PDF Python Reader

A tool for extracting and analyzing text from PDF documents with natural language processing capabilities.

Features

• PDF text extraction using PyMuPDF
• Text tokenization and embedding generation
• OCR for image-based PDFs using Tesseract
• Document classification using pre-trained models
• Pattern-based field extraction from insurance documents
• Embedding storage in CSV format

Installation

# Clone the repository
git clone https://github.com/yourusername/PDF-Python-Reader.git
cd PDF-Python-Reader

# Install dependencies
pip install -r requirements.txt

Usage

Basic PDF Reading

python pdf_reader.py

This will:

1. Read the first page of the default PDF document
2. Generate token embeddings using a pre-trained model
3. Save the embeddings to CSV

Insurance Document Processing

python main.py path/to/sample.pdf

This will:

1. Convert the PDF to images and extract text via OCR
2. Classify the document type
3. Extract key fields like policy numbers and claim information
4. Output a JSON summary

Project Structure

• main.py - Insurance document processor with OCR and classification
• pdf_reader.py - Basic PDF text extraction and embedding generation
• text_embeddings.py - Tokenization and embedding utilities
• helpers/ - Additional utility scripts
• pdf_documents/ - Sample PDF files for testing

Dependencies

• PyMuPDF (fitz)
• PyTesseract
• PDF2Image
• Transformers (Hugging Face)
• Pandas
• PyTorch

License

MIT License