use model=TRUE in cox fit

NEWS.md

@@ -1,3 +1,7 @@
+# mlr3proba dev
+
+* fix: `lrn("surv.coxph")` is now trained with `model=TRUE` which fixes an issue with using observation weights [stackoverflow link](https://stackoverflow.com/questions/79297386/mlr3-predicted-values-for-surv-coxph-learner-with-case-weights).
+
 # mlr3proba 0.7.1
 
 * cleanup: removed all `PipeOp`s and pipelines related to survival => regression reduction techniques (see #414)

R/LearnerSurvCoxPH.R

@@ -43,7 +43,7 @@ LearnerSurvCoxPH = R6Class("LearnerSurvCoxPH",
       }
 
       invoke(survival::coxph, formula = task$formula(), data = task$data(),
-             .args = pv, x = TRUE)
+             .args = pv, model = TRUE)
     },
 
     .predict = function(task) {

tests/testthat/test_mlr_learners_surv_coxph.R

@@ -2,22 +2,11 @@ test_that("autotest", {
   with_seed(42L, {
     learner = lrn("surv.coxph")
     expect_learner(learner)
-    ## no idea why weights check here fails, we test the same task
-    ## in the below test and it works!
-    result = suppressWarnings(
-      run_autotest(learner, exclude = "weights", check_replicable = FALSE, N = 10L)
-    )
+    result = run_autotest(learner, check_replicable = FALSE)
     expect_true(result, info = result$error)
   })
 })
 
-test_that("weights", {
-  learner = lrn("surv.coxph")
-  task = generate_tasks.LearnerSurv(learner)$weights
-  learner$train(task)
-  expect_equal(learner$model$weights, task$weights$weight)
-})
-
 test_that("missing", {
   learner = lrn("surv.coxph")
   lung = survival::lung