Skip to content

Latest commit

 

History

History
37 lines (24 loc) · 1.12 KB

File metadata and controls

37 lines (24 loc) · 1.12 KB

CLAUDE.md

Guidance for Claude Code when working in this repo.

What this is

Backing project for a learnwithparam.com course. See README.md for the subject and walkthrough.

Quick start

uv sync              # Python projects (or: npm install / go mod tidy)
cp .env.example .env # populate OPENROUTER_API_KEY and any other keys
uv run uvicorn main:app --reload --port 8111

Smoke test

This repo is covered by the shared sweep in ../smoke_test_all.sh. To run just this one:

bash ../smoke_test_all.sh streaming-llm-fastapi

Expect: 200 OK from the main endpoint + a real LLM response (not a stub).

Push workflow

Before every push, always:

  1. Run the smoke test above. It must pass.
  2. Verify .gitignore covers .env, __pycache__/, .venv/, data dirs, and OS files.
  3. git status — confirm no tracked secrets, no data artefacts, no local scratch files.
  4. Commit with a descriptive message.
  5. git push origin main — remote is git@github.com:learnwithparam/streaming-llm-fastapi.git.

Never force-push. Never commit .env or populated data directories.