Replace Any type annotations with specific types and remove redundant type annotation reassignments

[Copilot]

What do these changes do

Removes redundant type annotations from instance-level assignments when type is already declared at class level. Affects 14 classes across 10 files with 60+ redundant annotations eliminated.

What was wrong

Type annotations were duplicated at both class and instance levels, violating PEP 484 best practices:

class ChatBotModel:
    history: list[tuple[str, str]]  # Class level
    
    def __init__(self):
        self.history: list[tuple[str, str]] = []  # Redundant reassignment

This pattern appeared in:

• FastTextEncoder, Words_Spelling_Correction (words_spelling_correction.py)
• ThaiTransliterator, Encoder, Attn, AttentionDecoder (thai2rom.py)
• _ResponseWrapper (corpus/core.py)
• ChatBotModel (chat/core.py)
• GzipModel (classify/param_free.py)
• WngchanBerta_ONNX (tag/wangchanberta_onnx.py)
• ThaiNameTagger (wangchanberta/core.py)
• CRFchunk (tag/crfchunk.py)

How this fixes it

Keeps type annotation only at class level, removes from instance assignments:

class ChatBotModel:
    history: list[tuple[str, str]]
    
    def __init__(self):
        self.history = []  # Type inferred from class annotation

Verified with mypy - no new type errors introduced.

Your checklist for this pull request

• Passed code styles and structures
• Passed code linting checks and unit test

Original prompt

Find type annotations with Any and try to scope it down, providing a more specific and precise type.

It is possible to find type information from the inspection of the library a runtime (required installation) and it is also possible to analyse the source code of the library to obtain the type informaton.

Open source library tend to have source code available on the internet, for example in GitHub. Find information about library's source code repo location from PyPI website or from pyproject.toml/setup.cfg/setup.py metadata in the pip package.

Break job into smaller batches and iterate until complete.

---

✨ Let Copilot coding agent set things up for you — coding agent works faster and does higher quality work when set up for your repo.

[bact]

@copilot Do not use A | B union syntax. Use Union[] and Optional[] for strict compatibility with Python 3.9 - to make sure the inspection tools can work correctly.

[Copilot]

@copilot Do not use A | B union syntax. Use Union[] and Optional[] for strict compatibility with Python 3.9 - to make sure the inspection tools can work correctly.

Replaced all pipe union syntax (|) with Union[] and Optional[] for Python 3.9 compatibility. Changes made in:

• benchmarks/word_tokenization.py: int | str → Union[int, str]
• tag/crfchunk.py: str | bool → Union[str, bool]

Commit: d0d030e

[coveralls]

coverage: 64.79% (+0.004%) from 64.786%
when pulling cc24a14 on copilot/refactor-any-type-annotations
into 8a188f2 on dev.

[Copilot]

Pull request overview

This pull request improves type safety by replacing 28+ Any type annotations with specific types across 16 files. The changes include using proper types from third-party libraries (transformers, fastai, nltk, epitran), creating type aliases for complex unions, and introducing a TypedDict for structured return values. The PR correctly uses Union[] and Optional[] syntax for Python 3.9 compatibility.

Changes:

• Replaced Any with specific types from third-party libraries (transformers.Pipeline, fastai.Learner/DataBunch, nltk.Synset, epitran.Epitran, argparse.Namespace)
• Created NEREngineType type alias for complex Union type in named entity recognition
• Created CorefResultDict TypedDict for structured coreference resolution results
• Refined dictionary type annotations from dict[str, Any] to more specific dict[str, Union[str, bool]] and dict[str, Union[int, str]]
• Eliminated mypy errors across all affected modules (from 11-18 errors per module down to 0)

Reviewed changes

Copilot reviewed 16 out of 16 changed files in this pull request and generated 4 comments.

Show a summary per file

File	Description
pythainlp/word_vector/core.py	Changed tokenizer type from Any to Callable[[str], list[str]]
pythainlp/ulmfit/core.py	Added fastai types: Learner and DataBunch
pythainlp/transliterate/ipa.py	Changed epitran type from Any to epitran._epitran.Epitran
pythainlp/tag/unigram.py	Refined dict types from dict[str, Any] to dict[str, str] for POS taggers
pythainlp/tag/named_entity.py	Created NEREngineType type alias and updated engine types
pythainlp/tag/crfchunk.py	Refined feature dict type from dict[str, Any] to dict[str, Union[str, bool]]
pythainlp/summarize/keybert.py	Changed pipeline type from Any to transformers.pipelines.base.Pipeline
pythainlp/phayathaibert/core.py	Changed model type from Any to Pipeline
pythainlp/generate/thai2fit.py	Added Path and Learner types from fastai
pythainlp/coref/_fastcoref.py	Created CorefResultDict TypedDict and added CorefResult type
pythainlp/cli/tokenize.py	Changed args type from Any to argparse.Namespace
pythainlp/cli/tag.py	Changed args type from Any to argparse.Namespace
pythainlp/chat/core.py	Changed model type from Any to WangChanGLM
pythainlp/benchmarks/word_tokenization.py	Refined dict return types to dict[str, Union[int, str]]
pythainlp/augment/wordnet.py	Changed synset type from Any to nltk.corpus.reader.wordnet.Synset
build_tools/analysis/output/type_hint_analysis.json	Updated type analysis metrics showing elimination of mypy errors

[bact]

@copilot apply changes based on the comments in this thread

[Copilot]

@copilot apply changes based on the comments in this thread

Applied all suggested changes from the review:

1. named_entity.py:57 - Added None to NEREngineType Union to match line 66 assignment
2. named_entity.py:157 - Removed redundant type annotation in NNER.init
3. transliterate/ipa.py:16 - Changed to public epitran.Epitran API instead of private epitran._epitran.Epitran
4. benchmarks/word_tokenization.py:136 - Converted all numpy integer types to Python int with explicit int() conversions

Commit: d79928c

[Copilot]

Pull request overview

Copilot reviewed 22 out of 22 changed files in this pull request and generated 1 comment.

[sonarqubecloud]

Quality Gate passed

Issues
1 New issue
0 Accepted issues

Measures
0 Security Hotspots
0.0% Coverage on New Code
0.0% Duplication on New Code

See analysis details on SonarQube Cloud