You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
check_non_state_changing_method fail normally means that some estimator attribute is being changed in predict/transform. These are probably all related, and require either a fit method if fit is empty is true or some reworking. If we want to keep it like this, we should create an issue and comment it in the file #testing_config
"FLUSSSegmenter"
"InformationGainSegmenter"
"GreedyGaussianSegmenter"
"ClaSPSegmenter"
"HMMSegmenter"
"RSTSF"
"QuerySearch"
"SeriesSearch"
Last two have a comment
# Keeps length during predict to avoid recomputing means and std of data in fit
# if the next predict calls uses the same query length parameter.
which is fine, but maybe an issue expanding on it?
Describe your proposed solution
fix or create issue
Describe alternatives you've considered, if relevant
No response
Additional context
No response
The text was updated successfully, but these errors were encountered:
You can remove QuerySearch and SeriesSearch from the list, as they will be replaced in an upcoming PR with simpler and more generic objects.
For example, in QuerySearch, it's used to keep track of the size of the query given during predict. This is to avoid recomputing means and stds of the collection given in fit everytime we call predict. To do so we modify a _previous_query_length parameter during predict, causing the state change.
Describe the feature or idea you want to propose
check_non_state_changing_method fail normally means that some estimator attribute is being changed in predict/transform. These are probably all related, and require either a fit method if fit is empty is true or some reworking. If we want to keep it like this, we should create an issue and comment it in the file #testing_config
Last two have a comment
# Keeps length during predict to avoid recomputing means and std of data in fit
# if the next predict calls uses the same query length parameter.
which is fine, but maybe an issue expanding on it?
Describe your proposed solution
fix or create issue
Describe alternatives you've considered, if relevant
No response
Additional context
No response
The text was updated successfully, but these errors were encountered: