-
Notifications
You must be signed in to change notification settings - Fork 94
Description
Blocked by #1885, #1886.
Partially blocked by #1944 : the Project.summarize().reports()
can't work.
Is your feature request related to a problem? Please describe.
Today, we have project.summarize(), which outputs an extended pandas.DataFrame, containing only EstimatorReports.
Tomorrow, we would like to store and retrieve from the project both EstimatorReports and CrossValidationReports. Therefore, what should project.summarize() outputs?
-->
it's consistant with hub, where we have different tabs for Estimator and CrossValidation.
what is displayed makes sense.
in the parallel coordinates plot, the line for a given CrossValidationReport is the average of the metrics. For now, we do not want to display the std.
A crossvalidation report is made of several Estimator Reports. Those sub-Estimator Reports shouldn't be displayed not fetchable by project.summarize(report_type="Estimator").
Describe the solution you'd like
As presented in #1892, I would like to have:
project.summarize(report_type="Estimator")
. (or "CrossValidation")
report_type cannot be "all", because we cannot display both Estimator and CrossValidation reports in the same parallel plot, it wouldn't make sense.
Describe alternatives you've considered, if relevant
To give access to all reports in project.summarize()
, but it wouldn't make sense because these two objects are different by nature.
Additional context
No response