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
When users query public extractions parquet and provide their own processing period, we need to align existing extractions with the user-proposed season and make sure that the label (represented by the valid_date) is still preserved in the selected period.
While the current version of code is already doing that (refer to this part), it would be nice to add more context for the user and include the actual shift that has been made with the respect to the actual label position.
To do:
add an attribute to the resulting dataframe of the query_public_extraction function that preserves information about the shift
add a logger message/warning to the user that notifies the user about the number of samples where the shift was larger than a threshold
define the acceptable threshold of the shift
The text was updated successfully, but these errors were encountered:
When users query public extractions parquet and provide their own processing period, we need to align existing extractions with the user-proposed season and make sure that the label (represented by the valid_date) is still preserved in the selected period.
While the current version of code is already doing that (refer to this part), it would be nice to add more context for the user and include the actual shift that has been made with the respect to the actual label position.
To do:
query_public_extraction
function that preserves information about the shiftThe text was updated successfully, but these errors were encountered: