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
The argument for not using the filter at prediction time is that it is not clear what the row_filter is used for. It is a generic feature to cut data with pandas filters. Doing this during prediction time seems harsh. One could for example intentionally filter out data to build a certain model (low rpm model?), with the precise intention to see how it behaves under other circumstances.
The text was updated successfully, but these errors were encountered:
something should definitely change, just silently dropping the data is clearly bad. And my guess is that it probably screws up some date alignment as well since the server expects data without gaps
gordo/gordo/client/client.py
Line 543 in e04c446
Previously, the row_filter was not used when doing predictions. After 994a9f8#diff-e7f1b0bbb4d76747e62b1fb5d978a1feL505 I believe it is.
The argument for not using the filter at prediction time is that it is not clear what the row_filter is used for. It is a generic feature to cut data with pandas filters. Doing this during prediction time seems harsh. One could for example intentionally filter out data to build a certain model (low rpm model?), with the precise intention to see how it behaves under other circumstances.
The text was updated successfully, but these errors were encountered: