trading_test_sda

Trading strategy backtesting and analysis framework for PEA (French equity accounts).

Setup

python -m venv .venv
source .venv/bin/activate  # On Windows: .venv\Scripts\activate
pip install -e ".[dev]"

Quick Start

# Download market data (creates data/market_data.db - ~13MB, 87 tickers)
python src/data_collector.py

# Run backtest
python -c "from src.strategy import generate_portfolio_signals; r = generate_portfolio_signals(); print(r['backtest'])"

# Run tests
pytest tests/

Project Structure

• src/ - Source code
  • data_collector.py - OHLCV collection → DuckDB
  • indicators.py - SMA200, RSI14, volume, value scores
  • backtester.py - VectorBT backtesting with 0.5% PEA fees
  • portfolio_optimizer.py - PyPortfolioOpt with 5% max position
  • strategy.py - Orchestration
  • monitor.py - Telegram alerts
• config/ - PEA universe (85 stocks + 8 ETF)
• tests/ - Unit tests (5 tests)
• data/ - Market data (market_data.db - gitignored)

Features

• PEA-compliant universe: European stocks and ETFs only
• Smart Money Scanner: Trend + volume + value composite score
• Backtesting: 12 years history, 0.5% transaction costs
• Optimization: Max Sharpe with L2 regularization