Data Pipeline

Ingest, clean, and feature-engineer market data.

The strategy is data-agnostic. It accepts CSV files, yfinance downloads, or broker APIs via a single interface.

Supported sources

Source	Command	Status
Local CSV	python src/backtest.py --data csv --path data/sample.csv	Ready
yfinance	python src/backtest.py --data yfinance --ticker SPY	Ready
OANDA REST	data_feed.oanda_stub(...)	Stub; needs token

Required columns

• open, high, low, close, volume
• Optional: confirmation (0 to 1 external score)

Feature set

ATR, ADX, Hurst exponent, 20-day and 60-day returns, volume ratio, volatility percentile, and mean-reversion z-score. All features are computed inside features.py.