scan already matched 21 separate duplicated blocks between them, totalling at least 476 lines, which is the signature of one file having been copied from the other rather than of a helper waiting to be extracted. If that is what happened, the fix is to keep one copy and have the other call it (or delete it), which resolves this row and its siblings together — extracting one helper per block leaves the fork in place.
Duplicated block (13 lines × 4) hanlp/components/parsers/constituency/crf_constituency_parser.py:63— hanlp/components/parsers/constituency/crf_constituency_parser.py:63-75 | hanlp/components/parsers/ud/ud_parser.py:230-242 | hanlp/components/srl/span_bio/span_bio.py:105-117 | hanlp/components/taggers/tagger.py:83-95 — the copies span different directories, so extracting a shared function means choosing where it lives: put it somewhere all 4 call sites can already reach — a location they all depend on today, or a new shared one if there is none — and call it from each site; until then, every change has to be made 4 times. The matched lines also transfer control out of the body holding them, which cannot survive a move into a called unit unchanged: have the extracted unit return that decision and let each site act on it.
Duplicated block (13 lines × 3) hanlp/components/amr/amrbart/preprocess/read_and_process.py:34— hanlp/components/amr/amrbart/preprocess/read_and_process.py:34-46 | hanlp/components/amr/seq2seq/dataset/tokenization_bart.py:266-278 | hanlp/components/amr/seq2seq/dataset/tokenization_t5.py:257-269 — before extracting anything, compare `hanlp/components/amr/seq2seq/dataset/tokenization_bart.py` and `hanlp/components/amr/seq2seq/dataset/tokenization_t5.py` as WHOLE FILES: this scan already matched 21 separate duplicated blocks between them, totalling at least 476 lines, which is the signature of one file having been copied from the other rather than of a helper waiting to be extracted. If that is what happened, the fix is to keep one copy and have the other call it (or delete it), which resolves this row and its siblings together — extracting one helper per block leaves the fork in place.
Duplicated block (8–12 lines × 2) hanlp/components/amr/amrbart/bart_amr_parser.py:131— hanlp/components/amr/amrbart/bart_amr_parser.py:131-142 | hanlp/components/amr/seq2seq/seq2seq_amr_parser.py:276-283 — the copies span different directories, so extracting a shared function means choosing where it lives: put it somewhere both call sites can already reach — a location they all depend on today, or a new shared one if there is none — and call it from each site; until then, every change has to be made twice. Read the line range as the matched WINDOW rather than a finished unit: at `hanlp/components/amr/amrbart/bart_amr_parser.py:131` it begins part-way through the construct above it, so those exact lines cannot be lifted as they stand — widen the region to the smallest complete statement or declaration that contains it, and extract that.
Duplicated block (11–12 lines × 2) hanlp/metrics/chunking/conlleval.py:296— hanlp/metrics/chunking/conlleval.py:296-306 | hanlp/metrics/chunking/conlleval.py:313-324 — both copies are in the same file, so extract the block into one function there and call it from each site — the copies drift apart the first time only one of them is edited. Read the line range as the matched WINDOW rather than a finished unit: at `hanlp/metrics/chunking/conlleval.py:296` it begins part-way through the construct above it, so those exact lines cannot be lifted as they stand — widen the region to the smallest complete statement or declaration that contains it, and extract that.
Duplicated block (10 lines × 4) REDACTED:348— REDACTED:348-357 | REDACTED:418-427 | REDACTED:476-485 | REDACTED:541-550 — before extracting anything, compare `REDACTED` and `REDACTED` as WHOLE FILES: this scan already matched 15 separate duplicated blocks between them, totalling at least 404 lines, which is the signature of one file having been copied from the other rather than of a helper waiting to be extracted. The two sit in different directories, so one cannot simply be deleted in favour of the other while both are reached separately: hoist the shared part into a location both already depend on and have each file call it, and retire whichever file turns out to have no caller of its own left. Extracting one helper per block leaves the fork in place.
Duplicated block (10 lines × 3) hanlp/components/amr/amrbart/model_interface/tokenization_bart.py:176— hanlp/components/amr/amrbart/model_interface/tokenization_bart.py:176-185 | hanlp/components/amr/seq2seq/dataset/tokenization_bart.py:356-365 | hanlp/components/amr/seq2seq/dataset/tokenization_t5.py:349-358 — before extracting anything, compare `hanlp/components/amr/amrbart/model_interface/tokenization_bart.py` and `hanlp/components/amr/seq2seq/dataset/tokenization_bart.py` as WHOLE FILES: this scan already matched 10 separate duplicated blocks between them, totalling at least 329 lines, which is the signature of one file having been copied from the other rather than of a helper waiting to be extracted. The two sit in different directories, so one cannot simply be deleted in favour of the other while both are reached separately: hoist the shared part into a location both already depend on and have each file call it, and retire whichever file turns out to have no caller of its own left. Extracting one helper per block leaves the fork in place. ★ These copies have DRIFTED, and that is worth reading before extracting anything: just after the matched lines, `hanlp/components/amr/seq2seq/dataset/tokenization_t5.py:360` calls `Graph` and `hanlp/components/amr/seq2seq/dataset/tokenization_bart.py:367` does not — after which the two agree again for 3 more lines. One of those two behaviours is the intended one and the other is what a copy-paste left behind, so decide which BEFORE unifying them: extracting the shared part will silently settle it, and if the copy that skips the call is the wrong one, that bug is already live.
Duplicated block (8–10 lines × 3) hanlp/components/parsers/biaffine/biaffine_dep.py:208— hanlp/components/parsers/biaffine/biaffine_dep.py:208-217 | hanlp/components/parsers/constituency/crf_constituency_parser.py:58-65 | hanlp/components/taggers/tagger.py:78-85 — the copies span different directories, so extracting a shared function means choosing where it lives: put it somewhere all 3 call sites can already reach — a location they all depend on today, or a new shared one if there is none — and call it from each site; until then, every change has to be made 3 times.
