tabular_classification/PassiveAClf.py no predict_proba function

[delphieritas]

    # probs = self._clf.predict_proba(data) 
    # PassiveAggressiveClassifier does not have predict_proba function, we can use _predict_proba_lr() function,
    # pls see https://github.com/scikit-learn/scikit-learn/blob/0fb307bf3/sklearn/linear_model/_base.py#L261

pls merge dev...delphieritas:patch-3

[delphieritas]

To justify the input format changes:

When do prepare_X(queries), it was supposed to be pandas Dataframe. But the original input passes a list to that function, then it won't work

[pinpom]

@delphieritas: as discussed, please run those models on singa-auto system to see if there is any errors and start from there. I did run those models successfully on singa-auto before, but there have been quite some changes on the system, i.e. data format,... so need to check and update.