[AIR] switch to a common RunConfig. (#23076)

python/ray/ml/config.py

@@ -1,8 +1,49 @@
-from typing import Dict, Any
+from dataclasses import dataclass
+from typing import Any, Dict, List, Optional
 
-# Right now, RunConfig is just an arbitrary dict that specifies tune.run
-# kwargs.
-# TODO(xwjiang): After Tuner is implemented, define the schema
-RunConfig = Dict[str, Any]
+from ray.tune.callback import Callback
+from ray.util import PublicAPI
 
 ScalingConfig = Dict[str, Any]
+
+
+@dataclass
+@PublicAPI(stability="alpha")
+class FailureConfig:
+    """Configuration related to failure handling of each run/trial.
+    Args:
+        max_failures: Tries to recover a run at least this many times.
+            Will recover from the latest checkpoint if present.
+            Setting to -1 will lead to infinite recovery retries.
+            Setting to 0 will disable retries. Defaults to 0.
+    """
+
+    max_failures: int = 0
+
+
+@dataclass
+@PublicAPI(stability="alpha")
+class RunConfig:
+    """Runtime configuration for individual trials that are run.
+
+    This contains information that applies to individual runs of Trainable classes.
+    This includes both running a Trainable by itself or running a hyperparameter
+    tuning job on top of a Trainable (applies to each trial).
+
+    Args:
+        name: Name of the trial or experiment. If not provided, will be deduced
+            from the Trainable.
+        local_dir: Local dir to save training results to.
+            Defaults to ``~/ray_results``.
+        callbacks: Callbacks to invoke.
+            Refer to ray.tune.callback.Callback for more info.
+    """
+
+    # TODO(xwjiang): Clarify RunConfig behavior across resume. Is one supposed to
+    #  reapply some of the args here? For callbacks, how do we enforce only stateless
+    #  callbacks?
+    # TODO(xwjiang): Add more.
+    name: Optional[str] = None
+    local_dir: Optional[str] = None
+    callbacks: Optional[List[Callback]] = None
+    failure: Optional[FailureConfig] = None

python/ray/ml/run_config.py

@@ -1,47 +0,0 @@
-from dataclasses import dataclass
-from typing import List, Optional
-
-from ray.tune.callback import Callback
-from ray.util import PublicAPI
-
-
-@dataclass
-@PublicAPI(stability="alpha")
-class FailureConfig:
-    """Configuration related to failure handling of each run/trial.
-    Args:
-        max_failures: Tries to recover a run at least this many times.
-            Will recover from the latest checkpoint if present.
-            Setting to -1 will lead to infinite recovery retries.
-            Setting to 0 will disable retries. Defaults to 0.
-    """
-
-    max_failures: int = 0
-
-
-@dataclass
-@PublicAPI(stability="alpha")
-class RunConfig:
-    """Runtime configuration for individual trials that are run.
-
-    This contains information that applies to individual runs of Trainable classes.
-    This includes both running a Trainable by itself or running a hyperparameter
-    tuning job on top of a Trainable (applies to each trial).
-
-    Args:
-        name: Name of the trial or experiment. If not provided, will be deduced
-            from the Trainable.
-        local_dir: Local dir to save training results to.
-            Defaults to ``~/ray_results``.
-        callbacks: Callbacks to invoke.
-            Refer to ray.tune.callback.Callback for more info.
-    """
-
-    # TODO(xwjiang): Clarify RunConfig behavior across resume. Is one supposed to
-    #  reapply some of the args here? For callbacks, how do we enforce only stateless
-    #  callbacks?
-    # TODO(xwjiang): Add more.
-    name: Optional[str] = None
-    local_dir: Optional[str] = None
-    callbacks: Optional[List[Callback]] = None
-    failure: Optional[FailureConfig] = None

python/ray/ml/trainer.py

@@ -5,7 +5,7 @@ from typing import Dict, Union, Callable, Optional, TYPE_CHECKING, Type
 from ray.ml.preprocessor import Preprocessor
 from ray.ml.checkpoint import Checkpoint
 from ray.ml.result import Result
-from ray.ml.config import ScalingConfig, RunConfig
+from ray.ml.config import RunConfig, ScalingConfig
 from ray.tune import Trainable
 from ray.util import PublicAPI
 

python/ray/tune/tuner.py

@@ -13,7 +13,7 @@ class Tuner:
         tune_config: Tuning algorithm specific configs.
             Refer to ray.tune.tune_config.TuneConfig for more info.
         run_config: Runtime configuration that is specific to individual trials.
-            Refer to ray.ml.run_config.RunConfig for more info.
+            Refer to ray.ml.config.RunConfig for more info.
 
     Returns:
         ``ResultGrid`` object.