feat: add batch processing for image directories

[Chirudeva-Reddy]

Summary

Closes #2.

This PR adds directory-level batch processing for handwriting OCR transcription:

• Batch Processing Module (batch_process.py):
  • get_image_files(directory): Discovers and sorts all supported image formats (.jpg, .jpeg, .png, case-insensitive).
  • process_image(image_path, ...): Chains line detection with measure_text_lines(), crops and saves text lines with crop_and_save_lines(), transcribes lines, cleans up temporary crop files, and exports transcriptions to .txt and .pdf files named directly after the source image stem (output/<image_stem>.txt and output/<image_stem>.pdf) as well as raw/NLP variants.
  • batch_process(directory, ...): Processes an entire folder of handwriting images.
  • Safe intermediate crop cleanup via cleanup_line_crops() preserving final output artifacts.
  • export_to_pdf(): Robust Unicode punctuation sanitization (em-dashes, en-dashes, typographical quotes, bullets, ellipses) preventing FPDFUnicodeEncodingException under Latin-1 fonts.
• CLI & Pipeline Integration (main.py):
  • Adds argparse to main.py supporting positional directory/image path, --folder DIR / -f, --image FILE / -i, --output DIR / -o, and --model-dir PATH / -m.
• Unit Tests (tests/test_batch_process.py):
  • 9 automated tests covering file discovery, non-existent directories, export formatting, Unicode punctuation handling, custom transcription mock pipelines, end-to-end single image processing with exact naming checks, and multi-image directory batching.
• Code Quality:
  • Passes Flake8 checks without unused imports or line-length violations.
• Documentation:
  • Updated README.md setup & usage instructions with single image and batch directory processing commands.

Verification

• uv run pytest: All 13 tests pass cleanly (test_batch_process.py and test_count_lines.py).
• uv run flake8: 0 errors/warnings on new code.

[estherg12]

Thank you very much for your interest and collaboration!!

Everthing seems fine.