chore: tightened ruff rules

kreuzberg/_tesseract.py

@@ -6,7 +6,7 @@
 from enum import Enum
 from functools import partial
 from os import PathLike
-from typing import Literal, TypeVar, Union, cast
+from typing import Final, Literal, TypeVar, Union, cast
 
 from anyio import CapacityLimiter, create_task_group, to_process
 from anyio import Path as AsyncPath
@@ -23,6 +23,8 @@
 if sys.version_info < (3, 11):  # pragma: no cover
     from exceptiongroup import ExceptionGroup  # type: ignore[import-not-found]
 
+MINIMAL_SUPPORTED_TESSERACT_VERSION: Final[int] = 5
+
 version_ref = {"checked": False}
 
 T = TypeVar("T", bound=Union[Image, PathLike[str], str])
@@ -198,7 +200,7 @@ async def validate_tesseract_version() -> None:
         command = ["tesseract", "--version"]
         result = await run_sync(subprocess.run, command, capture_output=True)
         version_match = re.search(r"tesseract\s+v?(\d+)", result.stdout.decode())
-        if not version_match or int(version_match.group(1)) < 5:
+        if not version_match or int(version_match.group(1)) < MINIMAL_SUPPORTED_TESSERACT_VERSION:
             raise MissingDependencyError("Tesseract version 5 or above is required.")
 
         version_ref["checked"] = True

pyproject.toml

@@ -1,6 +1,6 @@
 [project]
 name = "kreuzberg"
-version = "1.7.0"
+version = "2.0.0"
 description = "A text extraction library supporting PDFs, images, office documents and more"
 readme = "README.md"
 keywords = [
@@ -75,41 +75,40 @@ format.docstring-code-line-length = 120
 format.docstring-code-format = true
 lint.select = [ "ALL" ]
 lint.ignore = [
-  "ANN401",   # Dynamically typed ANY
-  "ASYNC110", # Use Event instead of sleep in async loop
-  "C901",     # Cognitive Complexity
-  "COM812",   # Conflicts with formatter
-  "D100",     # Missing docstring in public module
-  "D104",     # Missing docstring in public package
-  "D107",     # Missing docstring in __init__
-  "D205",     # 1 blank line required between summary line and description
-  "E501",     # Line too long, handled by ruff format
-  "EM",       # Exception messages,
-  "FBT",      # Boolean-typed positional argument in function definition
-  "FIX",      # We allow todo and fixme comments
-  "ISC001",   # Conflicts with formatter
-  "PD011",    # Creates false positives
-  "PGH003",   # Required specific mypy ignores
-  "PLR0911",  # Pylint too many return statements
-  "PLR0912",  # Pylint too many branches
-  "PLR0913",  # Pylint too many arguments
-  "PLR2004",  # Magic variables, we allow them
-  "TD",       # We allow todo and fixme comments
-  "TRY",      # Try except block, rules are too strict
+  "ANN401", # Dynamically typed ANY for kwargs
+  "COM812", # Conflicts with formatter
+  "D100",   # Missing docstring in public module
+  "D104",   # Missing docstring in public package
+  "D107",   # Missing docstring in __init__
+  "D205",   # 1 blank line required between summary line and description
+  "E501",   # Line too long, handled by ruff format
+  "EM",     # Exception messages,
+  "FBT",    # Boolean-typed positional argument in function definition
+  "FIX",    # We allow todo and fixme comments
+  "ISC001", # Conflicts with formatter
+  "TD",     # We allow todo and fixme comments
+  "TRY",    # Try except block, rules are too strict
 ]
 lint.per-file-ignores."tests/**/*.*" = [
   "ARG001",
   "D",
   "N815",
+  "PGH003",
+  "PLR2004",
   "PT006",
   "PT007",
   "PT013",
   "RUF012",
   "S",
 ]
 lint.isort.known-first-party = [ "kreuzberg", "tests" ]
+lint.mccabe.max-complexity = 15
 lint.pydocstyle.convention = "google"
 
+lint.pylint.max-args = 10
+lint.pylint.max-branches = 15
+lint.pylint.max-returns = 10
+
 [tool.pyproject-fmt]
 keep_full_version = true
 max_supported_python = "3.13"