[FEATURE] Simplify some arguments of the evaluation

[Jannchie]

Description

I found that for metrics like map_at_k, ndcg_at_k, precision_at_k, recall_at_k, the col_rating argument seem unnecessary if the relevancy_method is "top_k". However, the current implementation will report an error if this argument is not passed in:

ValueError: Missing columns in true rating DataFrame

Expected behavior with the suggested feature

I think it's a bit confusing to pass in unnecessary arguments.

A simple but dirty fix would be to change the default arguments DEFAULT_RATING_COL to DEFAULT_USER_COL.

Other Comments

[miguelgfierro]

@ChuyangKe can you review this?

[yueguoguo]

Perhaps another way is to use *args and then check the existence of rating column in args when relevancy_method is not "top_k".

[AdityaSoni19031997]

i can work on this issue, Can someone point-me to the correct file? I think the changes has to be cascaded and checked as well as multiple modules might be using these metrics?
Thanks

[miguelgfierro]

@AdityaSoni190319 can you implement the solution proposed by Le (yueguoguo)?

metrics are here https://github.com/microsoft/recommenders/blob/main/recommenders/evaluation/python_evaluation.py

[AdityaSoni19031997]

@miguelgfierro I have tried to patch this fix and had run the tests for the given methods as well. LMK your feedback or any other changes needed.