Add type hints to foundation submodules (tools, cli, corpus)

pyproject.toml

@@ -363,3 +363,81 @@ basepython = "python"
 deps = ["ruff"]
 commands = [["ruff", "check", "pythainlp"], ["ruff", "format", "--check", "pythainlp"]]
 skip_install = true
+
+# Mypy configuration
+[tool.mypy]
+python_version = "3.9"
+warn_return_any = true
+warn_unused_configs = true
+disallow_untyped_defs = false
+disallow_incomplete_defs = false
+check_untyped_defs = true
+disallow_untyped_decorators = false
+no_implicit_optional = true
+warn_redundant_casts = true
+warn_unused_ignores = true
+warn_no_return = true
+warn_unreachable = true
+strict_equality = true
+show_error_codes = true
+show_column_numbers = true
+pretty = true
+
+# Per-module options for stricter checking on foundation modules
+[[tool.mypy.overrides]]
+module = [
+    "pythainlp.tools.*",
+    "pythainlp.cli.*",
+]
+disallow_untyped_defs = true
+disallow_incomplete_defs = true
+
+# Ignore missing imports for optional dependencies
+[[tool.mypy.overrides]]
+module = [
+    "attacut.*",
+    "bpemb.*",
+    "budoux.*",
+    "deepcut.*",
+    "emoji.*",
+    "epitran.*",
+    "esupar.*",
+    "fairseq.*",
+    "fastai.*",
+    "fastcoref.*",
+    "gensim.*",
+    "huggingface_hub.*",
+    "icu.*",
+    "khamyo.*",
+    "khanaa.*",
+    "multiel.*",
+    "nlpo3.*",
+    "nltk.*",
+    "numpy.*",
+    "onnxruntime.*",
+    "oskut.*",
+    "pandas.*",
+    "panphon.*",
+    "phunspell.*",
+    "pycrfsuite.*",
+    "pyicu.*",
+    "sacremoses.*",
+    "sefr_cut.*",
+    "sentencepiece.*",
+    "sentence_transformers.*",
+    "spacy.*",
+    "spacy_thai.*",
+    "ssg.*",
+    "symspellpy.*",
+    "thai_nner.*",
+    "tltk.*",
+    "torch.*",
+    "tqdm.*",
+    "transformers.*",
+    "ufal.chu_liu_edmonds.*",
+    "word2word.*",
+    "wtpsplit.*",
+    "wunsen.*",
+    "yaml.*",
+]
+ignore_missing_imports = true

pythainlp/cli/__init__.py

@@ -22,7 +22,7 @@
 CLI_NAME = "thainlp"
 
 
-def make_usage(command: str) -> dict:
+def make_usage(command: str) -> dict[str, str]:
     prog = f"{CLI_NAME} {command}"
 
     return {"prog": prog, "usage": f"{prog} [options]"}

pythainlp/cli/benchmark.py

@@ -3,22 +3,28 @@
 # SPDX-FileType: SOURCE
 # SPDX-License-Identifier: Apache-2.0
 
+from __future__ import annotations
+
 import argparse
 import json
 import os
+from typing import TYPE_CHECKING
 
 from pythainlp import cli
 from pythainlp.tools import safe_print
 
+if TYPE_CHECKING:
+    from collections.abc import Sequence
+
 
-def _read_file(path):
+def _read_file(path: str) -> list[str]:
     with open(path, encoding="utf-8") as f:
         lines = (r.strip() for r in f.readlines())
     return list(lines)
 
 
 class App:
-    def __init__(self, argv):
+    def __init__(self, argv: Sequence[str]) -> None:
         parser = argparse.ArgumentParser(
             prog="benchmark",
             description=(
@@ -45,8 +51,8 @@ def __init__(self, argv):
 
 
 class WordTokenizationBenchmark:
-    def __init__(self, name, argv):
-        parser = argparse.ArgumentParser(**cli.make_usage("benchmark " + name))
+    def __init__(self, name: str, argv: Sequence[str]) -> None:
+        parser = argparse.ArgumentParser(**cli.make_usage("benchmark " + name))  # type: ignore[arg-type]
 
         parser.add_argument(
             "--input-file",

pythainlp/cli/data.py

@@ -4,14 +4,20 @@
 """Command line for PyThaiNLP's dataset/corpus management.
 """
 
+from __future__ import annotations
+
 import argparse
+from typing import TYPE_CHECKING
 
 from pythainlp import corpus
 from pythainlp.tools import get_pythainlp_data_path
 
+if TYPE_CHECKING:
+    from collections.abc import Sequence
+
 
 class App:
-    def __init__(self, argv):
+    def __init__(self, argv: Sequence[str]) -> None:
         parser = argparse.ArgumentParser(
             prog="data",
             description="Manage dataset/corpus.",
@@ -43,7 +49,7 @@ def __init__(self, argv):
         args = parser.parse_args(argv[2:3])
         getattr(self, args.subcommand)(argv)
 
-    def get(self, argv):
+    def get(self, argv: Sequence[str]) -> None:
         parser = argparse.ArgumentParser(
             description="Download a dataset",
             usage="thainlp data get <dataset_name>",
@@ -59,7 +65,7 @@ def get(self, argv):
         else:
             print("Not found.")
 
