SignalDeck

Operations console for crypto strategy research, backtests, live process state, risk controls, and audit logs.

SignalDeck pairs a React frontend with a FastAPI backend. The backend stores operational data in PostgreSQL in deployment, can use SQLite for local smoke tests, and exposes REST and WebSocket endpoints for the dashboard.

This is engineering tooling, not a trading signal service or financial advice.

What It Includes

• Strategy registry, configs, compile jobs, and run state
• Start/stop controls for strategy processes
• Backtest runs with logs, trades, equity curves, and metrics
• Portfolio equity, positions, orders, fills, and market data views
• Spot/futures market structure view for order book skew, live OFI, taker flow, funding, open interest, liquidation stream status, and cross-asset correlation
• Risk limits, risk rehearsal, risk history, and audit events
• Health, metrics, REST APIs, and WebSocket live updates

Technology

• Frontend: React 18, TypeScript, Vite, Ant Design, TanStack Query, Zustand, ECharts, Lightweight Charts
• Backend: FastAPI, Uvicorn, pandas, NumPy, ccxt, psycopg, PostgreSQL, SQLite for local smoke tests
• Testing: Vitest, Testing Library, pytest
• Ops: Nginx-friendly static build, supervisor-compatible backend process, health and metrics endpoints

Repo Layout

.
├── backend/quant/                 # FastAPI backend and strategy runtime
│   ├── api_server.py              # REST, WebSocket, auth, persistence endpoints
│   ├── main.py                    # Strategy runner entrypoint
│   ├── statarb/                   # Backtest and strategy modules
│   ├── db_store.py                # SQLite repository implementation
│   ├── postgres_store.py          # PostgreSQL repository implementation
│   ├── tests/                     # Backend tests
│   ├── tools/                     # E2E, backup, and maintenance tools
│   └── README.md                  # Backend-specific notes
├── frontweb/www.zlsjtj.tech/      # React frontend
│   ├── src/
│   ├── package.json
│   ├── vite.config.ts
│   └── README.md                  # Frontend-specific notes
└── README.md

Local Setup

Backend

The deployed setup uses PostgreSQL. For local development, SQLite is the fastest way to start the API without creating a database first.

cd backend/quant
python3 -m venv .venv
source .venv/bin/activate
pip install -r requirements-dev.txt

export API_DB_BACKEND=sqlite
export API_DB_PATH=./logs/dev_quant_api.db
export API_AUTH_REQUIRED=false

uvicorn api_server:app --host 127.0.0.1 --port 8000 --reload

Health check:

curl http://127.0.0.1:8000/api/health

For a PostgreSQL-backed run:

export API_DB_BACKEND=postgres
export API_DB_PATH=/dev/null
export API_DB_POSTGRES_DSN='postgresql://quant_user:password@127.0.0.1:5432/quant_db'
uvicorn api_server:app --host 127.0.0.1 --port 8000

Frontend

cd frontweb/www.zlsjtj.tech
npm install
cat > .env.local <<'EOF'
VITE_API_BASE_URL=http://127.0.0.1:8000/api
VITE_WS_URL=ws://127.0.0.1:8000/ws
VITE_USE_MOCK=false
VITE_API_PROFILE=quant-api-server
VITE_MARKET_CONFIG_PATH=config_market.yaml
VITE_MARKET_POLL_MS=1000
EOF
npm run dev

Open http://127.0.0.1:5173.

Useful Commands

Backend:

cd backend/quant
source .venv/bin/activate
python -m pytest -q
python -m py_compile api_server.py

Frontend:

cd frontweb/www.zlsjtj.tech
npm run lint
npm run typecheck
npm run test:run
npm run build

API Surface

The frontend talks to the backend through /api:

• Auth: POST /auth/login, GET /auth/status, POST /auth/logout
• Health and metrics: GET /health, GET /metrics
• Strategies: GET/POST /strategies, GET/PUT /strategies/{id}, POST /strategies/{id}/start, POST /strategies/{id}/stop
• Backtests: GET/POST /backtests, GET /backtests/{id}, GET /backtests/{id}/logs
• Live state: GET /portfolio, GET /positions, GET /orders, GET /fills
• Risk: GET/PUT /risk, GET /risk/history, GET /risk/events
• Audit and logs: GET /audit/logs, GET /logs
• Market data: GET /market/ticks, GET /market/klines
• Live stream: WS /ws

Data and Secrets

Runtime data and credentials are intentionally not part of the repository. Local .env files, .secrets/, databases, logs, generated backtest results, build output, and exchange keys should stay on the machine where the system runs.

Sample configuration files are for development and testing. Review the risk settings before connecting any strategy to a real exchange account.

Checks

The main local checks are:

• Backend: python -m pytest -q
• Frontend: npm run lint, npm run typecheck, npm run test:run, npm run build

License

No license has been selected yet.