How stormbreaker06/reviewing- works
This is a FastAPI service for inspecting GitHub pull requests and producing AI code reviews. A caller can fetch pull-request metadata or a parsed diff; the review workflow fetches the PR head, retrieves its diff, parses changed files, loads each file’s content, and reviews files concurrently through an LLM. Review text is parsed into issue results. The graph reflects the sampled implementation; no README was provided, and the Pydantic result models are defined but their use in runtime responses is not shown.
Main entry points: main.py FastAPI routes · client.py GitHub client · llm.py LLM reviewer · diff_parser.py Diff parser
Components 10
HTTP API
- FastAPI routes
app/main.py - Review orchestration
app/main.py
GitHub access
- GitHub REST API
- GitHub client
app/github/client.py
Review workflow
- Diff parser
app/github/diff_parser.py - Review parser
app/github/diff_parser.py - Review models
app/model.py
AI review
- LLM reviewer
app/llm.py - LLM service
Other components
- API caller
Connections 14
- API caller → FastAPI routes: requests
app/main.py - FastAPI routes → GitHub client: fetches PR
app/main.py - FastAPI routes → GitHub client: fetches diff
app/main.py - GitHub client → GitHub REST API: calls
app/github/client.py - FastAPI routes → Diff parser: parses diff
app/main.py - FastAPI routes → Review orchestration: dispatches review
app/main.py - Review orchestration → GitHub client: fetches PR
app/main.py - Review orchestration → GitHub client: fetches diff
app/main.py - Review orchestration → Diff parser: parses diff
app/main.py - Review orchestration → GitHub client: loads file
app/main.py - Review orchestration → LLM reviewer: reviews files
app/main.py - LLM reviewer → LLM service: requests review
app/llm.py - LLM reviewer → Review parser: parses response
app/llm.py - FastAPI routes → API caller: returns response
app/main.py