pydocai

Extract text from PDF documents using pypdfium2 with automatic OCR fallback via pytesseract.

Features

• Fast native text extraction using pypdfium2 (Apache 2.0 licensed)
• Automatic OCR fallback for scanned documents using pytesseract
• Smart detection of sparse/scanned PDFs that need OCR
• Memory-efficient processing with lazy imports and temporary file handling
• Page limit to prevent processing extremely large documents (default: 15 pages)

Installation

pip install pydocai

System Dependencies

This package requires Tesseract OCR to be installed on your system for the OCR fallback feature:

macOS:

brew install tesseract

Ubuntu/Debian:

sudo apt-get install tesseract-ocr

Windows: Download and install from: https://github.com/UB-Mannheim/tesseract/wiki

Usage

from pydocai import extract_pdf_text

# Extract text and save to file
success = extract_pdf_text("document.pdf", "output.txt")

# Or let it auto-generate the output filename
success = extract_pdf_text("document.pdf")  # Creates document_extracted.txt

How it works

1. First attempts native text extraction using pypdfium2
2. Checks if extracted text has sufficient content (>= 5 lines per page)
3. If content is sparse (likely a scanned document), falls back to OCR
4. Saves extracted text to the specified output file

Configuration

You can import and check the default configuration values:

from pydocai import (
    OCR_DPI,           # DPI for OCR rendering (default: 100)
    MAX_PDF_PAGES,     # Maximum pages to process (default: 15)
    MIN_LINES_PER_PAGE # Minimum lines to consider page valid (default: 5)
)

Development

# Clone the repository
git clone https://github.com/catherine/pydocai.git
cd pydocai

# Install in development mode with dev dependencies
pip install -e ".[dev]"

# Run tests
pytest

License

Apache License 2.0 - see LICENSE for details.