How rahulv1130/codenova works
Codenova is a GitHub-connected codebase navigator with a web UI and terminal client. Users authenticate, add a repository for background indexing, then ask codebase questions, generate documentation, review pull requests, or request verified code edits that can become GitHub PRs. The backend combines API routes and worker tasks with relational, vector, and graph stores. README statements establish the major workflows; source excerpts and the index support some internal calls, but several route and agent implementations were not sampled, so those links are documented or left unverified.
Main entry points: main.py FastAPI app · repo_analyzer.py Ingestion orchestration · tasks.py Celery tasks · vector_store.py Qdrant vectors
Components 30
User clients
- Web app
frontend/src/app/page.tsx - CLI
cli/src/codenova/main.py - CLI API client
cli/src/codenova/api/client.py
API and jobs
- FastAPI app
backend/app/main.py - Repository routes
backend/app/api/routes/repositories.py - Chat routes
backend/app/api/routes/chat.py - Documentation routes
backend/app/api/routes/docs.py - PR routes
backend/app/api/routes/pull_request.py - Edit routes
backend/app/api/routes/code_edit.py - Celery tasks
backend/app/tasks.py
Repository indexing
- Ingestion orchestration
backend/app/agents/repo_analyzer.py - GitHub loading
backend/app/services/github_service.py - AST extraction
backend/app/services/ast_extractor.py - API surface extraction
backend/app/services/api_surface_extractor.py
AI workflows
- Repository chat
backend/app/agents/chat_agent.py - Documentation generation
backend/app/agents/doc_agent.py - PR review
backend/app/agents/pr_agent.py - Edit planner
backend/app/agents/coding_agent/planner.py - Edit agent
backend/app/agents/coding_agent/agentic_editor.py - Edit verification
backend/app/agents/coding_agent/verify.py - Container session
backend/app/services/container_session.py
Knowledge and state
- PostgreSQL
backend/app/core/database.py - Qdrant vectors
backend/app/services/vector_store.py - Neo4j graph
backend/app/services/graph_store.py - Redis status and queue
backend/app/core/redis.py
Other components
- Engineer
- GitHub
- OpenRouter models
- OpenAI embeddings
- Docker sandbox
Connections 44
- Engineer → Web app: uses
README.md - Engineer → CLI: uses
README.md - Web app → FastAPI app: REST and SSE
README.md - CLI → CLI API client: calls
cli/src/codenova/main.py - CLI API client → FastAPI app: HTTP requests
README.md - FastAPI app → Repository routes: registers
backend/app/main.py - FastAPI app → Chat routes: registers
backend/app/main.py - FastAPI app → Documentation routes: registers
backend/app/main.py - FastAPI app → PR routes: registers
backend/app/main.py - FastAPI app → Edit routes: registers
backend/app/main.py - Repository routes → Celery tasks: enqueues ingestion
- Celery tasks → Ingestion orchestration: runs indexing
backend/app/tasks.py - Ingestion orchestration → GitHub loading: loads repository
backend/app/agents/repo_analyzer.py - Ingestion orchestration → AST extraction: extracts structure
backend/app/agents/repo_analyzer.py - Ingestion orchestration → API surface extraction: extracts surfaces
backend/app/agents/repo_analyzer.py - Ingestion orchestration → Qdrant vectors: indexes chunks
backend/app/agents/repo_analyzer.py - Ingestion orchestration → Neo4j graph: writes graph
backend/app/agents/repo_analyzer.py - Chat routes → Repository chat: invokes
backend/app/api/routes/chat.py - Repository chat → Qdrant vectors: searches code
README.md - Repository chat → Neo4j graph: adds graph context
README.md - Repository chat → OpenRouter models: generates answers
README.md - Documentation routes → Documentation generation: invokes
- Documentation generation → Qdrant vectors: retrieves context
README.md - Documentation generation → Neo4j graph: retrieves context
README.md - Documentation generation → OpenRouter models: generates topics
README.md - PR routes → PR review: invokes
- PR review → GitHub loading: fetches PR data
README.md - PR review → Neo4j graph: analyzes impact
README.md - PR review → OpenRouter models: analyzes and synthesizes
README.md - Edit routes → Celery tasks: queues edit
backend/app/api/routes/code_edit.py - Celery tasks → Edit planner: starts planning
backend/app/tasks.py - Edit planner → Qdrant vectors: retrieves context
backend/app/agents/coding_agent/planner.py - Celery tasks → Edit agent: runs edit agent
backend/app/tasks.py - Edit agent → Container session: uses session
backend/app/agents/coding_agent/agentic_editor.py - Celery tasks → Edit verification: verifies changes
backend/app/tasks.py - Edit verification → Container session: runs checks
backend/app/agents/coding_agent/verify.py - Celery tasks → GitHub loading: pushes approved changes
README.md - Edit agent → OpenRouter models: uses coding model
README.md - FastAPI app → PostgreSQL: initializes database
backend/app/main.py - FastAPI app → Redis status and queue: connects
backend/app/main.py - FastAPI app → Qdrant vectors: ensures collection
backend/app/main.py - FastAPI app → Neo4j graph: creates indexes
backend/app/main.py - Ingestion orchestration → OpenAI embeddings: embeds chunks
README.md - Container session → Docker sandbox: runs isolated container
README.md