Added a method to compute data drift on a ME

[fterrazzoni]

Relates to https://github.com/dataiku/dip/pull/13469

[shortcut-integration]

This pull request has been linked to Clubhouse Story #68855: Replace the "Compute data drift" scenario step by an API to compute the drift between two "model likes".

[lpenet]

Playing with it, I felt quite incomfortable with the API, despite the fact that it works.
The user can not easily get the referenceId of a ME or a SMV through the API.
We discussed this with @fterrazzoni and agreed that it would be nice to also allow to pass a ModelEvaluation or a DSSTrainedPredictionModelDetails as a parameter.
It would also be nice to allow the user to access the fullId easily. But as this is an API change, I guess your opinion is required, @instanceofme

[lpenet]

Almost good. One last remark on the doc. Seems important as a user could get perplex, else.

[lpenet]

LGTM+tested with Python 2.7 and 3.7. Would like to enrich API doc, but will turn that in an opportunity of learning for a newcomer. :-)