Document AI Toolkit

Production-ready document processing toolkit with AI capabilities for text extraction, table detection, OCR, entity recognition, and document classification.

Features

• Multi-Format Support: PDF, DOCX, HTML, Markdown, TXT, images, and more
• Text Extraction: Simple, layout-aware, and structured extraction modes
• Table Detection: Automatic table detection with cell-level extraction
• OCR Integration: Built-in OCR support with Tesseract, EasyOCR, PaddleOCR
• Entity Extraction: Named entity recognition (persons, organizations, dates, etc.)
• Document Classification: Automatic document type detection
• Layout Analysis: Detect headers, footers, paragraphs, lists, and more
• Zero Dependencies Core: Core functionality works without heavy dependencies

Installation

pip install document-ai-toolkit          # Core
pip install document-ai-toolkit[pdf]     # PDF support
pip install document-ai-toolkit[ocr]     # OCR support
pip install document-ai-toolkit[full]    # All features

Quick Start

from document_ai_toolkit import DocumentProcessor, ProcessingConfig

# Basic processing
processor = DocumentProcessor()
result = processor.process("document.pdf")

print(result.document.content)
print(f"Pages: {result.document.metadata.page_count}")

# With tables
config = ProcessingConfig(extract_tables=True)
processor = DocumentProcessor(config)
result = processor.process("report.pdf")

for table in result.document.tables:
    print(table.to_dict())

# Classification
from document_ai_toolkit import DocumentClassifier
classifier = DocumentClassifier()
result = classifier.classify("document.pdf")
print(f"Type: {result.document_type.value} ({result.confidence:.0%})")

# Comparison
from document_ai_toolkit import DocumentComparator
comparator = DocumentComparator()
result = comparator.compare("v1.docx", "v2.docx")
print(f"Similarity: {result.similarity_score:.0%}")

Supported Formats

License

MIT License - Pranay M