Allow `Y.hat` input for `causal_survival_forest()`?

[bcjaeger]

Hello,

Thank you for developing grf. It's great!

Would it be feasible to allow Y.hat to be an input for causal survival forests? I see from code in causal_survival_forest that a couple of intermediate values are taken from the output of survival_forest(), so it may not be straightforward to just plug in Y.hat from a separate routine. I am interested in using a forest object from aorsf::orsf()(https://github.com/ropensci/aorsf). If aorsf::orsf() could provide those intermediate outputs, would it be feasible to allow a forest object from aorsf to be used here?