feat: add support for version 2.42 for RoV Bigquery read/write

PiperOrigin-RevId: 738125614

Author: vertex-sdk-bot

google/cloud/aiplatform/vertex_ray/bigquery_datasink.py

@@ -51,7 +51,7 @@
 )
 
 
-# BigQuery write for Ray 2.33.0 and 2.9.3
+# BigQuery write for Ray 2.42.0, 2.33.0, and 2.9.3
 if Datasink is None:
     _BigQueryDatasink = None
 else:

google/cloud/aiplatform/vertex_ray/data.py

@@ -57,26 +57,25 @@ def read_bigquery(
         dataset: The name of the dataset hosted in BigQuery in the format of
             ``dataset_id.table_id``. Both the dataset_id and table_id must exist
             otherwise an exception will be raised.
-        query: The query to execute.
-            The dataset is created from the results of executing the query if provided.
-            Otherwise, the entire dataset is read. For query syntax guidelines, see
+        query: The query to execute. The dataset is created from the results of
+            executing the query if provided. Otherwise, the entire dataset is read.
+            For query syntax guidelines, see
             https://cloud.google.com/bigquery/docs/reference/standard-sql/query-syntax
-        parallelism:
-            2.9.3: The requested parallelism of the read. If -1, it will be
-            automatically chosen based on the available cluster resources
-            and estimated in-memory data size.
-            2.33.0: This argument is deprecated. Use ``override_num_blocks`` argument.
+        parallelism: 2.33.0, 2.42.0: This argument is deprecated. Use
+            ``override_num_blocks`` argument. 2.9.3: The requested parallelism of
+            the read. If -1, it will be automatically chosen based on the available
+            cluster resources and estimated in-memory data size.
         ray_remote_args: kwargs passed to ray.remote in the read tasks.
-        concurrency: Not supported in 2.9.3.
-            2.33.0: The maximum number of Ray tasks to run concurrently. Set this
-            to control number of tasks to run concurrently. This doesn't change the
-            total number of tasks run or the total number of output blocks. By default,
-            concurrency is dynamically decided based on the available resources.
-        override_num_blocks: Not supported in 2.9.3.
-            2.33.0: Override the number of output blocks from all read tasks.
-            By default, the number of output blocks is dynamically decided based on
-            input data size and available resources. You shouldn't manually set this
-            value in most cases.
+        concurrency: Supported for 2.33.0 and 2.42.0 only: The maximum number of
+            Ray tasks to run concurrently. Set this to control number of tasks to
+            run concurrently. This doesn't change the total number of tasks run or
+            the total number of output blocks. By default, concurrency is
+            dynamically decided based on the available resources.
+        override_num_blocks: Supported for 2.33.0 and 2.42.0 only: Override the
+            number of output blocks from all read tasks. By default, the number of
+            output blocks is dynamically decided based on input data size and
+            available resources. You shouldn't manually set this value in most
+            cases.
 
     Returns:
         Dataset producing rows from the results of executing the query
@@ -96,7 +95,7 @@ def read_bigquery(
             parallelism=parallelism,
             ray_remote_args=ray_remote_args,
         )
-    elif ray.__version__ == "2.33.0":
+    elif ray.__version__ in ("2.33.0", "2.42.0"):
         return ray.data.read_datasource(
             datasource=datasource,
             parallelism=parallelism,
@@ -107,7 +106,7 @@ def read_bigquery(
     else:
         raise ImportError(
             f"[Ray on Vertex AI]: Unsupported version {ray.__version__}."
-            + "Only 2.33.0 and 2.9.3 are supported."
+            + "Only 2.42.0, 2.33.0, and 2.9.3 are supported."
         )
 
 
@@ -135,19 +134,19 @@ def write_bigquery(
             The default number of retries is 10.
         ray_remote_args: kwargs passed to ray.remote in the write tasks.
         overwrite_table: Not supported in 2.9.3.
-            2.33.0: Whether the write will overwrite the table if it already
+            2.33.0, 2.42.0: Whether the write will overwrite the table if it already
             exists. The default behavior is to overwrite the table.
             If false, will append to the table if it exists.
         concurrency: Not supported in 2.9.3.
-            2.33.0: The maximum number of Ray tasks to run concurrently. Set this
+            2.33.0, 2.42.0: The maximum number of Ray tasks to run concurrently. Set this
             to control number of tasks to run concurrently. This doesn't change the
             total number of tasks run or the total number of output blocks. By default,
             concurrency is dynamically decided based on the available resources.
     """
     if ray.__version__ == "2.4.0":
         raise RuntimeError(_V2_4_WARNING_MESSAGE)
 
-    elif ray.__version__ == "2.9.3" or ray.__version__ == "2.33.0":
+    elif ray.__version__ in ("2.9.3", "2.33.0", "2.42.0"):
         if ray.__version__ == "2.9.3":
             warnings.warn(_V2_9_WARNING_MESSAGE, DeprecationWarning, stacklevel=1)
         if ray_remote_args is None:
@@ -174,7 +173,7 @@ def write_bigquery(
                 datasink=datasink,
                 ray_remote_args=ray_remote_args,
             )
-        elif ray.__version__ == "2.33.0":
+        elif ray.__version__ in ("2.33.0", "2.42.0"):
             datasink = _BigQueryDatasink(
                 project_id=project_id,
                 dataset=dataset,
@@ -189,5 +188,5 @@ def write_bigquery(
     else:
         raise ImportError(
             f"[Ray on Vertex AI]: Unsupported version {ray.__version__}."
-            + "Only 2.33.0 and 2.9.3 are supported."
+            + "Only 2.42.0, 2.33.0 and 2.9.3 are supported."
         )

google/cloud/aiplatform/vertex_ray/util/_validation_utils.py

@@ -25,15 +25,17 @@
 from google.cloud.aiplatform import initializer
 from google.cloud.aiplatform.utils import resource_manager_utils
 
-SUPPORTED_RAY_VERSIONS = immutabledict({"2.9": "2.9.3", "2.33": "2.33.0"})
+SUPPORTED_RAY_VERSIONS = immutabledict(
+    {"2.9": "2.9.3", "2.33": "2.33.0", "2.42": "2.42.0"}
+)
 SUPPORTED_PY_VERSION = ["3.10"]
 _V2_4_WARNING_MESSAGE = (
-    "After google-cloud-aiplatform>1.53.0, using Ray version = 2.4 will result in an error. "
-    "Please use Ray version = 2.33.0 (default)."
+    "After google-cloud-aiplatform>1.53.0, using Ray version = 2.4 will result"
+    " in an error. Please use Ray version = 2.33.0 or 2.42.0 (default) instead."
 )
 _V2_9_WARNING_MESSAGE = (
     "In March 2025, using Ray version = 2.9 will result in an error. "
-    "Please use Ray version = 2.33.0 (default) instead."
+    "Please use Ray version = 2.33.0 or 2.42.0 (default) instead."
 )
 
 

tests/unit/vertex_ray/test_cluster_init.py

@@ -54,7 +54,7 @@
 
 _TEST_V2_4_WARNING_MESSAGE = (
     "After google-cloud-aiplatform>1.53.0, using Ray version = 2.4 will result"
-    " in an error. Please use Ray version = 2.33.0 (default)."
+    " in an error. Please use Ray version = 2.33.0 or 2.42.0 (default) instead."
 )