 claude_dev
|
db2cc8c531
|
fix(factor): compute_factors 时区对齐（边界 Asia/Shanghai aware，真数据跑通因子管线）
- Root cause: vnpy.alpha's to_datetime() creates naive datetimes from strings,
causing SchemaError when comparing with timezone-aware DataFrame columns
- Fix: Convert period boundaries to Asia/Shanghai-aware datetimes + localize
DataFrame datetime column before passing to AlphaDataset
- Restore data_adapter.py to fa7237b (removed ineffective tz stripping)
- Add test_compute_factors_passes_aware_periods_to_alpha_dataset
- Real data verification: 600000.SSE ma5 factor analysis successful
- Container tests: 67 passed
Co-Authored-By: Claude <noreply@anthropic.com>
|
2026-07-06 22:18:51 +08:00 |
|
|
claude_dev
|
fa7237b996
|
feat(polish): 因子报告 IC 提取 + periods 提参 + 缓存上限 + 代码整洁
- analyzer.py: 提取 IC 值到 ic_summary (mean/std/icir/t_stat)，periods 提参 (默认 1,5,10)
- alpha_lab.py: _loaded_bars 缓存 LRU 上限 (_MAX_CACHED_SYMBOLS=50)
- runner.py: 统一阶段文案 (参数优化中/因子分析中)，worker 类型标注，_wait_future 文档
- pool.py: submit_work 添加 task_id debug 日志
Co-Authored-By: Claude <noreply@anthropic.com>
|
2026-07-06 21:32:27 +08:00 |
|
|
claude_dev
|
8972d1058f
|
feat(factor): alpha_lab compute_factors 完整化
添加 compute_factors 方法到 AlphaLabSession：
- 在 __init__ 添加 _loaded_symbols 和 _loaded_bars 缓存
- load_symbols 现在缓存 bar 数据供 compute_factors 使用
- compute_factors 使用缓存的 bars 调用 AlphaDataset
- 使用懒导入避免本地 Python 3.14 缺少 polars/vnpy.alpha 的 ImportError
测试 (container only):
- test_compute_factors_calls_prepare_and_fetch 验证 AlphaDataset 调用
Co-Authored-By: Claude <noreply@anthropic.com>
|
2026-07-06 19:14:12 +08:00 |
|
|
claude_dev
|
2379159818
|
feat(factor): AlphaLab 封装 + Alphalens 分析器
- 新增 alpha_lab.py: AlphaLabSession 类（lazy import vnpy.alpha）
- 新增 analyzer.py: run_factor_analysis + FactorReport（lazy import alphalens）
- 测试: test_alpha_lab.py (2 passed) + test_analyzer.py (3 passed)
- 策略: 本地无 alphalens/polars，函数内 lazy import 避免 ImportError
- 骨架: run_factor_analysis 返回 FactorReport，alphalens 集成待完整实现
|
2026-07-06 10:36:46 +08:00 |
|