Running Training

Training processes your pypscript strategy on historical data to create a deployable .pyp skill.

How Training Works

1. Open your project in the editor
2. Click "Train" button in the toolbar
3. Configure pairs, timeframes, date range
4. Wait for training to complete

Training API

POST /api/projects/{projectId}/training-runs

Request body:

{
  "pairs": ["BTCUSDT", "ETHUSDT"],
  "timeframes": ["1h", "4h"],
  "startDate": "2024-01-01",
  "endDate": "2024-12-31",
  "initialBalance": 10000
}

Response

{
  "id": "run_abc123",
  "status": "completed",
  "createdAt": "2024-01-15T10:30:00Z",
  "completedAt": "2024-01-15T10:35:00Z",
  "pypFiles": [
    "skills/user_123/strategy_1h.pyp",
    "skills/user_123/strategy_4h.pyp"
  ]
}

Training Statuses

WebSocket Updates

Training progress is streamed via WebSocket:

{
  "type": "progress",
  "progress": 75,
  "logs": ["Loading data...", "Computing indicators...", "Training model..."]
}

When complete:

{
  "type": "complete",
  "results": { "pypFiles": [...] }
}