Fix type hints for Python 3.9 runtime compatibility

pythainlp/benchmarks/word_tokenization.py

@@ -196,7 +196,7 @@ def compute_stats(ref_sample: str, raw_sample: str) -> dict:
     }
 
 
-def _binary_representation(txt: str, verbose: bool = False):
+def _binary_representation(txt: str, verbose: bool = False) -> np.ndarray:
     """Transform text into {0, 1} sequence.
 
     where (1) indicates that the corresponding character is the beginning of

pythainlp/coref/core.py

@@ -8,7 +8,7 @@
 
 def coreference_resolution(
     texts: list[str], model_name: str = "han-coref-v1.0", device: str = "cpu"
-):
+) -> list[dict]:
     """Coreference Resolution
 
     :param List[str] texts: list of texts to apply coreference resolution to

pythainlp/spell/core.py

@@ -8,12 +8,16 @@
 
 import itertools
 from functools import lru_cache
+from typing import TYPE_CHECKING
 
 from pythainlp.spell import DEFAULT_SPELL_CHECKER
 
+if TYPE_CHECKING:
+    from pythainlp.spell.pn import NorvigSpellChecker
+
 
 @lru_cache
-def default_spell_checker():
+def default_spell_checker() -> "NorvigSpellChecker":
     """Lazy load default spell checker with cache"""
     return DEFAULT_SPELL_CHECKER()
 

pythainlp/spell/symspellpy.py

@@ -33,7 +33,7 @@
 _load_lock = threading.Lock()  # Thread safety for lazy loading
 
 
-def _get_sym_spell():
+def _get_sym_spell() -> SymSpell:
     """Lazy load the symspell instance.
 
     This function uses a lock to ensure thread-safe initialization.

pythainlp/tag/perceptron.py

@@ -32,36 +32,36 @@
 _TUD_TAGGER = None
 
 
-def _orchid_tagger():
+def _orchid_tagger() -> PerceptronTagger:
     global _ORCHID_TAGGER
     if not _ORCHID_TAGGER:
         _ORCHID_TAGGER = PerceptronTagger(path=_ORCHID_PATH)
     return _ORCHID_TAGGER
 
 
-def _pud_tagger():
+def _pud_tagger() -> PerceptronTagger:
     global _PUD_TAGGER
     if not _PUD_TAGGER:
         _PUD_TAGGER = PerceptronTagger(path=_PUD_PATH)
     return _PUD_TAGGER
 
 
-def _blackboard_tagger():
+def _blackboard_tagger() -> PerceptronTagger:
     global _BLACKBOARD_TAGGER
     if not _BLACKBOARD_TAGGER:
         path = get_corpus_path(_BLACKBOARD_NAME)
         _BLACKBOARD_TAGGER = PerceptronTagger(path=path)
     return _BLACKBOARD_TAGGER
 
 
-def _tdtb():
+def _tdtb() -> PerceptronTagger:
     global _TDTB_TAGGER
     if not _TDTB_TAGGER:
         _TDTB_TAGGER = PerceptronTagger(path=_TDTB_PATH)
     return _TDTB_TAGGER
 
 
-def _tud_tagger():
+def _tud_tagger() -> PerceptronTagger:
     global _TUD_TAGGER
     if not _TUD_TAGGER:
         _TUD_TAGGER = PerceptronTagger(path=_TUD_PATH)

pythainlp/tag/thainer.py

@@ -201,5 +201,5 @@ def get_ner(
         return sent_ner
 
     @staticmethod
-    def __extract_features(doc: list[str]) -> list[dict[str, str | bool]]:
+    def __extract_features(doc: list[str]) -> list[dict[str, Union[str, bool]]]:
         return [_doc2features(doc, i) for i in range(len(doc))]

pythainlp/tag/unigram.py

@@ -33,23 +33,23 @@
 _TUD_TAGGER = None
 
 
-def _orchid_tagger():
+def _orchid_tagger() -> dict:
     global _ORCHID_TAGGER
     if not _ORCHID_TAGGER:
         with open(_ORCHID_PATH, encoding="utf-8-sig") as fh:
             _ORCHID_TAGGER = json.load(fh)
     return _ORCHID_TAGGER
 
 
-def _pud_tagger():
+def _pud_tagger() -> dict:
     global _PUD_TAGGER
     if not _PUD_TAGGER:
         with open(_PUD_PATH, encoding="utf-8-sig") as fh:
             _PUD_TAGGER = json.load(fh)
     return _PUD_TAGGER
 
 
-def _blackboard_tagger():
+def _blackboard_tagger() -> dict:
     global _BLACKBOARD_TAGGER
     if not _BLACKBOARD_TAGGER:
         path = get_corpus_path(_BLACKBOARD_NAME)
@@ -58,15 +58,15 @@ def _blackboard_tagger():
     return _BLACKBOARD_TAGGER
 
 
-def _thai_tdtb():
+def _thai_tdtb() -> dict:
     global _TDTB_TAGGER
     if not _TDTB_TAGGER:
         with open(_TDTB_PATH, encoding="utf-8-sig") as fh:
             _TDTB_TAGGER = json.load(fh)
     return _TDTB_TAGGER
 
 
-def _tud_tagger():
+def _tud_tagger() -> dict:
     global _TUD_TAGGER
     if not _TUD_TAGGER:
         with open(_TUD_PATH, encoding="utf-8-sig") as fh:

pythainlp/tokenize/__init__.py

@@ -3,6 +3,7 @@
 # SPDX-License-Identifier: Apache-2.0
 """Tokenizers at different levels of linguistic analysis.
 """
+from __future__ import annotations
 
 __all__ = [
     "thai2fit_tokenizer",

pythainlp/tokenize/han_solo.py

@@ -10,6 +10,7 @@
 
 import threading
 from importlib.resources import as_file, files
+from typing import Optional
 
 try:
     import pycrfsuite
@@ -47,7 +48,7 @@ def _get_tagger() -> pycrfsuite.Tagger:
 class Featurizer:
     #  This class from ssg at https://github.com/ponrawee/ssg.
 
-    def __init__(self, N: int = 2, sequence_size: int = 1, delimiter: str | None = None) -> None:
+    def __init__(self, N: int = 2, sequence_size: int = 1, delimiter: Optional[str] = None) -> None:
         self.N = N
         self.delimiter = delimiter
         self.radius = N + sequence_size

pythainlp/transliterate/core.py

@@ -3,6 +3,8 @@
 # SPDX-License-Identifier: Apache-2.0
 from __future__ import annotations
 
+from typing import Callable
+
 DEFAULT_ROMANIZE_ENGINE = "royin"
 DEFAULT_TRANSLITERATE_ENGINE = "thaig2p"
 DEFAULT_PRONUNCIATE_ENGINE = "w2p"
@@ -70,7 +72,7 @@ def romanize(
 
     """
 
-    def select_romanize_engine(engine: str):
+    def select_romanize_engine(engine: str) -> Callable[[str], str]:
         if engine == "thai2rom":
             from pythainlp.transliterate.thai2rom import romanize
         elif engine == "thai2rom_onnx":

pythainlp/word_vector/core.py

@@ -43,7 +43,7 @@ def __init__(self, model_name: str = "thai2fit_wv") -> None:
         """
         self.load_wordvector(model_name)
 
-    def load_wordvector(self, model_name: str):
+    def load_wordvector(self, model_name: str) -> None:
         """Load word vector model.
 
         :param str model_name: model name