Fix mypy errors in CI: remove unsupported python_version, add types-regex, fix type annotations

- Remove python_version=3.9 from mypy config (no longer supported)
- Install regex and types-regex in lint CI for proper type checking
- Update bpe.pyi Trainer.__init__ to accept list[bytes | bytearray]
- Add explicit type annotation in simple.py for text_bytes_list

Author: neluca

.github/workflows/lint.yml

@@ -24,7 +24,7 @@ jobs:
       - name: Install dependencies
         run: |
           python -m pip install --upgrade pip
-          python -m pip install ruff mypy
+          python -m pip install ruff mypy regex types-regex
       - name: Run ruff (lint)
         run: ruff check tinybpe/ --output-format=github
       - name: Run ruff (format check)

pyproject.toml

@@ -55,7 +55,6 @@ indent-style = "space"
 
 [tool.mypy]
 strict = true
-python_version = "3.9"
 
 [tool.pytest.ini_options]
 testpaths = ["tests"]

tinybpe/bpe.pyi

@@ -22,7 +22,7 @@ class Tokenizer:
 class Trainer:
     """This is a BPE trainer implemented in C ."""
 
-    def __init__(self, list_bytes: list[bytes]): ...
+    def __init__(self, list_bytes: list[bytes | bytearray]): ...
     @property
     def n_merges(self) -> int: ...
     @property

tinybpe/simple.py

@@ -62,7 +62,7 @@ def __init__(
         callback: Callable[[int, int, tuple[int, int], int, int], None] | None = None,
     ) -> None:
         if preprocess is None:
-            text_bytes_list = [text.encode("utf-8")]
+            text_bytes_list: list[bytes | bytearray] = [text.encode("utf-8")]
         else:
             text_bytes_list = preprocess(text)