-    def rm(self, argv):
+    def rm(self, argv: Sequence[str]) -> None:
         parser = argparse.ArgumentParser(
             description="Remove a dataset",
             usage="thainlp data rm <dataset_name>",
@@ -75,7 +81,7 @@ def rm(self, argv):
         else:
             print("Not found.")
 
-    def info(self, argv):
+    def info(self, argv: Sequence[str]) -> None:
         parser = argparse.ArgumentParser(
             description="Print information about a dataset",
             usage="thainlp data info <dataset_name>",
@@ -92,14 +98,14 @@ def info(self, argv):
         else:
             print("Not found.")
 
-    def catalog(self, argv):
+    def catalog(self, argv: Sequence[str]) -> None:
         """Print dataset/corpus available for download."""
-        corpus_db = corpus.get_corpus_db(corpus.corpus_db_url())
-        corpus_db = corpus_db.json()
-        corpus_names = sorted(corpus_db.keys())
+        corpus_db_response = corpus.get_corpus_db(corpus.corpus_db_url())
+        corpus_db_dict: dict[str, dict[str, str]] = corpus_db_response.json()  # type: ignore[union-attr]
+        corpus_names = sorted(corpus_db_dict.keys())
         print("Dataset/corpus available for download:")
         for name in corpus_names:
-            print(f"- {name} {corpus_db[name]['latest_version']}", end="")
+            print(f"- {name} {corpus_db_dict[name]['latest_version']}", end="")
             corpus_info = corpus.get_corpus_db_detail(name)
             if corpus_info:
                 print(f"  (Local: {corpus_info['version']})")
@@ -111,6 +117,6 @@ def catalog(self, argv):
             "Example: thainlp data get crfcut\n"
         )
 
-    def path(self, argv):
+    def path(self, argv: Sequence[str]) -> None:
         """Print path of local dataset."""
         print(get_pythainlp_data_path())

pythainlp/cli/misspell.py

@@ -2,15 +2,21 @@
 # SPDX-FileType: SOURCE
 # SPDX-License-Identifier: Apache-2.0
 
+from __future__ import annotations
+
 import argparse
 import os
 import random
+from typing import TYPE_CHECKING
 
 from pythainlp.tools.misspell import misspell
 
+if TYPE_CHECKING:
+    from collections.abc import Sequence
+
 
 class App:
-    def __init__(self, argv):
+    def __init__(self, argv: Sequence[str]) -> None:
         parser = argparse.ArgumentParser(
             prog="misspell",
             description="Generate misspelled texts from a given file.",

pythainlp/cli/soundex.py

@@ -6,14 +6,20 @@
 It takes input text from the command line.
 """
 
+from __future__ import annotations
+
 import argparse
+from typing import TYPE_CHECKING
 
 from pythainlp.soundex import DEFAULT_SOUNDEX_ENGINE, soundex
 from pythainlp.tools import safe_print
 
+if TYPE_CHECKING:
+    from collections.abc import Sequence
+
 
 class App:
-    def __init__(self, argv):
+    def __init__(self, argv: Sequence[str]) -> None:
         parser = argparse.ArgumentParser(
             prog="soundex",
             description="Convert a text to its sound-based index.",

pythainlp/cli/tag.py

@@ -4,16 +4,25 @@
 """Command line for PyThaiNLP's taggers.
 """
 
+from __future__ import annotations
+
 import argparse
+from typing import TYPE_CHECKING
 
 from pythainlp import cli
 from pythainlp.tag import pos_tag
 from pythainlp.tools import safe_print
 
