Yuxuan Yan
07ed6ad917
reversal_rank weights the 5-day reversal signal by bounded cross-sectional rank instead of z-score, so a few extreme A-share pct_change outliers (newly listed / post-suspension / limit-up names) can no longer dominate the book. compute_alpha gains an optional per-date investable-universe mask (tradable, non-ST, seasoned, top-liquidity) applied to the signal before weighting, exposed via --liquid-universe/--universe-top-n. combo combine now accepts a single alpha as an identity passthrough so a one-alpha pipeline run needs no synthetic second input. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
13 lines
334 B
Python
13 lines
334 B
Python
"""Built-in alpha library.
|
|
|
|
Importing this package imports each alpha module, which registers the alpha via
|
|
the ``@register_alpha`` decorator. Add a new built-in by dropping a module here
|
|
and importing it below.
|
|
"""
|
|
from pipeline.alpha.library import ( # noqa: F401
|
|
momentum,
|
|
reversal,
|
|
reversal_rank,
|
|
reversal_vol,
|
|
)
|