Transcript Index Manager

This application provides functionality to manage and search through transcript data. It supports both building indexes on-the-fly and working with pre-generated flat index files.

Installation

1. Clone the repository
2. Install dependencies:

   pip install -r requirements.txt

Usage

Command Line Interface

The application provides a CLI for managing index files:

1. Generate an index file from transcript data:

   python -m app.cli generate-index /path/to/data/dir explore-index.json.gz

2. Validate an existing index file:

   python -m app.cli validate-index explore-index.json.gz

Running the Web Application

You can run the web application in two modes:

1. Building index on-the-fly (default):

   export FLASK_APP=app
   flask run

2. Using a pre-generated index file:

   export FLASK_APP=app
   export INDEX_FILE=/path/to/explore-index.json.gz
   flask run

Environment Variables

• INDEX_FILE: Path to a pre-generated index file (optional)
• FLASK_APP: Set to "app" to run the Flask application
• FLASK_ENV: Set to "development" for development mode
• SECRET_KEY: Secret key for Flask sessions
• POSTHOG_API_KEY: PostHog API key for analytics (optional)
• POSTHOG_HOST: PostHog host URL (optional)
• DISABLE_ANALYTICS: Set to "true" to disable analytics

Directory Structure

• app/: Main application package
  • services/: Core services including index management
  • routes/: Web application routes
  • templates/: HTML templates
  • static/: Static files
  • cli.py: Command line interface
• data/: Data directory
  • json/: Transcript JSON files
  • audio/: Audio files

Project Structure

• app/ - Main application code
  • routes/ - Flask route definitions
  • services/ - Business logic and data services
  • static/ - CSS, JavaScript, and images
  • templates/ - HTML templates

Data Format

The application expects JSON transcript files with the following structure:

[
  {
    "start": 0.0,
    "text": "Transcript text segment"
  },
  ...
]

License

This project is licensed under the MIT License. The data accessible through this application is licensed under the ivrit.ai license.

Contact

For support, help, ideas, and contributions, please contact us at info@ivrit.ai.