skatedplrn issues

Results 18 issues of


                                            skatedplrn

Actuals should be sent to endpoint only if multipredict_enabled is TRUE

bug

Method to build/get metrics for the trial

Now it is only visible in UI

enhancement

Support API for R language

Customer request

enhancement

feature_divergence models should be recreated with new sklearn

and fix issue with feature_divergence

enhancement

Build docker with azure

1. create setup.azure.py and include only azure there 2. add step to build docker for azure

enhancement

Add baseline_scores to actual endpoint

Now we have baseline_target inside actuals data, it require to keep all predicted values on client side baseline_scores: {actual_at: {'r2': 0.2, 'mda': 0.54}}

enhancement

MLRAM: review command should return as much information as possible

now it return only accuracy, but should return all available metrics and ROI

enhancement

Tasks which handle datasets should be run in separate process

if task failed with OOM , hub does not get any result, since we do not use result backend see trial worker in auger-ml

enhancement

send providers to create project file

see https://github.com/deeplearninc/hub/issues/3476

enhancement

support postgres url as data source

enhancement