auto_trader

Intraday trading app: backtesting, paper trading, and (later) live trading via Capital.com, with a TradingView-style chart.

• Backend — Python (FastAPI + asyncio). Broker adapters, a backtest engine, and strategies. Designed so one strategy runs unchanged across backtest → paper → live; only the data source and order executor get swapped.
• Frontend — React + TypeScript + Vite. Charting via @klinecharts/pro — a TradingView-style shell (symbol search, timeframe bar, indicator menu, drawing toolbar) wired to our backend through a Datafeed. (Migrated from TradingView lightweight-charts, which lacks built-in drawing tools / indicators.)

Architecture

backend/auto_trader/
  brokers/    MarketDataBroker interface + Capital.com adapter (capital.com first)
  core/       domain models (Candle, Signal, Trade, Fill) — UTC everywhere
  strategy/   Strategy interface + SmaCross reference strategy
  engine/     event-driven backtest engine (no lookahead, next-open fills)
  api/        FastAPI REST: /api/candles, /api/backtest
frontend/src/ React UI: toolbar + lightweight-charts candlestick chart w/ markers

Milestone 1 (done): backtesting + chart. Historical candles from the Capital.com demo API → engine replays them point-in-time → SMA-cross strategy emits trades → REST serves candles + markers → chart renders with SMA overlays and an equity-curve pane. Plus instrument search and a live WebSocket stream of real-time candles onto the chart. No live orders yet.

API surface

Method	Path	Purpose
GET	/api/markets?q=	search instruments (epic, name, status)
GET	/api/candles?epic=&resolution=&bars=	recent OHLCV candles
GET	/api/backtest?epic=&resolution=&bars=&fast=&slow=	candles + trade markers + equity + summary
WS	/ws/candles?epic=&resolution=	live mid-price candles as they form

Candles use recent-bars mode (max without a date window), which is weekend-proof — a fixed date range 404s when the market is closed. Most epics are CLOSED on weekends; the _W epics (e.g. EURUSD_W) trade then and are good for testing the live stream.

Setup

Backend

cd backend
python3 -m venv .venv && source .venv/bin/activate
pip install -e ".[dev]"
cp .env.example .env        # then fill in your Capital.com DEMO credentials

API keys are environment-specific. Switch your Capital.com app to the Demo account before generating the key (Settings → API integrations), or it won't authenticate against the demo host.

Smoke-test the data path, then run the server:

python scripts/check_capital.py     # smoke test: auth + pull 5m candles
uvicorn auto_trader.api.app:app --reload --port 8000
pytest                              # engine correctness tests

Frontend

cd frontend
npm install
npm run dev

Then open the URL Vite prints (http://localhost:5173). Set the symbol/resolution/SMA periods in the toolbar and hit Run backtest. Override the API base with VITE_API_BASE if the backend isn't on :8000.

Paste commands without the trailing # ... comments — interactive zsh doesn't treat # as a comment and will pass it as an argument.

Roadmap

• Live price streaming over WebSocket (Capital.com OHLC subscription)
• TradingView-style chart shell (klinecharts/pro): drawing tools, indicator library, multi-timeframe, instrument search, live data
• Re-add backtest trade markers + equity curve as chart overlays
• Real alerts (fire via live stream + draggable) and on-chart order ticket
• Paper trading executor sharing the engine's strategy interface
• More indicators / overlays, additional strategies
• Additional broker adapters behind MarketDataBroker / execution seam
• Live order execution (gated, demo-first)