Add frames to return hazard and/or survival functions to coxPH corresponding to input dataset

[wendycwong]

Current CoxPH predict function will only return exp(beta * covariate). However, it will be helpful to also return the hazard or survival functions as well if the user will choose to have it.

[wendycwong]

A good reference on what to include for our CoxPH model is:

They did mention one library based on XgBoost as a good example of full tool set for survival analysis: https://github.com/loft-br/xgboost-survival-embeddings

[wendycwong]

actually, coxPH already return baseline hazard frame and baseline survival frames. If your model is called model, you get them by calling:

model.baseline_hazard_frame and model.baseline_survival_frame.

However, these are calculated as a function of time unit t and not correspond to each row of the dataset.