GeoExtract: Open-Source Geological Report Data Extraction System

GeoExtract is a Python-based CLI tool and web interface that uses OCR + LLMs to automatically extract structured geological data from legacy PDF reports (1950s-2000s). The tool handles scanned documents, typed reports, and mixed formats, outputting standardized geospatial data formats.

Features

• Multi-format Input: PDFs, scanned images (TIF, JPG, PNG)
• Advanced OCR: PaddleOCR + Tesseract with layout preservation
• LLM Extraction: Local Ollama or OpenAI API integration
• Geological Focus: Specialized for mining/exploration reports
• Multiple Outputs: GeoJSON, CSV, GeoPackage, JSON-LD
• Web Interface: Streamlit-based UI for easy document processing
• API Server: FastAPI with async job processing
• Batch Processing: Handle directories of documents

Quick Start

Installation

# Install with Poetry
git clone https://github.com/your-org/geoextract.git
cd geoextract
poetry install

# Or install with pip
pip install geoextract

Basic Usage

# Process a single PDF
geoextract process report.pdf --output results/

# Batch process directory
geoextract batch input_dir/ --output results/ --format geojson,csv

# Start web interface
geoextract ui

# Start API server
geoextract serve --port 8000

Supported Data Types

• Location Data: Coordinates (decimal degrees, DMS, UTM, township-range-section)
• Geological Data: Rock types, mineral occurrences, assay results
• Drilling Data: Hole IDs, depth intervals, sample data
• Temporal Data: Report dates, survey dates
• Metadata: Report titles, authors, references

Output Formats

• GeoJSON: For immediate mapping with proper CRS metadata
• CSV: Tabular data for analysis
• GeoPackage: For GIS integration
• JSON-LD: With schema.org geological vocabulary

Configuration

# Set LLM provider
geoextract config --set llm.provider=ollama
geoextract config --set llm.model=llama3.1:8b

# Configure OCR engine
geoextract config --set ocr.engine=paddle
geoextract config --set ocr.confidence_threshold=0.8

Development

# Install development dependencies
poetry install --with dev

# Run tests
pytest

# Format code
black geoextract/
isort geoextract/

# Type checking
mypy geoextract/

License

MIT License - see LICENSE file for details.

Contributing

Contributions welcome! Please see CONTRIBUTING.md for guidelines.

Support

• GitHub Issues: Report bugs and request features
• Documentation: Full documentation
• Community: Join our discussions