Survival learner doc updates + miscellaneous stuff

[bblodfon]

Overview of changes

• Add prediction types @section for all survival learners, making sure it is well documented and consistent Refine prediction type doc for survival learners #347
• All survival learners have crank as main prediction (and it is always returned) Set default predict_type = crank for all survival learners #331
• Other stuff:
  • Set n_thread = 1 for surv.aorsf and use unique event time points for predicted S(t) @bcjaeger
  • Add selected_features() for surv.penalized
  • Fix surv.prioritylasso learner + add distr predictions via Breslow Survival Priority Lasso Update #344
  • SSVM gamma.mu was splitted to gamma and mu
  • Harmonized the use of times points for prediction as much as possible across survival learners Time points used in survival learners predicted matrix (distr) #387
    • parametric and akritas get a proper ntime argument
  • parametric is now used by default with discrete = TRUE

List of survival learners (WIP)

• aorsf, BART, 2 x coxboost, survflex, gbm, glmnet x 2, mboost x 4 {gamboost, glmboost, mboost, blackboost}, cforest, ctree, penalized, prioritylasso, survival SVM, rfsrc, ranger, akritas, parametric, nelson, 6 x DL survival models (coxtime, deepsurv, loghaz, deephit, dnnsurv, pchazard), 2 x xgboost (cox, aft)

[sebffischer]

looking great, just some minor comments / questions