Duplicated block (7–9 lines × 2) hanlp/components/parsers/alg.py:139— hanlp/components/parsers/alg.py:139-145 | hanlp/components/parsers/alg.py:629-637 — both copies are in the same file, so extract the block into one function there and call it from each site — the copies drift apart the first time only one of them is edited. ★ These copies have DRIFTED, and that is worth reading before extracting anything: just before the matched lines, `hanlp/components/parsers/alg.py:628` calls `new_zeros` and `hanlp/components/parsers/alg.py:138` does not — after which the two agree again for 2 more lines. One of those two behaviours is the intended one and the other is what a copy-paste left behind, so decide which BEFORE unifying them: extracting the shared part will silently settle it, and if the copy that skips the call is the wrong one, that bug is already live.
Duplicated block (8 lines × 3) hanlp/components/amr/amrbart/model_interface/tokenization_bart.py:74— hanlp/components/amr/amrbart/model_interface/tokenization_bart.py:74-81 | hanlp/components/amr/seq2seq/dataset/tokenization_bart.py:95-102 | hanlp/components/amr/seq2seq/dataset/tokenization_t5.py:87-94 — before extracting anything, compare `hanlp/components/amr/amrbart/model_interface/tokenization_bart.py` and `hanlp/components/amr/seq2seq/dataset/tokenization_bart.py` as WHOLE FILES: this scan already matched 10 separate duplicated blocks between them, totalling at least 329 lines, which is the signature of one file having been copied from the other rather than of a helper waiting to be extracted. The two sit in different directories, so one cannot simply be deleted in favour of the other while both are reached separately: hoist the shared part into a location both already depend on and have each file call it, and retire whichever file turns out to have no caller of its own left. Extracting one helper per block leaves the fork in place. Read the line range as the matched WINDOW rather than a finished unit: at `hanlp/components/amr/amrbart/model_interface/tokenization_bart.py:74` it begins part-way through the construct above it, so those exact lines cannot be lifted as they stand — widen the region to the smallest complete statement or declaration that contains it, and extract that. The `return` at the foot of the matched lines is the enclosing body's own terminal exit, not an early one: it moves with them unchanged, and each site calls the extracted unit from the position that `return` occupied — no decision has to be handed back and re-acted on.
Duplicated block (2–8 lines × 6) hanlp/components/amr/amrbart/model_interface/tokenization_bart.py:72— hanlp/components/amr/amrbart/model_interface/tokenization_bart.py:72-73 | hanlp/components/amr/amrbart/model_interface/tokenization_bart.py:88-95 | hanlp/components/amr/seq2seq/dataset/tokenization_bart.py:93-94 | hanlp/components/amr/seq2seq/dataset/tokenization_bart.py:111-118 | hanlp/components/amr/seq2seq/dataset/tokenization_t5.py:85-86 | hanlp/components/amr/seq2seq/dataset/tokenization_t5.py:102-104 — before extracting anything, compare `hanlp/components/amr/amrbart/model_interface/tokenization_bart.py` and `hanlp/components/amr/seq2seq/dataset/tokenization_bart.py` as WHOLE FILES: this scan already matched 10 separate duplicated blocks between them, totalling at least 329 lines, which is the signature of one file having been copied from the other rather than of a helper waiting to be extracted. The two sit in different directories, so one cannot simply be deleted in favour of the other while both are reached separately: hoist the shared part into a location both already depend on and have each file call it, and retire whichever file turns out to have no caller of its own left. Extracting one helper per block leaves the fork in place. The `return` at the foot of the matched lines is the enclosing body's own terminal exit, not an early one: it moves with them unchanged, and each site calls the extracted unit from the position that `return` occupied — no decision has to be handed back and re-acted on. Note that the copies do not run to the end of the range shown: their LAST lines are different code, not the same code under different names — the matched region ends inside that line. Extract the lines above it, and read the last line of each site separately. ★ These copies have DRIFTED, and that is worth reading before extracting anything: just before the matched lines, `hanlp/components/amr/amrbart/model_interface/tokenization_bart.py:88` calls `extend` and `hanlp/components/amr/amrbart/model_interface/tokenization_bart.py:72` does not — after which the two agree again for 3 more lines. One of those two behaviours is the intended one and the other is what a copy-paste left behind, so decide which BEFORE unifying them: extracting the shared part will silently settle it, and if the copy that skips the call is the wrong one, that bug is already live.
Duplicated block (7 lines × 4) hanlp/components/amr/seq2seq/seq2seq_amr_parser.py:168— hanlp/components/amr/seq2seq/seq2seq_amr_parser.py:168-174 | hanlp/components/parsers/constituency/crf_constituency_parser.py:72-78 | hanlp/components/srl/span_bio/span_bio.py:114-120 | hanlp/components/taggers/tagger.py:92-98 — the copies span different directories, so extracting a shared function means choosing where it lives: put it somewhere all 4 call sites can already reach — a location they all depend on today, or a new shared one if there is none — and call it from each site; until then, every change has to be made 4 times.
Duplicated block (7 lines × 3) hanlp/components/parsers/constituency/crf_constituency_parser.py:49— hanlp/components/parsers/constituency/crf_constituency_parser.py:49-55 | hanlp/components/parsers/ud/ud_parser.py:216-222 | hanlp/components/srl/span_bio/span_bio.py:91-97 — the copies span different directories, so extracting a shared function means choosing where it lives: put it somewhere all 3 call sites can already reach — a location they all depend on today, or a new shared one if there is none — and call it from each site; until then, every change has to be made 3 times.
