[tune][train] Improve ray.train method deprecations called from Tune functions

[justinvyu]

Description

Ray Tune users need to stop using ray.train.get_context(), ray.train.get_checkpoint() and ray.train.report. This PR improves the error messages raise if they try to call these instead of the ray.tune counterparts. Note that we've already soft-deprecated this usage with a warning message for 6+ months.

Related issues

#49454

Additional information

Before:

ray.exceptions.RayTaskError(RuntimeError): ray::ImplicitFunc.train() (pid=3034, ip=127.0.0.1, actor_id=b88a8a6f297b83cd04bc6a5001000000, repr=tune_fn)
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/Users/justin/Developer/ray/python/ray/tune/trainable/trainable.py", line 331, in train
    raise skipped from exception_cause(skipped)
  File "/Users/justin/Developer/ray/python/ray/air/_internal/util.py", line 107, in run
    self._ret = self._target(*self._args, **self._kwargs)
                ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/Users/justin/Developer/ray/python/ray/tune/trainable/function_trainable.py", line 44, in <lambda>
    training_func=lambda: self._trainable_func(self.config),
                          ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/Users/justin/Developer/ray/python/ray/tune/trainable/function_trainable.py", line 249, in _trainable_func
    output = fn()
             ^^^^
  File "/Users/justin/Developer/ray/python/ray/tune/tests/test_api_migrations.py", line 25, in tune_fn
    ray.train.get_checkpoint()
  File "/Users/justin/Developer/ray/python/ray/train/v2/api/train_fn_utils.py", line 182, in get_checkpoint
    return get_train_fn_utils().get_checkpoint()
           ^^^^^^^^^^^^^^^^^^^^
  File "/Users/justin/Developer/ray/python/ray/train/v2/_internal/execution/train_fn_utils.py", line 264, in get_train_fn_utils
    raise RuntimeError("TrainFnUtils has not been initialized.")
RuntimeError: TrainFnUtils has not been initialized.

After:

ray.exceptions.RayTaskError(DeprecationWarning): ray::ImplicitFunc.train() (pid=97276, ip=127.0.0.1, actor_id=c9cd63f30250e52577c0b4c601000000, repr=tune_fn)
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/Users/justin/Developer/ray/python/ray/tune/trainable/trainable.py", line 331, in train
    raise skipped from exception_cause(skipped)
  File "/Users/justin/Developer/ray/python/ray/air/_internal/util.py", line 107, in run
    self._ret = self._target(*self._args, **self._kwargs)
                ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/Users/justin/Developer/ray/python/ray/tune/trainable/function_trainable.py", line 44, in <lambda>
    training_func=lambda: self._trainable_func(self.config),
                          ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/Users/justin/Developer/ray/python/ray/tune/trainable/function_trainable.py", line 249, in _trainable_func
    output = fn()
             ^^^^
  File "/Users/justin/Developer/ray/python/ray/tune/tests/test_api_migrations.py", line 44, in tune_fn
    ray.train.get_checkpoint()
  File "/Users/justin/Developer/ray/python/ray/train/v2/api/train_fn_utils.py", line 203, in get_checkpoint
    raise DeprecationWarning(
DeprecationWarning: `ray.train.get_checkpoint` is deprecated when running in a function passed to Ray Tune. Please use `ray.tune.get_checkpoint` instead. See this issue for more context: https://github.com/ray-project/ray/issues/49454

[gemini-code-assist]

Code Review

This pull request improves the deprecation warnings for ray.train APIs when used within Ray Tune, which is a great enhancement for user experience. The implementation correctly raises DeprecationWarning and the accompanying tests cover both the new and old behavior. My review includes a couple of suggestions to improve code maintainability by reducing duplication in both the implementation and the tests. Overall, the changes are solid.