Skip to content

IMPLEMENTATION_PLAN.md

Latest commit

 

History

History
61 lines (47 loc) · 3.59 KB

IMPLEMENTATION_PLAN.md

File metadata and controls

61 lines (47 loc) · 3.59 KB

Advanced Backtesting & Strategy Studio Upgrade

Goal Description

Enhance the existing "Strategy Studio" to rival frameworks like "Jesse" or "VectorBT" in terms of visualization and capability. This involves integrating vectorbt for high-performance backtesting and parameter optimization, and creating a rich frontend dashboard to visualize equity curves, trade lists, and optimization heatmaps.

User Review Required

Important

This plan introduces vectorbt as a dependency. It is a powerful library but can be heavy. We will aim for the standard (non-PRO) version which is open source.

Note

We will use recharts for the frontend visualization as it is already in the project, creating a consistent look and feel with the existing dashboard.

Proposed Changes

Backend

[MODIFY] requirements.txt

  • Add vectorbt and schedule (if needed for advanced timing, though vectorbt usually suffices for backtest).

[MODIFY] backend/backtesting_agent.py

  • Fully implement run_backtest using vectorbt.
  • Add support for Parameter Optimization: Detect if parameters are lists/ranges and run vbt.Portfolio.from_signals in batch mode.
  • Return comprehensive JSON:
    • metrics: Existing scalar metrics.
    • equity: Time series of equity.
    • trades: List of individual trades (entry/exit time, price, PnL).
    • optimization_results: (Optional) Table of Params vs. Return/Sharpe for heatmaps.

[NEW] backend/optimizer_utils.py

  • Helper functions to parse natural language ranges (e.g., "fast 10 to 50 step 10") into python lists for vectorbt.

Frontend

[NEW] frontend/src/components/BacktestDashboard.tsx

  • A detailed dashboard component replacing the simple table.
  • Charts: Equity Curve (Recharts AreaChart), Drawdown (Recharts AreaChart).
  • Tables: Trade List (Scrollable table with Win/Loss coloring).
  • Optimization View: If multiple results exist, show a heatmap or sorted table of best parameter sets.

[MODIFY] frontend/src/pages/StrategyStudio/index.tsx

  • Integrate BacktestDashboard.
  • Add UI inputs for parameter ranges (or rely on the Chatbot to parse "optimize ...").
  • Update runBacktest to handle the richer response.

Verification Plan

Automated Tests

  • Backend Unit Test: Create backend/tests/test_vectorbt.py to ensure vectorbt runs a simple crossover and returns the expected JSON structure.
  • API Check: Use curl to POST to /api/agent/execute_task with a backtest task and verify the JSON response contains equity and trades.

Manual Verification

  1. Optimization Flow:
    • Go to Strategy Studio.
    • Type: "Optimize SMA strategy on XAUUSD H1. Fast from 10 to 50 step 10, Slow from 60 to 100 step 10."
    • Verify the backend runs multiple combinations.
    • Verify the Frontend displays a heatmap/table of results.
    • Click a result to see its specific Equity Curve.
  2. Visual Check:
    • Run a single backtest.
    • Check that the Equity Curve looks correct (starts at 0% or 100%, fluctuates).
    • Check the Trade List matches the chart logic.