[python] refine categorical feature support

[wxchan]

to do list:

• remove categorical features = 'auto' -> use None instead
• let user confirm categorical features are set successfully: maybe add some debug level log? categorical_feature seems does not work? #893 (but do no use too much logging: UserWarning when using lightgbm.Dataset from pandas.DataFrame if both train and valid_sets are specified #2157)
• check train, valid, predict consistency
• pandas category: [python] refine pandas support #960
• auto convert string type or object type: Categorical features not accepted #1020
• test_category.py

some other issues:

• sklearn categorical handling of early stopping #921
• Categorical features use causes crash returning from train function #994

[StrikerRUS]

Closed in favor of being in #2302. We decided to keep all feature requests in one place.

Welcome to contribute this feature! Please re-open this issue (or post a comment if you are not a topic starter) if you are actively working on implementing this feature.