Duplicated block (5 lines × 5) hanlp/components/parsers/constituency/crf_constituency_parser.py:51— hanlp/components/parsers/constituency/crf_constituency_parser.py:51-55 | hanlp/components/parsers/ud/ud_parser.py:218-222 | hanlp/components/srl/span_bio/span_bio.py:93-97 | hanlp/components/sts/transformer_sts.py:82-86 | hanlp/components/taggers/tagger.py:71-75 — the copies span different directories, so extracting a shared function means choosing where it lives: put it somewhere all 5 call sites can already reach — a location they all depend on today, or a new shared one if there is none — and call it from each site; until then, every change has to be made 5 times.
Duplicated block (3–5 lines × 6) hanlp/components/amr/amrbart/model_interface/tokenization_bart.py:67— hanlp/components/amr/amrbart/model_interface/tokenization_bart.py:67-71 | hanlp/components/amr/amrbart/model_interface/tokenization_bart.py:85-87 | hanlp/components/amr/seq2seq/dataset/tokenization_bart.py:88-92 | hanlp/components/amr/seq2seq/dataset/tokenization_bart.py:108-110 | hanlp/components/amr/seq2seq/dataset/tokenization_t5.py:80-84 | hanlp/components/amr/seq2seq/dataset/tokenization_t5.py:100-102 — before extracting anything, compare `hanlp/components/amr/amrbart/model_interface/tokenization_bart.py` and `hanlp/components/amr/seq2seq/dataset/tokenization_bart.py` as WHOLE FILES: this scan already matched 10 separate duplicated blocks between them, totalling at least 329 lines, which is the signature of one file having been copied from the other rather than of a helper waiting to be extracted. The two sit in different directories, so one cannot simply be deleted in favour of the other while both are reached separately: hoist the shared part into a location both already depend on and have each file call it, and retire whichever file turns out to have no caller of its own left. Extracting one helper per block leaves the fork in place. Note that the copies do not run to the end of the range shown: their LAST lines are different code, not the same code under different names — the matched region ends inside that line. Extract the lines above it, and read the last line of each site separately. ★ These copies have DRIFTED, and that is worth reading before extracting anything: just after the matched lines, `hanlp/components/amr/amrbart/model_interface/tokenization_bart.py:71` calls `rsplit` and `hanlp/components/amr/amrbart/model_interface/tokenization_bart.py:87` does not — after which the two agree again for 3 more lines. One of those two behaviours is the intended one and the other is what a copy-paste left behind, so decide which BEFORE unifying them: extracting the shared part will silently settle it, and if the copy that skips the call is the wrong one, that bug is already live.
Duplicated block (44 lines × 2) hanlp/components/amr/amrbart/common/penman_interface.py:37— hanlp/components/amr/amrbart/common/penman_interface.py:37-80 | hanlp/components/amr/amrbart/preprocess/penman_interface.py:37-80 — `hanlp/components/amr/amrbart/common/penman_interface.py` and `hanlp/components/amr/amrbart/preprocess/penman_interface.py` are one unit implemented once per sibling directory, so they are most likely parallel implementations of one contract rather than a copy of each other — this scan matched 1 separate duplicated blocks between them, totalling at least 44 lines. If both are selected at run time, neither can be retired in favour of the other, and the lines that DIFFER between them are the reason both exist. The move that pays here is to hoist the identical part into a shared location the whole family can reach and give what differs a parameter or a seam, so a change lands once instead of once per sibling; extracting one helper per block leaves every sibling to drift on its own.
X17 · Uncapped recursion over a caller-supplied document· Recursive walk over a caller-supplied document has no depth cap · ×1
Recursive walk over a caller-supplied document has no depth cap plugins/hanlp_common/hanlp_common/configurable.py:27— `from_config` is public and takes `config: dict` from its caller, and it calls itself at line 27 (`self.from_config`) with a piece of `config` (and at line 32) — and nowhere in `from_config` is there a depth parameter, a descent counter, or any other cap. The walk therefore spends one stack frame per LEVEL of the value it is handed, so the nesting of the caller's document — not this code — decides how deep the stack goes. Python's parsers accept far more nesting than its stack allows: `json.loads` turns 3,000 bytes of brackets into 1,500 levels and `xml.etree.ElementTree` parses tens of thousands, while a walk like this one raises `RecursionError` once it passes the interpreter's default limit of 1,000 frames. Wherever that exception lands, the work fails: a request errors, and outside any `try` the worker or the program exits with it. Raising `sys.setrecursionlimit` only moves the number the document has to beat. Thread a depth argument through `from_config` and reject the value past a fixed maximum, or rewrite the descent with an explicit stack.
Interpolated log message defeats structured logging REDACTED:339— `debug` at line 339 is handed an f-string as its message, so the logger receives one finished string — the values in it cannot be queried, filtered or aggregated as fields, every occurrence reads as a different message, and the string is built even when the record is filtered out. Pass the values beside a constant message instead: `logger.debug("order %s moved to %s", order_id, status)` with the standard library's `logging`, which keeps the message and its arguments apart on the record, or `log.debug("order moved", order_id=order_id, status=status)` with structlog.
Interpolated log message defeats structured logging hanlp/components/taggers/rnn_tagger.py:52— `info` at line 52 is handed an f-string as its message, so the logger receives one finished string — the values in it cannot be queried, filtered or aggregated as fields, every occurrence reads as a different message, and the string is built even when the record is filtered out. Pass the values beside a constant message instead: `logger.info("order %s moved to %s", order_id, status)` with the standard library's `logging`, which keeps the message and its arguments apart on the record, or `log.info("order moved", order_id=order_id, status=status)` with structlog.
