documentation for dynamic survival probabilities #979
Description
predict.model_fit()will generate a data frame with a list column with predictions at each time point.censoring_weights_graf.model_fit()can take the prediction results and add columns related to IPCWaugment.model_fit()will do both (if there is an outcome column).
We need to document this in the functions to reduce confusion. The main point to make related to predict() is that this S3 method does not ever involve the outcome (which is needed to compute censoring weights).
If there is an outcome, we should document that the required columns for yardstick can be made using two steps (predict() then censoring_weights_graf()) or with one step (augment()).