A document redaction system that helps protect sensitive information in your documents using AI-powered redaction techniques.

The backend is built with Flask and consists of several key components:

1. app.py - Main Application Entry Point

   • Handles all HTTP routes and API endpoints
   • Manages file uploads and document processing
   • Implements email notifications
   • Key endpoints:
     • /email/send - Sends notification emails
     • /document/add - Handles document uploads
     • /documents - Lists all documents
     • /document/hash/<hash> - Retrieves documents by hash
     • /structured - Processes structured data
     • /redact - Handles document redaction

2. redaction_service.py

   • Core redaction logic implementation
   • Handles PDF processing and text extraction
   • Manages redaction patterns and rules

3. ocr_redaction.py

   • OCR (Optical Character Recognition) implementation
   • Processes scanned documents
   • Extracts text from images within PDFs

4. ollamahandler.py

   • Integration with Ollama AI model
   • Handles AI-powered text analysis
   • Manages model interactions and responses

5. preprocessor.py

   • Document preprocessing utilities
   • Text cleaning and normalization
   • Format conversion helpers

6. auto_emailer.py

   • Email notification system
   • Template management
   • Email sending utilities

7. config.py

   • Configuration management
   • Environment variables
   • System settings

8. model.py

   • Database models
   • Data structures
   • Schema definitions

• SQLite database implementation
• Document storage and retrieval
• Hash management for documents

• temp_uploads/ - Temporary file storage
• document_storage/ - Permanent document storage

The frontend is built with React, TypeScript, and Vite, featuring a modern component-based architecture.

1. App.tsx - Root Component

   • Application routing
   • Theme management
   • Global state setup
   • Toast notifications

2. Components Directory (components/)

   • Reusable UI components
   • Theme toggle
   • Form elements
   • Layout components

3. Pages Directory (pages/)

   • Route-based components
   • Main application views
   • Error pages

4. Hooks Directory (hooks/)

   • Custom React hooks
   • API integration hooks
   • State management hooks

5. Types Directory (types/)

   • TypeScript type definitions
   • Interface declarations
   • Type utilities

6. Lib Directory (lib/)

   • Utility functions
   • API clients
   • Helper functions

• Docker
• Docker Compose
• Ollama

Ollama running with llama3.2:latest default model.

1. Clone the repository:

git clone [your-repo-url]
cd Hackfest25-23

2. Start the application:

docker-compose up --build

The application will be available at:

• Frontend: http://localhost:5173
• Backend API: http://localhost:5000

1. Upload Document

POST /document/add
Content-Type: multipart/form-data
Body:
  - file: PDF file
  - email: [email protected]

2. List Documents

GET /documents?[email protected]

3. Get Document by Hash

GET /document/hash/<hash>

4. Process Document

POST /redact
Content-Type: multipart/form-data
Body:
  - files: PDF file(s)
  - method: redaction method

POST /email/send
Content-Type: application/json
Body:
{
  "email": "[email protected]",
  "subject": "Document Verification",
  "contents": "Please verify the redacted document"
}

• File upload validation
• Secure file storage
• Hash-based document retrieval
• Rate limiting
• CORS protection
• Input sanitization

1. Follow PEP 8 style guide
2. Add docstrings to all functions
3. Implement error handling
4. Write unit tests for new features

1. Use TypeScript for type safety
2. Follow React best practices
3. Implement responsive design
4. Maintain component reusability

This project is licensed under the MIT License.