How noah-duncan/attendee works
Project Type and Purpose
• Project Type: A Django-based backend API platform.
• Purpose: To manage meeting bots that join meetings, capture transcripts, handle media streams, and interact with third-party services like Zoom (via OAuth), Deepgram (for transcription), Google Meet, and Microsoft Teams.
• Key Use Cases:
• Creating bots for meetings via API requests.
• Managing bot lifecycle and media stream processing.
• Retrieving transcripts and meeting recordings.
File Structure
• Top-Level Directories:
• accounts: Handles user registration, authentication, and account-specific operations (templates, forms, models, views).
• attendee: Contains Django project settings, URL configuration, and add environment-specific settings (development, production, testing).
• bots: Contains the core logic related to the meeting bots functionality. This includes:
– Bot controllers (for handling media streams, audio input/output, transcription via Deepgram).
– Bot adapters (separate modules for Zoom, Google Meet, Teams, and web bots).
– Management commands to launch, run, and manage bots.
– Celery tasks and other asynchronous job handling.
• docs, static, and templates: Provide documentation and static assets that support the UI elements (if any) and potential API response templates.
• Configuration Files:
• Dockerfile, docker-compose files (e.g., dev.docker-compose.yaml): Indicate containerized deployment.
• Procfile, heroku.yml, and Makefile: Show deployment and build/deployment automation.
• Environment settings (attendee/settings/*): Reflect best practice separation of configuration for development, staging, and production.
• Architectural Indicators:
• Separation of concerns with distinct Django apps (accounts for user operations, bots for meeting bot operations).
• Usage of management commands and Celery tasks implies asynchronous processing components.
• Presence of migration files indicates a relational database (likely Postgres based on README) as a storage backend.
README for Additional Insights
• The README details how to use the Attendee API including key endpoints such as POST /bots to join a meeting and GET /bots/<id> to monitor state, plus GET /bots/<id>/transcript to get transcripts.
• It outlines external dependencies:
• Postgres and Redis are required for storage and task-queue respectively.
• Zoom OAuth credentials and Deepgram API key for streaming and transcription.
• The instructions for “self-hosting” indicate that the project runs in a single Docker container for the Django app with external services required as separate containers or managed services.