Skip to content

CLAUDE.md

Latest commit

 

History

History
54 lines (41 loc) · 2.07 KB

CLAUDE.md

File metadata and controls

54 lines (41 loc) · 2.07 KB
Raw

Lyrebird

Distributed microservices for real-time Bluesky sentiment analysis.

Architecture

NestJS monorepo with RabbitMQ message passing:

  • apps/gateway - REST API (port 3000)
  • apps/ingestion - Bluesky scraper (port 3001)
  • apps/analysis - Sentiment engine (port 3002)
  • apps/dashboard - React frontend (Vite)
  • libs/* - Shared modules (database, rabbitmq, bluesky, logger, shared-types, testing)

Commands

pnpm start:all                             # Run all services with hot-reload
pnpm test                                  # All unit tests
pnpm test -- --testPathPattern=<pattern>   # Single backend test
pnpm validate                              # Full validation (lint + type-check + test + knip + build)
docker compose up -d                       # Start infrastructure (Postgres, RabbitMQ)
pnpm db:studio                             # Drizzle Studio for DB inspection

Conventions

  • Conventional commits required
  • Libs use @app/* path aliases (e.g., @app/database)
  • Tests colocated with source (.spec.ts)
  • Services communicate via RabbitMQ patterns defined in libs/shared-types

Dashboard (apps/dashboard)

  • Vite + React 19 + Tailwind + shadcn/ui (new-york style, zinc base)
  • TanStack Query for server state, Recharts + Lightweight Charts for visualization
  • Uses @/* path aliases (maps to ./src/*)
  • Vitest + Testing Library for unit tests, Playwright for e2e
  • MSW for API mocking in tests
  • API requests proxied to gateway via Vite dev server (/api -> localhost:3000)
pnpm --filter dashboard dev              # Dev server (port 5173)
pnpm --filter dashboard test:run         # All unit tests
pnpm --filter dashboard test:run <file>  # Single unit test
pnpm --filter dashboard test:e2e         # Playwright e2e tests

Gotchas

  • Backend uses Jest, dashboard uses Vitest - don't mix test APIs
  • Dashboard is a separate pnpm workspace - always use pnpm --filter dashboard <cmd>
  • Run docker compose up -d before starting services (needs Postgres + RabbitMQ)
  • ML models cached in models-cache/ - first run downloads ~500MB