How c4-job-cheat/server works
Project type and purpose
• This is a Django-based monolithic backend API service (“job_cheat”) composed of multiple Django apps (api, core, cover_letters, interviews, job_search, personas).
• Purpose: expose REST endpoints for job-search workflows—persona generation, cover-letter drafting, interview prep, competency evaluation, and health checks—while orchestrating calls to AI services (ChatGPT, Cohere, Whisper, Gemini), a vector store (Pinecone), and Firebase (Auth, Firestore, Storage).
File structure
• Top-level: manage.py, ASGI/WSGI entrypoints, job_cheat/job_cheat (settings, URLs).
• Apps:
• api: high-level entrypoints and URL routing
• core: shared models, serializers, data, and “services” layer invoking AI or storage backends
• cover_letters, interviews, job_search, personas: domain-specific views, serializers, services
• docs/: API specifications, database schema, feature/architecture notes.
• Configuration: pyproject.toml, requirements.txt, .python-version, .gitignore.
• Testing: each service under core/services/tests and per-app test modules.
Architecture insights (from docs/miscl/ARCHITECTURE.md & database-schema.md)
• Single Django process served via ASGI (uvicorn/daphne) or WSGI (gunicorn).
• PostgreSQL (or similar relational DB) for core models.
• Firebase Firestore for persona/user data sync and Storage for file assets.
• Pinecone for similarity search embeddings.
• External AI endpoints: OpenAI, Cohere, Google Gemini, Whisper.
Summary
• Layer in domain apps and async processes (RAG embedding jobs).
• Django app and service is represented.
• With technology names and endpoints per docs/api-doc.
• Validate against docs/miscl/ARCHITECTURE.md and database-schema.md to capture data relationships.