- Project structure created (src/clawrity layout)
- pyproject.toml with all dependencies
- .gitignore configured
- Pydantic v2 models (all phases)
- Custom exception hierarchy
- Settings (pydantic-settings)
- Client config loader (YAML)
- Mock data service
- Logging + formatters
- Docker + CI/CD
- Unit tests (54 passing)
- Data Service — Unified access (mock + BigQuery backends)
- NL-to-SQL Engine — Pattern-based query generation
- Gen Agent — Newsletter summaries via DeepSeek
- QA Agent — Hallucination scoring with JSON parsing
- Agent Orchestrator — Gen→QA pipeline with retry logic
- FastAPI App — Chat, digest, clients, health endpoints
- Slack Channel — Base adapter with Socket Mode
- Unit tests — 54 tests passing
- Integration tests (full pipeline test)
- End-to-end demo with real DeepSeek API
- Vector store service
- RAG agent
- Recommendation endpoint
- ML forecast service (Prophet)
- Prediction cache
- Scenario endpoint
| Commit | Description | Date |
|---|---|---|
| 85e5e51 | feat: Phase 0 — Project foundation | 2026-05-01 |
| 5d5ecb9 | chore: Remove leftover src/pulse | 2026-05-01 |
| c2cd2b4 | docs: Update progress.md | 2026-05-01 |
| 111510a | feat: Phase 1 — Core services, agents, API | 2026-05-01 |
| ced2871 | docs: Update progress.md — Phase 1 complete | 2026-05-01 |
| 50ddb9c | feat: Groq (free) for non-critical + Azure Terraform | 2026-05-01 |
- Coverage reporting shows 0% (tests import from installed package, not source)
- Read this file to understand current progress
- Read context.md for architectural decisions (especially LLM strategy)
- Check git log:
git log --oneline -10 - Run quality gates:
ruff check . && pytest tests/unit/ - Get Groq API key (free at https://console.groq.com)
- Test end-to-end with real API keys
- Deploy to Azure:
cd infra && terraform apply