AgentSkillsCN

trading-data-pipeline

为本仓库运行并验证BTCUSDT OHLCV数据的摄入。当用户希望下载、刷新、验证或排查artifacts/raw与artifacts/reports中的原始市场数据时使用。

SKILL.md
--- frontmatter
name: trading-data-pipeline
description: Run and verify BTCUSDT OHLCV ingestion for this repository. Use when the user asks to download, refresh, validate, or troubleshoot raw market data in artifacts/raw and artifacts/reports.

Trading Data Pipeline

Run the repository's data pipeline in this order:

  1. •Activate virtual environment: source .venv/bin/activate.
  2. •Download data: make run-download.
  3. •Validate data: make run-validate.
  4. •Read report: artifacts/reports/data_quality.json.

If make is unavailable, run scripts directly:

  • •python src/data/download_data.py
  • •python src/data/validate_data.py

When reporting results, include:

  1. •Input symbol, timeframe, and date range from configs/config.yaml.
  2. •Row count and first/last timestamp from artifacts/raw/BTCUSDT_15m.csv.
  3. •Validation status and errors from artifacts/reports/data_quality.json.

If download fails:

  1. •Confirm internet access and exchange name in config.
  2. •Confirm symbol format required by CCXT (BTC/USDT).
  3. •Retry once, then report the exact exception and failing step.

For quick command references, read references/commands.md.