Interpolated log message defeats structured logging hanlp/components/taggers/rnn_tagger.py:76— `info` at line 76 is handed an f-string as its message, so the logger receives one finished string — the values in it cannot be queried, filtered or aggregated as fields, every occurrence reads as a different message, and the string is built even when the record is filtered out. Pass the values beside a constant message instead: `logger.info("order %s moved to %s", order_id, status)` with the standard library's `logging`, which keeps the message and its arguments apart on the record, or `log.info("order moved", order_id=order_id, status=status)` with structlog.
Interpolated log message defeats structured logging hanlp/components/taggers/rnn_tagger.py:77— `info` at line 77 is handed an f-string as its message, so the logger receives one finished string — the values in it cannot be queried, filtered or aggregated as fields, every occurrence reads as a different message, and the string is built even when the record is filtered out. Pass the values beside a constant message instead: `logger.info("order %s moved to %s", order_id, status)` with the standard library's `logging`, which keeps the message and its arguments apart on the record, or `log.info("order moved", order_id=order_id, status=status)` with structlog.
Interpolated log message defeats structured logging hanlp/components/taggers/tagger.py:75— `info` at line 75 is handed an f-string as its message, so the logger receives one finished string — the values in it cannot be queried, filtered or aggregated as fields, every occurrence reads as a different message, and the string is built even when the record is filtered out. Pass the values beside a constant message instead: `logger.info("order %s moved to %s", order_id, status)` with the standard library's `logging`, which keeps the message and its arguments apart on the record, or `log.info("order moved", order_id=order_id, status=status)` with structlog.
Interpolated log message defeats structured logging hanlp/components/taggers/tagger.py:96— `info` at line 96 is handed an f-string as its message, so the logger receives one finished string — the values in it cannot be queried, filtered or aggregated as fields, every occurrence reads as a different message, and the string is built even when the record is filtered out. Pass the values beside a constant message instead: `logger.info("order %s moved to %s", order_id, status)` with the standard library's `logging`, which keeps the message and its arguments apart on the record, or `log.info("order moved", order_id=order_id, status=status)` with structlog.
Interpolated log message defeats structured logging hanlp/components/taggers/tagger.py:97— `info` at line 97 is handed an f-string as its message, so the logger receives one finished string — the values in it cannot be queried, filtered or aggregated as fields, every occurrence reads as a different message, and the string is built even when the record is filtered out. Pass the values beside a constant message instead: `logger.info("order %s moved to %s", order_id, status)` with the standard library's `logging`, which keeps the message and its arguments apart on the record, or `log.info("order moved", order_id=order_id, status=status)` with structlog.
Interpolated log message defeats structured logging hanlp/components/taggers/tagger.py:98— `info` at line 98 is handed an f-string as its message, so the logger receives one finished string — the values in it cannot be queried, filtered or aggregated as fields, every occurrence reads as a different message, and the string is built even when the record is filtered out. Pass the values beside a constant message instead: `logger.info("order %s moved to %s", order_id, status)` with the standard library's `logging`, which keeps the message and its arguments apart on the record, or `log.info("order moved", order_id=order_id, status=status)` with structlog.
Interpolated log message defeats structured logging hanlp/components/taggers/transformers/transformer_tagger.py:160— `info` at line 160 is handed an f-string as its message, so the logger receives one finished string — the values in it cannot be queried, filtered or aggregated as fields, every occurrence reads as a different message, and the string is built even when the record is filtered out. Pass the values beside a constant message instead: `logger.info("order %s moved to %s", order_id, status)` with the standard library's `logging`, which keeps the message and its arguments apart on the record, or `log.info("order moved", order_id=order_id, status=status)` with structlog.
Interpolated log message defeats structured logging hanlp/components/taggers/transformers/transformer_tagger.py:181— `info` at line 181 is handed an f-string as its message, so the logger receives one finished string — the values in it cannot be queried, filtered or aggregated as fields, every occurrence reads as a different message, and the string is built even when the record is filtered out. Pass the values beside a constant message instead: `logger.info("order %s moved to %s", order_id, status)` with the standard library's `logging`, which keeps the message and its arguments apart on the record, or `log.info("order moved", order_id=order_id, status=status)` with structlog.
Interpolated log message defeats structured logging hanlp/components/sts/transformer_sts.py:86— `info` at line 86 is handed an f-string as its message, so the logger receives one finished string — the values in it cannot be queried, filtered or aggregated as fields, every occurrence reads as a different message, and the string is built even when the record is filtered out. Pass the values beside a constant message instead: `logger.info("order %s moved to %s", order_id, status)` with the standard library's `logging`, which keeps the message and its arguments apart on the record, or `log.info("order moved", order_id=order_id, status=status)` with structlog.
Interpolated log message defeats structured logging hanlp/components/sts/transformer_sts.py:98— `info` at line 98 is handed an f-string as its message, so the logger receives one finished string — the values in it cannot be queried, filtered or aggregated as fields, every occurrence reads as a different message, and the string is built even when the record is filtered out. Pass the values beside a constant message instead: `logger.info("order %s moved to %s", order_id, status)` with the standard library's `logging`, which keeps the message and its arguments apart on the record, or `log.info("order moved", order_id=order_id, status=status)` with structlog.
Interpolated log message defeats structured logging hanlp/components/srl/span_rank/span_rank.py:97— `info` at line 97 is handed an f-string as its message, so the logger receives one finished string — the values in it cannot be queried, filtered or aggregated as fields, every occurrence reads as a different message, and the string is built even when the record is filtered out. Pass the values beside a constant message instead: `logger.info("order %s moved to %s", order_id, status)` with the standard library's `logging`, which keeps the message and its arguments apart on the record, or `log.info("order moved", order_id=order_id, status=status)` with structlog.
