How redfox1899/gitunderstand works
Project type and purpose
• From the file tree and minimal README, we see:
• A src/api folder with routers (ingest, summary, health) and middleware ⇒ a web/API server (likely FastAPI).
• A templates folder plus static/js ⇒ server-rendered frontend with client-side scripts.
• A core package with modules for cloning, chunking, parsing, AI summarization, formatting, and ingestion ⇒ business-logic layer that processes Git repositories and generates AI-powered summaries.
• A storage package supporting local and GCS backends ⇒ pluggable persistence layer.
• Docker, Docker Compose, CI/CD workflows ⇒ containerized full-stack application with automated deployment.
• Purpose: Expose an HTTP interface to ingest or query Git repositories, chunk their contents, send to an AI summarizer, format results, and store/retrieve data.
File structure
• Top-level:
• Dockerfile, docker-compose.yml, .github/workflows ⇒ build, test, and deploy pipelines.
• requirements.txt, pyproject.toml ⇒ Python dependencies.
• src/api:
• main.py and __main__.py: app entrypoint.
• routers: endpoints for ingest, summary, health, index.
• middleware, config, models, shared ⇒ API-level components (request/response schemas, config injection, auth, error handling).
• src/core:
• entrypoint.py: CLI or worker entrypoint.
• clone, chunking, parser, ai_summary, formatter, ingestion modules ⇒ core domain services.
• schemas and utils ⇒ data definitions and helper functions.
• src/storage:
• factory.py, base.py, local.py, gcs.py ⇒ abstract factory for storage adapters.
• templates + static: server-rendered HTML pages and associated JavaScript.
• tests: unit tests for each core and storage component.
Elements
• Labels: “Browser”, “API Server”, “Core Logic”, “Storage Adapter”, “GCS / Local FS”, “AI Service”, “Git Remote”
– Green for core domain services
– Yellow for storage/persistence