neuradoc 0.1.2

A Python package for parsing and transforming various document formats into LLM-ready data with element classification capabilities

NeuraDoc

NeuraDoc is a Python package for parsing and transforming various document formats into LLM-ready data with element classification capabilities. The library intelligently extracts and classifies content from documents for AI/ML workflows.

Features

• Multi-format Support: Parse at least 10 different document types (PDF, Word, TXT, etc.)
• Element Extraction: Extract text, images, tables, and diagrams from documents
• Classification: Classify document elements by type
• Smart Positioning: Position and organize extracted elements intelligently
• LLM Integration: Convert extracted data to LLM-ready formats (tokenized structures)
• Memory Efficiency: Optimized for processing large documents
• Configurable Parsing: Control extraction behavior with custom configurations
• Parsing Profiles: Use predefined profiles for different extraction needs (fast, detailed, etc.)
• Batch Processing: Process multiple documents with consistent settings
• Performance Metrics: Get detailed processing statistics and timing information

Supported Document Formats

NeuraDoc supports the following document formats:

• PDF (.pdf)
• Microsoft Word (.docx, .doc)
• Plain Text (.txt)
• Microsoft Excel (.xlsx, .xls)
• HTML (.html, .htm)
• XML (.xml)
• Images (.jpg, .jpeg, .png, .gif)
• Microsoft PowerPoint (.pptx, .ppt)
• CSV (.csv)
• JSON (.json)
• Markdown (.md)

Installation

Basic Installation

pip install neuradoc

Installation with Optional Dependencies

# Install with OCR support
pip install neuradoc[ocr]

# Install with advanced table extraction
pip install neuradoc[tables]

# Install with NLP capabilities
pip install neuradoc[nlp]

# Install with transformer model support
pip install neuradoc[transformers]

# Install with web interface
pip install neuradoc[web]

# Install with all optional dependencies
pip install neuradoc[ocr,tables,nlp,transformers,web]

Quick Start

Basic Usage

import neuradoc

# Load and parse a document
doc = neuradoc.load_document("path/to/your/document.pdf")

# Get all text content
text = doc.get_text_content()

# Get tables
tables = doc.get_tables()

# Get images
images = doc.get_images()

# Save extracted content in different formats
doc.save("output.json", format="json")
doc.save("output.md", format="markdown")
doc.save("output.txt", format="text")

Advanced Usage

import neuradoc
from neuradoc.models.element import ElementType
from neuradoc.transformers.llm_transformer import chunk_document

# Load document
doc = neuradoc.load_document("document.docx")

# Filter elements by type
headings = doc.get_elements_by_type(ElementType.HEADING)
code_blocks = doc.get_elements_by_type(ElementType.CODE)

# Transform document into chunks for LLM processing
chunks = chunk_document(doc, max_chunk_size=1000, overlap=100)

# Process chunks with your LLM
for chunk in chunks:
    # Process each chunk with your LLM implementation
    print(f"Chunk: {len(chunk)} characters")

Web Interface

NeuraDoc includes a web interface for document processing:

# Install web dependencies
pip install neuradoc[web]

# Run the web server
python -m neuradoc.web.app

Contributing

Contributions are welcome! Please feel free to submit a Pull Request.

License

This project is licensed under the MIT License - see the LICENSE file for details.