Interpolated log message defeats structured logging hanlp/components/srl/span_bio/span_bio.py:97— `info` at line 97 is handed an f-string as its message, so the logger receives one finished string — the values in it cannot be queried, filtered or aggregated as fields, every occurrence reads as a different message, and the string is built even when the record is filtered out. Pass the values beside a constant message instead: `logger.info("order %s moved to %s", order_id, status)` with the standard library's `logging`, which keeps the message and its arguments apart on the record, or `log.info("order moved", order_id=order_id, status=status)` with structlog.
Interpolated log message defeats structured logging hanlp/components/srl/span_bio/span_bio.py:118— `info` at line 118 is handed an f-string as its message, so the logger receives one finished string — the values in it cannot be queried, filtered or aggregated as fields, every occurrence reads as a different message, and the string is built even when the record is filtered out. Pass the values beside a constant message instead: `logger.info("order %s moved to %s", order_id, status)` with the standard library's `logging`, which keeps the message and its arguments apart on the record, or `log.info("order moved", order_id=order_id, status=status)` with structlog.
Interpolated log message defeats structured logging hanlp/components/srl/span_bio/span_bio.py:119— `info` at line 119 is handed an f-string as its message, so the logger receives one finished string — the values in it cannot be queried, filtered or aggregated as fields, every occurrence reads as a different message, and the string is built even when the record is filtered out. Pass the values beside a constant message instead: `logger.info("order %s moved to %s", order_id, status)` with the standard library's `logging`, which keeps the message and its arguments apart on the record, or `log.info("order moved", order_id=order_id, status=status)` with structlog.
Interpolated log message defeats structured logging hanlp/components/srl/span_bio/span_bio.py:120— `info` at line 120 is handed an f-string as its message, so the logger receives one finished string — the values in it cannot be queried, filtered or aggregated as fields, every occurrence reads as a different message, and the string is built even when the record is filtered out. Pass the values beside a constant message instead: `logger.info("order %s moved to %s", order_id, status)` with the standard library's `logging`, which keeps the message and its arguments apart on the record, or `log.info("order moved", order_id=order_id, status=status)` with structlog.
Interpolated log message defeats structured logging hanlp/components/srl/span_bio/span_bio.py:289— `info` at line 289 is handed an f-string as its message, so the logger receives one finished string — the values in it cannot be queried, filtered or aggregated as fields, every occurrence reads as a different message, and the string is built even when the record is filtered out. Pass the values beside a constant message instead: `logger.info("order %s moved to %s", order_id, status)` with the standard library's `logging`, which keeps the message and its arguments apart on the record, or `log.info("order moved", order_id=order_id, status=status)` with structlog.
Interpolated log message defeats structured logging hanlp/components/parsers/ud/ud_parser.py:222— `info` at line 222 is handed an f-string as its message, so the logger receives one finished string — the values in it cannot be queried, filtered or aggregated as fields, every occurrence reads as a different message, and the string is built even when the record is filtered out. Pass the values beside a constant message instead: `logger.info("order %s moved to %s", order_id, status)` with the standard library's `logging`, which keeps the message and its arguments apart on the record, or `log.info("order moved", order_id=order_id, status=status)` with structlog.
Interpolated log message defeats structured logging hanlp/components/parsers/ud/ud_parser.py:243— `info` at line 243 is handed an f-string as its message, so the logger receives one finished string — the values in it cannot be queried, filtered or aggregated as fields, every occurrence reads as a different message, and the string is built even when the record is filtered out. Pass the values beside a constant message instead: `logger.info("order %s moved to %s", order_id, status)` with the standard library's `logging`, which keeps the message and its arguments apart on the record, or `log.info("order moved", order_id=order_id, status=status)` with structlog.
Interpolated log message defeats structured logging hanlp/components/parsers/ud/ud_parser.py:244— `info` at line 244 is handed an f-string as its message, so the logger receives one finished string — the values in it cannot be queried, filtered or aggregated as fields, every occurrence reads as a different message, and the string is built even when the record is filtered out. Pass the values beside a constant message instead: `logger.info("order %s moved to %s", order_id, status)` with the standard library's `logging`, which keeps the message and its arguments apart on the record, or `log.info("order moved", order_id=order_id, status=status)` with structlog.
Interpolated log message defeats structured logging hanlp/components/parsers/ud/ud_parser.py:245— `info` at line 245 is handed an f-string as its message, so the logger receives one finished string — the values in it cannot be queried, filtered or aggregated as fields, every occurrence reads as a different message, and the string is built even when the record is filtered out. Pass the values beside a constant message instead: `logger.info("order %s moved to %s", order_id, status)` with the standard library's `logging`, which keeps the message and its arguments apart on the record, or `log.info("order moved", order_id=order_id, status=status)` with structlog.
Interpolated log message defeats structured logging hanlp/components/parsers/constituency/crf_constituency_parser.py:55— `info` at line 55 is handed an f-string as its message, so the logger receives one finished string — the values in it cannot be queried, filtered or aggregated as fields, every occurrence reads as a different message, and the string is built even when the record is filtered out. Pass the values beside a constant message instead: `logger.info("order %s moved to %s", order_id, status)` with the standard library's `logging`, which keeps the message and its arguments apart on the record, or `log.info("order moved", order_id=order_id, status=status)` with structlog.
Interpolated log message defeats structured logging hanlp/components/parsers/constituency/crf_constituency_parser.py:76— `info` at line 76 is handed an f-string as its message, so the logger receives one finished string — the values in it cannot be queried, filtered or aggregated as fields, every occurrence reads as a different message, and the string is built even when the record is filtered out. Pass the values beside a constant message instead: `logger.info("order %s moved to %s", order_id, status)` with the standard library's `logging`, which keeps the message and its arguments apart on the record, or `log.info("order moved", order_id=order_id, status=status)` with structlog.
