feat(ci+tests): deploy safety audit — linting, rollback, smoke tests, 50+ new tests (#120 )

Comprehensive deploy safety audit implementing 19 improvements across CI/CD pipeline, test coverage, and source code.

### CI/CD Pipeline
- ruff + mypy added to both release.yml and keboola-deploy.yml (continue-on-error)
- Smoke test added to keboola-deploy.yml (was missing)
- Automatic rollback on smoke test failure in release.yml
- Expanded smoke-test.sh with catalog, admin/tables, marketplace.zip, metrics
- Required status checks via .github/settings.yml
- Dependabot + CODEOWNERS + pre-commit hooks + ruff config

### Source Code
- DB schema version check in /api/health (db_schema: ok/mismatch/unhealthy)
- Config versioning (config_version: 1 in instance.yaml, non-blocking validation)
- BigQuery extractor ATTACH error handling (try/except around INSTALL+ATTACH)
- Post-deploy smoke test script for prod VM validation

### Test Coverage (~50 new tests)
- v13->v14 migration, Email magic link TTL, PAT, Marketplace ZIP/Git,
  Jira webhooks, Hybrid Query BQ, Keboola/BQ extractor failure modes,
  Orchestrator failure modes

Co-Authored-By: Devin <158243242+devin-ai-integration[bot]@users.noreply.github.com>

2026-04-29 09:18:55 +02:00

2.4 KiB

Raw Blame History

Quick Start Guide

Prerequisites

Python 3.10+
Docker + Docker Compose (for production deployment)
Data source credentials (Keboola token, BigQuery project, etc.)

Local Development Setup

Clone the repository:

git clone <repo-url>
cd ai-data-analyst

Create virtual environment and install dependencies:

python3 -m venv .venv && source .venv/bin/activate
uv pip install ".[dev]"

Configure your instance:

cp config/instance.yaml.example config/instance.yaml
# Edit config/instance.yaml with your settings

Set up environment variables:

cp config/.env.template .env
# Edit .env with your data source credentials

# Via CLI
da admin register-table --source-type keboola --bucket "in.c-crm" --table "company" --query-mode local

# Or start the server and use the web UI at /admin/tables

Start the FastAPI server:
```
uvicorn app.main:app --reload
```

Trigger a data sync:

curl -X POST http://localhost:8000/api/sync/trigger
# Or: da sync

Docker Deployment

# Start app + scheduler
docker compose up

# Include telegram bot
docker compose --profile full up

# HTTPS mode — Caddy + corporate-CA certs
docker compose -f docker-compose.yml -f docker-compose.prod.yml -f docker-compose.tls.yml \
    --profile tls up -d

See DEPLOYMENT.md for full server setup instructions.

Using with Claude Code

Open the project in Claude Code. The CLAUDE.md file will guide the AI assistant through setup and analysis workflows.

Analyst Setup

Visit your instance URL (e.g., https://data.yourcompany.com)
Sign in with your company email
Access data through the API or download parquets for local analysis

Analysis Workflow

Sync latest data: curl -X POST https://data.yourcompany.com/api/sync/trigger
Open Claude Code in your project directory
Ask Claude to analyze your data using DuckDB

Hackathon

See HACKATHON.md for the deploy-and-develop playbook. Per-developer dev VMs are the supported pattern — point your VM at your branch image with gcloud compute ssh <vm> --command "sudo sed -i 's/^AGNES_TAG=.*/AGNES_TAG=dev-<slug>/' /opt/agnes/.env && sudo /usr/local/bin/agnes-auto-upgrade.sh".

2.4 KiB Raw Blame History