dr_score and plug_in_t_score document learner as optional but raise without it

[jeongyoonlee]

The docstrings of dr_score and plug_in_t_score (causalml/metrics/cate_scoring.py) describe learner as optional, but _resolve_outcome_learners (cate_scoring.py:37) raises unless learner or both of control_outcome_learner/treatment_outcome_learner are passed:

from causalml.metrics import dr_score
dr_score(df, X=X, outcome_col="y", treatment_col="w")
# ValueError: Either `learner` or both `control_outcome_learner` and
# `treatment_outcome_learner` must be specified.

Observed on master (18ebfb0).

Suggestion: either supply a default learner (the sibling metrics in causalml.metrics default to LightGBM elsewhere) or change the docstrings to mark the argument required. Independent of #1026/#1027; touches only the learner-resolution path of cate_scoring.py.

🤖 Generated with Claude Code

[su-jin1425]

Hi! I’d like to work on this. I’ll first verify the intended learner API against the other CATE scoring functions, then either make the learner optional with an appropriate default or correct the documentation if it is intended to be required. I’ll add a regression test covering the current failure and update the relevant documentation.

mark it as done on or before 19/Aug