Zipline 日频回测

Zipline 日频回测:使用 Zipline 框架执行日频股票策略回测,支持多市场数据接入、因子研究、可视化绩效分析,默认本金千万级。 含 25 条反模式约束。

✓ 0 人报告成功·v0.1.0·

晶体简介

使用 Zipline 框架执行日频股票策略回测,支持多市场数据接入、因子研究、可视化绩效分析,默认本金千万级。。 本 skill 基于开源项目构建并集成 25 条 anti-pattern 约束。访问 doramagic.ai/r/zipline-daily-backtest 查看中英双语完整文档和触发场景。适用于 Doramagic 生态(Claude Code / Cursor / openclaw / ChatGPT / Gemini 等)。

Blueprint Source

finance-bp-088

zvtvz/zvt6360a632 source files

Constraints

97total
72fatal
72 must-not-violate

Evidence Quality

Confidence89%

Medium confidence — review before critical use

72 条不可违反的约束

FATALdomain_rulefinance-C-001

WHENWhen writing daily bar data to bcolz format

ACTIONverify each asset has exactly one row per trading session with no gaps or duplicates

CONSEQUENCEThe BcolzDailyBarWriter raises an AssertionError if the number of rows per asset does not match the expected session count. Missing or extra sessions cause incorrect data alignment during backtesting, leading to wrong price lookups.

FATALdomain_rulefinance-C-002

WHENWhen validating OHLCV price values during ingestion

ACTIONverify each price and volume values fit within uint32 bounds after scaling

CONSEQUENCEPrice columns are scaled by 1000 and stored as uint32. Values exceeding UINT32_MAX (~4.29 billion) cause data corruption, zeroing out prices or causing silent data loss depending on invalid_data_behavior setting.

FATALdomain_rulefinance-C-003

WHENWhen providing start_session or end_session to BcolzDailyBarWriter

ACTIONverify provided session dates are valid trading days on the specified calendar

CONSEQUENCEInvalid session dates that are not trading days cause ValueError to be raised, preventing bundle ingestion from completing.

常见问题

讨论 (0)

暂无讨论,成为第一个发言的人吧!

更新历史

v0.1.02026-04-23·贡献者: tangweigang-jpg

v0.1.0: 首次发布到 Doramagic.ai。基于 zvtvz/zvt 的自动化 batch-v1 元数据 + 自动生成 FAQ。