Interpolated log message defeats structured logging hanlp/components/parsers/constituency/crf_constituency_parser.py:77— `info` at line 77 is handed an f-string as its message, so the logger receives one finished string — the values in it cannot be queried, filtered or aggregated as fields, every occurrence reads as a different message, and the string is built even when the record is filtered out. Pass the values beside a constant message instead: `logger.info("order %s moved to %s", order_id, status)` with the standard library's `logging`, which keeps the message and its arguments apart on the record, or `log.info("order moved", order_id=order_id, status=status)` with structlog.
Duplicated predicate hanlp/components/mtl/tasks/tok/tag_tok.py:173— `dictionary is not None and not isinstance(dictionary, DictInterface)` appears character-identically in 4 files — hanlp/components/mtl/tasks/tok/tag_tok.py, hanlp/components/ner/transformer_ner.py, hanlp/components/taggers/tagger.py, hanlp/components/tokenizers/transformer.py. It is one line, so the duplication detector's token window never sees it; the copies drift when only one is corrected. Give the condition a name and one home.
Duplicated predicate REDACTED:261— `edge_index is None or edge_index == (l - 1)` appears character-identically in 2 files — REDACTED, REDACTED. It is one line, so the duplication detector's token window never sees it; the copies drift when only one is corrected. Give the condition a name and one home.
Duplicated predicate hanlp/components/amr/seq2seq/seq2seq_amr_parser.py:186— `gradient_accumulation and gradient_accumulation > 1` appears character-identically in 9 files — hanlp/components/amr/seq2seq/seq2seq_amr_parser.py, hanlp/components/mtl/multi_task_learning.py, hanlp/components/ner/biaffine_ner/biaffine_ner.py, hanlp/components/parsers/constituency/crf_constituency_parser.py. It is one line, so the duplication detector's token window never sees it; the copies drift when only one is corrected. Give the condition a name and one home.
Duplicated predicate hanlp/components/amr/amrbart/model_interface/tokenization_bart.py:214— `isinstance(fol, str) and (fol.startswith(':') or (fol == ')'))` appears character-identically in 3 files — hanlp/components/amr/amrbart/model_interface/tokenization_bart.py, hanlp/components/amr/seq2seq/dataset/tokenization_bart.py, hanlp/components/amr/seq2seq/dataset/tokenization_t5.py. It is one line, so the duplication detector's token window never sees it; the copies drift when only one is corrected. Give the condition a name and one home.
Duplicated predicate hanlp/components/amr/amrbart/model_interface/tokenization_bart.py:193— `isinstance(nxt, str) and nxt.startswith('<pointer:')` appears character-identically in 3 files — hanlp/components/amr/amrbart/model_interface/tokenization_bart.py, hanlp/components/amr/seq2seq/dataset/tokenization_bart.py, hanlp/components/amr/seq2seq/dataset/tokenization_t5.py. It is one line, so the duplication detector's token window never sees it; the copies drift when only one is corrected. Give the condition a name and one home.
Duplicated predicate REDACTED:98— `isinstance(tokens[-1], str) and rex_spc.match(tokens[-1])` appears character-identically in 2 files — REDACTED, REDACTED. It is one line, so the duplication detector's token window never sees it; the copies drift when only one is corrected. Give the condition a name and one home.
Duplicated predicate hanlp/common/torch_component.py:127— `isinstance(v, dict) and 'classpath' in v` appears character-identically in 2 files — hanlp/common/torch_component.py, plugins/hanlp_common/hanlp_common/configurable.py. It is one line, so the duplication detector's token window never sees it; the copies drift when only one is corrected. Give the condition a name and one home.
Duplicated predicate hanlp/components/amr/amrbart/model_interface/tokenization_bart.py:72— `len(recats) == 2 and recats[0] in self.recategorizations and ('_' + recats[1]) in self.encoder` appears character-identically in 3 files — hanlp/components/amr/amrbart/model_interface/tokenization_bart.py, hanlp/components/amr/seq2seq/dataset/tokenization_bart.py, hanlp/components/amr/seq2seq/dataset/tokenization_t5.py. It is one line, so the duplication detector's token window never sees it; the copies drift when only one is corrected. Give the condition a name and one home.
Duplicated predicate hanlp/components/amr/seq2seq/dataset/tokenization_bart.py:314— `lst == '(' and linearized_nodes[i + 1] == '/'` appears character-identically in 2 files — hanlp/components/amr/seq2seq/dataset/tokenization_bart.py, hanlp/components/amr/seq2seq/dataset/tokenization_t5.py. It is one line, so the duplication detector's token window never sees it; the copies drift when only one is corrected. Give the condition a name and one home.
Duplicated predicate REDACTED:403— `n.startswith('"') and n.endswith('"') and len(n) > 2` appears character-identically in 2 files — REDACTED, REDACTED. It is one line, so the duplication detector's token window never sees it; the copies drift when only one is corrected. Give the condition a name and one home.
Duplicated predicate hanlp/components/amr/amrbart/model_interface/tokenization_bart.py:179— `n.startswith('<') and n.endswith('>') and (not n.startswith('<pointer:'))` appears character-identically in 3 files — hanlp/components/amr/amrbart/model_interface/tokenization_bart.py, hanlp/components/amr/seq2seq/dataset/tokenization_bart.py, hanlp/components/amr/seq2seq/dataset/tokenization_t5.py. It is one line, so the duplication detector's token window never sees it; the copies drift when only one is corrected. Give the condition a name and one home.
