feat: Implement TrainerClient Backends & Local Process

kubeflow/trainer/api/trainer_client_test.py

@@ -28,9 +28,10 @@
 from unittest.mock import Mock, patch
 
 import pytest
-from kubeflow.trainer import TrainerClient
+from kubeflow.trainer.backends import K8SBackend
 from kubeflow.trainer.constants import constants
 from kubeflow.trainer.types import types
+from kubeflow.trainer.types.backends import K8SBackendConfig
 from kubeflow.trainer.utils import utils
 from kubeflow_trainer_api import models
 
@@ -100,7 +101,7 @@ def trainer_client(request):
             read_namespaced_pod_log=Mock(side_effect=mock_read_namespaced_pod_log),
         ),
     ):
-        yield TrainerClient()
+        yield K8SBackend(K8SBackendConfig())
 
 
 # --------------------------

kubeflow/trainer/backends/__init__.py

@@ -0,0 +1,28 @@
+# Copyright 2025 The Kubeflow Authors.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+from kubeflow.trainer.backends.k8s import K8SBackend
+from kubeflow.trainer.backends.local_process import LocalProcessBackend
+from kubeflow.trainer.types.backends import K8SBackendConfig, LocalProcessBackendConfig
+
+TRAINER_BACKEND_REGISTRY = {
+    "kubernetes": {
+        "backend_cls": K8SBackend,
+        "config_cls": K8SBackendConfig,
+    },
+    "local": {
+        "backend_cls": LocalProcessBackend,
+        "config_cls": LocalProcessBackendConfig,
+    }
+}

kubeflow/trainer/backends/base.py

@@ -0,0 +1,88 @@
+# Copyright 2025 The Kubeflow Authors.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+import abc
+
+from typing import Dict, List, Optional, Set
+from kubeflow.trainer.constants import constants
+from kubeflow.trainer.types import types
+
+
+class TrainingBackend(abc.ABC):
+
+    @abc.abstractmethod
+    def list_runtimes(self) -> List[types.Runtime]:
+        raise NotImplementedError()
+
+    @abc.abstractmethod
+    def get_runtime(self, name: str) -> Optional[types.Runtime]:
+        raise NotImplementedError()
+
+    @abc.abstractmethod
+    def train(self,
+              runtime: types.Runtime,
+              initializer: Optional[types.Initializer] = None,
+              trainer: Optional[types.RuntimeTrainer] = None,
+              ) -> str:
+        raise NotImplementedError()
+
+    @abc.abstractmethod
+    def list_jobs(
+            self, runtime: Optional[types.Runtime] = None
+    ) -> List[types.TrainJob]:
+        raise NotImplementedError()
+
+    @abc.abstractmethod
+    def get_job(self, name: str) -> Optional[types.TrainJob]:
+        raise NotImplementedError()
+
+    @abc.abstractmethod
+    def get_job_logs(self,
+                     name: str,
+                     follow: Optional[bool] = False,
+                     step: str = constants.NODE,
+                     node_rank: int = 0,
+                     ) -> Dict[str, str]:
+        raise NotImplementedError()
+
+    @abc.abstractmethod
+    def delete_job(self, name: str) -> None:
+        raise NotImplementedError()
+
+    @abc.abstractmethod
+    def wait_for_job_status(
+        self,
+        name: str,
+        status: Set[str] = {constants.TRAINJOB_COMPLETE},
+        timeout: int = 600,
+        polling_interval: int = 2,
+    ) -> types.TrainJob:
+        """Wait for TrainJob to reach the desired status
+
+        Args:
+            name: Name of the TrainJob.
+            status: Set of expected statuses. It must be subset of Created, Running, Complete, and
+                Failed statuses.
+            timeout: How many seconds to wait until TrainJob reaches one of the expected conditions.
+            polling_interval: The polling interval in seconds to check TrainJob status.
+
+        Returns:
+            TrainJob: The training job that reaches the desired status.
+
+        Raises:
+            ValueError: The input values are incorrect.
+            RuntimeError: Failed to get TrainJob or TrainJob reaches unexpected Failed status.
+            TimeoutError: Timeout to wait for TrainJob status.
+        """
+        raise NotImplementedError()