Add cap to seasonal period to include decomposition #4147
Conversation
eccabay
changed the title
Codecov Report
@@ Coverage Diff @@
## main #4147 +/- ##
=======================================
+ Coverage 99.7% 99.7% +0.1%
=======================================
Files 349 349
Lines 37780 37781 +1
=======================================
+ Hits 37663 37664 +1
Misses 117 117
Help us with your feedback. Take ten seconds to tell us how you rate us. Have a feature suggestion? Share it here. |
eccabay
requested review from
christopherbunn,
jeremyliweishih,
fjlanasa,
asniyaz and
chukarsten
evalml/pipelines/utils.py
Outdated
| @@ -241,7 +241,7 @@ def _get_decomposer(X, y, problem_type, estimator_class, sampler_name=None): | |||
| y, | |||
| rel_max_order=order, | |||
| ) | |||
| if seasonal_period is not None: | |||
| if seasonal_period is not None and seasonal_period <= 1000: | |||
There was a problem hiding this comment.
in the make_pipeline tests can you make sure that the decomposer is not added if the seasonal period is > 1000? I think we can just add a case here by mocking seasonal_period.
There was a problem hiding this comment.
Looking back, I realized I completely forgot about the test_make_pipeline_controls_decomposer_time_series - I added a case in there!
There was a problem hiding this comment.
Ugh, my only issue with this is the hardcoded 1000 period. Particularly the ramifications for if a user does show up with a super high frequency dataset like the one that triggered this.
There was a problem hiding this comment.
Very fair. I'll drop it over to be a constant.
Closes #4146
Justification for this change can be found in the issue directly. I ran performance tests, there were no changes as all our test datasets have a period of <1000.