Duplicated predicate hanlp/components/amr/amrbart/model_interface/tokenization_bart.py:459— `node.startswith('"') and node.endswith('"')` appears character-identically in 3 files — hanlp/components/amr/amrbart/model_interface/tokenization_bart.py, hanlp/components/amr/seq2seq/dataset/tokenization_bart.py, hanlp/components/amr/seq2seq/dataset/tokenization_t5.py. It is one line, so the duplication detector's token window never sees it; the copies drift when only one is corrected. Give the condition a name and one home.
Duplicated predicate hanlp/components/parsers/alg_tf.py:60— `old is None or not tf.reduce_all(c == old)` appears character-identically in 2 files — hanlp/components/parsers/alg_tf.py, hanlp/components/parsers/biaffine_tf/alg.py. It is one line, so the duplication detector's token window never sees it; the copies drift when only one is corrected. Give the condition a name and one home.
Duplicated predicate hanlp/components/amr/amrbart/preprocess/read_and_process.py:37— `piece.startswith('"') and piece.endswith('"')` appears character-identically in 3 files — hanlp/components/amr/amrbart/preprocess/read_and_process.py, hanlp/components/amr/seq2seq/dataset/tokenization_bart.py, hanlp/components/amr/seq2seq/dataset/tokenization_t5.py. It is one line, so the duplication detector's token window never sees it; the copies drift when only one is corrected. Give the condition a name and one home.
Duplicated predicate hanlp/components/amr/amrbart/model_interface/tokenization_bart.py:309— `piece.startswith('"') or piece.startswith('"') or '"' in piece.strip('"')` appears character-identically in 3 files — hanlp/components/amr/amrbart/model_interface/tokenization_bart.py, hanlp/components/amr/seq2seq/dataset/tokenization_bart.py, hanlp/components/amr/seq2seq/dataset/tokenization_t5.py. It is one line, so the duplication detector's token window never sees it; the copies drift when only one is corrected. Give the condition a name and one home.
Duplicated predicate hanlp/components/amr/seq2seq/dataset/tokenization_bart.py:458— `self._classify(next) == 'VAR' and next in remap and (i < len(nodes) - 1) and nodes[i + 1] != '/'` appears character-identically in 2 files — hanlp/components/amr/seq2seq/dataset/tokenization_bart.py, hanlp/components/amr/seq2seq/dataset/tokenization_t5.py. It is one line, so the duplication detector's token window never sees it; the copies drift when only one is corrected. Give the condition a name and one home.
Duplicated predicate REDACTED:49— `tok.endswith('"') and (not tok.startswith('"'))` appears character-identically in 2 files — REDACTED, REDACTED. It is one line, so the duplication detector's token window never sees it; the copies drift when only one is corrected. Give the condition a name and one home.
Duplicated predicate REDACTED:47— `tok.startswith('"') and (not tok.endswith('"'))` appears character-identically in 2 files — REDACTED, REDACTED. It is one line, so the duplication detector's token window never sees it; the copies drift when only one is corrected. Give the condition a name and one home.
Duplicated predicate hanlp/components/amr/amrbart/model_interface/tokenization_bart.py:106— `tokk.startswith('"') and tokk.endswith('"')` appears character-identically in 3 files — hanlp/components/amr/amrbart/model_interface/tokenization_bart.py, hanlp/components/amr/seq2seq/dataset/tokenization_bart.py, hanlp/components/amr/seq2seq/dataset/tokenization_t5.py. It is one line, so the duplication detector's token window never sees it; the copies drift when only one is corrected. Give the condition a name and one home.
Duplicated predicate hanlp/components/amr/amrbart/model_interface/tokenization_bart.py:103— `tokk.startswith(':') and tokk.endswith('-of')` appears character-identically in 3 files — hanlp/components/amr/amrbart/model_interface/tokenization_bart.py, hanlp/components/amr/seq2seq/dataset/tokenization_bart.py, hanlp/components/amr/seq2seq/dataset/tokenization_t5.py. It is one line, so the duplication detector's token window never sees it; the copies drift when only one is corrected. Give the condition a name and one home.
Duplicated predicate hanlp/components/amr/amrbart/model_interface/tokenization_bart.py:102— `tokk.startswith('<') and tokk.endswith('>')` appears character-identically in 3 files — hanlp/components/amr/amrbart/model_interface/tokenization_bart.py, hanlp/components/amr/seq2seq/dataset/tokenization_bart.py, hanlp/components/amr/seq2seq/dataset/tokenization_t5.py. It is one line, so the duplication detector's token window never sees it; the copies drift when only one is corrected. Give the condition a name and one home.
Duplicated predicate hanlp/components/amr/seq2seq/dataset/linearization.py:153— `v1 in name_vars and rel.startswith(':op')` appears character-identically in 2 files — hanlp/components/amr/seq2seq/dataset/linearization.py, REDACTED. It is one line, so the duplication detector's token window never sees it; the copies drift when only one is corrected. Give the condition a name and one home.
D26 · Project Cohesion· Projects may be oversized for their cohesion · ×1
Projects may be oversized for their cohesion — 1 of 5 project(s) overshoot their size bounds, lowering Project Cohesion to 6.0/10. The most over is `(repository root)` (44085 LoC, 476 module-visible types across 67 directories). Review these for cohesion — draw the boundary inside the module first (group each responsibility into its own package or directory and keep the cross-boundary members non-public), since splitting a published package moves types between packages and breaks consumers.
