Epstein Unredactor
Home
Architecture
API
GitHub
Home
Architecture
API
GitHub
  • Getting Started

    • Setup & Deployment
    • Local Development
    • Production Deployment
    • /setup-and-deployment/TROUBLESHOOTING.html
  • Architecture

    • Epstein Unredactor — Architecture Overview
  • Redaction Processing

    • Redaction Processing — Backend Logic
    • ProcessRedactions.py
    • BoxDetector.py Documentation
    • SurroundingWordWidth.py Documentation
    • Scale & Size Detection
    • Artifact Visualizer — Documentation
    • width_calculator.py
    • extract_fonts — Font Detection Module
  • Frontend Implementation

    • Frontend — JavaScript Module Reference
    • State Management — state.js
    • PDF Viewer — pdf-viewer.js
    • API & Candidate Logic — api.js
    • UI Events — ui-events.js
    • WebGL Mask — webgl-mask.js
    • Formatting Bridge — text-tool.js
  • API Reference

    • API Reference

Setup & Deployment

This section covers everything needed to run the Epstein Unredactor, from local development to production server deployment.

Contents

DocumentDescription
Local DevelopmentSetting up a dev environment on Windows or Linux
Production DeploymentFull Linux server setup with Gunicorn, Nginx, and SSL

Prerequisites

  • Python 3.10+
  • pip and venv
  • System libraries: libgl1, libglib2.0-0, libharfbuzz-dev (Linux)
  • Microsoft Core Fonts (for accurate width calculation)

Dependencies

PackagePurpose
djangoWeb framework
opencv-python-headlessImage analysis (box detection)
numpyArray operations for image processing
pymupdfPDF parsing and image extraction
pillowImage format conversions, mask generation
uharfbuzzPrecision text shaping with kerning/ligature support
gunicornProduction WSGI server
Edit this page
Last Updated: 3/25/26, 11:14 AM
Contributors: JaguarM
Next
Local Development