[breaking] Add prediction fucntion for DMatrix and use inplace predict for dask. (#6668)

* Add a new API function for predicting on `DMatrix`.  This function aligns
with rest of the `XGBoosterPredictFrom*` functions on semantic of function
arguments.
* Purge `ntree_limit` from libxgboost, use iteration instead.
* [dask] Use `inplace_predict` by default for dask sklearn models.
* [dask] Run prediction shape inference on worker instead of client.

The breaking change is in the Python sklearn `apply` function, I made it to be
consistent with other prediction functions where `best_iteration` is used by
default.

tests/ci_build/conda_env/cpu_test.yml

@@ -34,6 +34,7 @@ dependencies:
 - llvmlite
 - pip:
   - shap
+  - ipython                     # required by shap at import time.
   - guzzle_sphinx_theme
   - datatable
   - modin[all]