+if TYPE_CHECKING:
+    from collections.abc import Callable, Sequence
+
 
 class SubAppBase:
-    def __init__(self, name, argv):
-        parser = argparse.ArgumentParser(**cli.make_usage("tag " + name))
+    separator: str
+    run: Callable[[list[str]], list[tuple[str, str]]]
+
+    def __init__(self, name: str, argv: Sequence[str]) -> None:
+        parser = argparse.ArgumentParser(**cli.make_usage("tag " + name))  # type: ignore[arg-type]
         parser.add_argument(
             "text",
             type=str,
@@ -39,15 +48,15 @@ def __init__(self, name, argv):
 
 
 class POSTaggingApp(SubAppBase):
-    def __init__(self, *args, **kwargs):
+    def __init__(self, *args: str, **kwargs: str) -> None:
         self.separator = "|"
         self.run = pos_tag
 
         super().__init__(*args, **kwargs)
 
 
 class App:
-    def __init__(self, argv):
+    def __init__(self, argv: Sequence[str]) -> None:
         parser = argparse.ArgumentParser(
             prog="tag",
             description="Annotate a text with linguistic information",
@@ -72,6 +81,6 @@ def __init__(self, argv):
         argv = argv[3:]
 
         if tag_type == "pos":
-            POSTaggingApp("Part-of-Speech tagging", argv)
+            POSTaggingApp("Part-of-Speech tagging", argv)  # type: ignore[arg-type]
         else:
             print(f"Tag type not available: {tag_type}")

pythainlp/cli/tokenize.py

@@ -3,7 +3,10 @@
 # SPDX-License-Identifier: Apache-2.0
 """Command line for PyThaiNLP's tokenizers."""
 
+from __future__ import annotations
+
 import argparse
+from typing import TYPE_CHECKING
 
 from pythainlp import cli
 from pythainlp.tokenize import (
@@ -16,15 +19,22 @@
 )
 from pythainlp.tools import safe_print
 
+if TYPE_CHECKING:
+    from collections.abc import Callable, Sequence
+
 DEFAULT_SENT_TOKEN_SEPARATOR = "@@"  # noqa: S105
 DEFAULT_SUBWORD_TOKEN_SEPARATOR = "/"  # noqa: S105
 DEFAULT_SYLLABLE_TOKEN_SEPARATOR = "~"  # noqa: S105
 DEFAULT_WORD_TOKEN_SEPARATOR = "|"  # noqa: S105
 
 
 class SubAppBase:
-    def __init__(self, name, argv):
-        parser = argparse.ArgumentParser(**cli.make_usage("tokenize " + name))
+    separator: str
+    algorithm: str
+    run: Callable[..., list[str]]
+
+    def __init__(self, name: str, argv: Sequence[str]) -> None:
+        parser = argparse.ArgumentParser(**cli.make_usage("tokenize " + name))  # type: ignore[arg-type]
         parser.add_argument(
             "text",
             type=str,
@@ -74,7 +84,7 @@ def __init__(self, name, argv):
 
 
 class WordTokenizationApp(SubAppBase):
-    def __init__(self, *args, **kwargs):
+    def __init__(self, *args: str, **kwargs: str) -> None:
         self.keep_whitespace = True
         self.algorithm = DEFAULT_WORD_TOKENIZE_ENGINE
         self.separator = DEFAULT_WORD_TOKEN_SEPARATOR
@@ -83,7 +93,7 @@ def __init__(self, *args, **kwargs):
 
 
 class SentenceTokenizationApp(SubAppBase):
-    def __init__(self, *args, **kwargs):
+    def __init__(self, *args: str, **kwargs: str) -> None:
         self.keep_whitespace = True
         self.algorithm = DEFAULT_SENT_TOKENIZE_ENGINE
         self.separator = DEFAULT_SENT_TOKEN_SEPARATOR
@@ -92,7 +102,7 @@ def __init__(self, *args, **kwargs):
 
 
 class SubwordTokenizationApp(SubAppBase):
-    def __init__(self, *args, **kwargs):
+    def __init__(self, *args: str, **kwargs: str) -> None:
         self.keep_whitespace = True
         self.algorithm = DEFAULT_SUBWORD_TOKENIZE_ENGINE
         self.separator = DEFAULT_SUBWORD_TOKEN_SEPARATOR
@@ -101,7 +111,7 @@ def __init__(self, *args, **kwargs):
 
 
 class App:
-    def __init__(self, argv):
+    def __init__(self, argv: Sequence[str]) -> None:
         parser = argparse.ArgumentParser(
             prog="tokenize",
             description="Break a text into small units (tokens).",
@@ -137,10 +147,10 @@ def __init__(self, argv):
 
         argv = argv[3:]
         if token_type.startswith("w"):
-            WordTokenizationApp("word", argv)
+            WordTokenizationApp("word", argv)  # type: ignore[arg-type]
         elif token_type.startswith("su"):
-            SubwordTokenizationApp("subword", argv)
+            SubwordTokenizationApp("subword", argv)  # type: ignore[arg-type]
         elif token_type.startswith("se"):
-            SentenceTokenizationApp("sent", argv)
+            SentenceTokenizationApp("sent", argv)  # type: ignore[arg-type]
         else:
             safe_print(f"Token type not available: {token_type}")