Add typing hints for ocr() function

src/ocrmypdf/api.py

@@ -21,6 +21,7 @@ import sys
 import warnings
 from enum import IntEnum
 from pathlib import Path
+from typing import List, Optional, Dict
 
 from tqdm import tqdm
 
@@ -165,52 +166,52 @@ def create_options(*, input_file, output_file, **kwargs):
 
 
 def ocr(  # pylint: disable=unused-argument
-    input_file,
-    output_file,
+    input_file: os.PathLike,
+    output_file: os.PathLike,
     *,
-    language=None,
-    image_dpi=None,
+    language: List[str] = None,
+    image_dpi: int = None,
     output_type=None,
-    sidecar=None,
-    jobs=None,
-    use_threads=None,
-    title=None,
-    author=None,
-    subject=None,
-    keywords=None,
-    rotate_pages=None,
-    remove_background=None,
-    deskew=None,
-    clean=None,
-    clean_final=None,
-    unpaper_args=None,
-    oversample=None,
-    remove_vectors=None,
-    threshold=None,
-    force_ocr=None,
-    skip_text=None,
-    redo_ocr=None,
-    skip_big=None,
-    optimize=None,
-    jpg_quality=None,
-    png_quality=None,
-    jbig2_lossy=None,
-    jbig2_page_group_size=None,
-    pages=None,
-    max_image_mpixels=None,
-    tesseract_config=None,
-    tesseract_pagesegmode=None,
-    tesseract_oem=None,
+    sidecar: os.PathLike = None,
+    jobs: int = None,
+    use_threads: bool = None,
+    title: str = None,
+    author: str = None,
+    subject: str = None,
+    keywords: str = None,
+    rotate_pages: bool = None,
+    remove_background: bool = None,
+    deskew: bool = None,
+    clean: bool = None,
+    clean_final: bool = None,
+    unpaper_args: str = None,
+    oversample: int = None,
+    remove_vectors: bool = None,
+    threshold: bool = None,
+    force_ocr: bool = None,
+    skip_text: bool = None,
+    redo_ocr: bool = None,
+    skip_big: float = None,
+    optimize: int = None,
+    jpg_quality: int = None,
+    png_quality: int = None,
+    jbig2_lossy: bool = None,
+    jbig2_page_group_size: int = None,
+    pages: str = None,
+    max_image_mpixels: float = None,
+    tesseract_config: List[str] = None,
+    tesseract_pagesegmode: int = None,
+    tesseract_oem: int = None,
     pdf_renderer=None,
-    tesseract_timeout=None,
-    rotate_pages_threshold=None,
+    tesseract_timeout: float = None,
+    rotate_pages_threshold: float = None,
     pdfa_image_compression=None,
-    user_words=None,
-    user_patterns=None,
-    fast_web_view=None,
-    keep_temporary_files=None,
-    progress_bar=None,
-    tesseract_env=None,
+    user_words: os.PathLike = None,
+    user_patterns: os.PathLike = None,
+    fast_web_view: float = None,
+    keep_temporary_files: bool = None,
+    progress_bar: bool = None,
+    tesseract_env: Dict[str, str] = None,
 ):
     """Run OCRmyPDF on one PDF or image.