D28 · Secrets (history)· Rotate the exposed credentials · ×1
No ADRs — No Architecture Decision Records found — no conventional ADR directory, no numbered `NNNN-title` documents in any markup this check reads, and nothing ADR-shaped by content. Design rationale recorded elsewhere (a design-notes tree, a mailing list, pull-request discussion) is not visible to this check and is not re-findable per decision, so a future maintainer cannot ask why one choice was made and get an answer.
No src/ separation — Production code isn't grouped under a src/ folder — it's spread across several top-level directories, so there's no one place that says 'this is the product'.
P10 · Library API & versioning· Large public API surface · ×1
Large public API surface — 11/11 types (100%) declared in the published library are public. For a library, every public type is a stability contract — keep implementation types off the surface and expose only the intended API.
No SAST — No static application security testing detected. For this repository's stack, add bandit, `semgrep --config=p/python`, or CodeQL's python pack as a CI step. What was searched, so you can tell an absence from a miss: the 2244 CI workflow file(s) in this repository, and the scanner and linter configuration checked in beside them. A scan that runs outside CI, one configured in your forge's web UI rather than in a committed file, or a tool whose name is none of those this check carries, is not seen — if that is your case the row is wrong, and saying so is more useful than adding a second scanner.
No release approval gate — The release is automated and no gate that pauses it for a human is DECLARED IN THIS REPOSITORY'S PIPELINE FILES. What was read: every file under `.github/workflows/`, `.forgejo/workflows/`, `.gitea/workflows/`, `.azuredevops/` and `.azure-pipelines/`, plus `.gitlab-ci*` and `azure-pipelines*` — with comment text stripped, so documenting a gate is not declaring one. What would have counted: GitLab's `when: manual`, CircleCI's `type: approval`, an Azure `ManualValidation@` task or an `approvals:` block, a Jenkins `input` step, a `uses:` step naming an approval action, an `environment:` paired with `reviewers` / `required_reviewers` / `protection` / `wait-timer` / `deployment_branch_policy`, a draft-release step, a `workflow_dispatch` promotion, or a release-event gate. ★ What this cannot see, because none of it is a file: a GitHub environment whose required reviewers are configured in repo SETTINGS, a branch protection rule, or an organisation deployment policy — all of them real, enforced gates that live outside the repository. If yours is one of those, this row is wrong and nothing in the tree could have told us. Otherwise: whatever the release trigger points at is published to users unreviewed, so a mistagged or unverified commit ships and the only remedy is a follow-up release.
No changelog — No CHANGELOG/HISTORY/RELEASES file — what shipped when isn't easy to reconstruct for support or audit. (Versioning/tagging makes releases traceable, but a changelog records the what.)
X5 · Nullable reference types· Java packages are not null-marked everywhere · ×1
Java packages are not null-marked everywhere — 0/2 of this repository's Java package directories are under a non-null default: a `package-info.java` declaring `@NullMarked` (JSpecify) or `@NonNullApi`, `@ParametersAreNonnullByDefault`, `@NonNullByDefault`, `@DefaultQualifier(NonNull.class)`; a `module-info.java` declaring `@NullMarked`; every type in it `@NullMarked`; or a build that runs NullAway with the package in `AnnotatedPackages`, or the Checker Framework's `NullnessChecker`. Elsewhere every reference admits null and no checker warns on the dereference that throws a `NullPointerException`.
Appendix B — Reproduction & audit trail
Every external tool invocation behind a deep-scan dimension — the tool, its captured version, the exact command, how many findings it yielded, and a link to the retained raw output. To reproduce any finding: check out the same commit and run the command shown (repo-relative — never an absolute scratch path). The complete raw scanner output is retained verbatim under artifacts/raw/ (indexed in artifacts/raw/index.json); per-invocation exit codes and wall-clock durations are in sidecar.json — kept out of this table so the rendered report stays byte-identical across runs of the same commit.
trivy: not applicable — No Infrastructure-as-Code or container manifests found (Dockerfile, Docker Compose, Terraform, Kubernetes/Helm, CloudFormation, ARM, Bicep, Ansible); nothing to scan.
semgrep: not applicable — No personal data was found crossing a boundary the PII/GDPR ruleset checks — nothing written to a log or console sink, placed in a URL or query string, or persisted to browser storage. That is a clean result for the LEAK surface only: this ruleset detects personal data escaping, it does not inventory the personal data a repository holds, so it is not evidence that this repository has no personal-data surface. The personal-data map (Appendix C) and the C1-C5 compliance cards are what speak to that.
disclosure: not applicable — No vulnerability-disclosure policy file found (SECURITY.md/.markdown/.rst/.txt at root or under .github/.forgejo/.gitea/docs, .well-known/security.txt). A coordinated-disclosure policy may live off-repo, so this is not evidenced rather than failed.
runtime-hardening: not applicable — No Kubernetes/orchestration workloads found in the repository manifests; network egress policy is a cluster-native control that may live at the platform/firewall layer, so there is nothing to assess here.
runtime-hardening: not applicable — No Kubernetes/orchestration workloads found in the repository manifests; seccomp/AppArmor/SELinux confinement is a workload-level control, so there is nothing to assess here.
runtime-hardening: not applicable — No Kubernetes/orchestration workloads found in the repository manifests; runtime threat-detection and admission-control policy are cluster-level controls, so there is nothing to assess here.
Run 01a17221-1573-7685-992b-798e445d66d3 · every finding is also locatable in findings.md, and the complete scoring record (with exit codes + durations) in sidecar.json.
Issues: 14 · Warnings: 389 · Recommendations: 77 — Appendix A · all findings · full markdown report.
Generated by Watchdog — deterministic code-health analysis. 25-10-2026 @ 08:15 UTC.
Downloadable artifacts
Machine-readable and reproducible from this commit + frozen rubric — drop them straight into a contract appendix, a CRA dossier, or a downstream SCA / VEX tool.