feat: Add support for developer connect based deployment

.kokoro/test-samples-impl.sh

@@ -34,7 +34,7 @@ env | grep KOKORO
 
 # Install nox
 # `virtualenv==20.26.6` is added for Python 3.7 compatibility
-python3 -m pip install --upgrade --quiet nox uv virtualenv==20.26.6
+python3 -m pip install --upgrade --quiet nox virtualenv==20.26.6
 
 # Use secrets acessor service account to get secrets
 if [[ -f "${KOKORO_GFILE_DIR}/secrets_viewer_service_account.json" ]]; then

google/cloud/aiplatform_v1/__init__.py

@@ -123,15 +123,21 @@
 from .types.content import LogprobsResult
 from .types.content import ModalityTokenCount
 from .types.content import ModelArmorConfig
+from .types.content import MultiSpeakerVoiceConfig
 from .types.content import Part
+from .types.content import PrebuiltVoiceConfig
+from .types.content import ReplicatedVoiceConfig
 from .types.content import RetrievalMetadata
 from .types.content import SafetyRating
 from .types.content import SafetySetting
 from .types.content import SearchEntryPoint
 from .types.content import Segment
+from .types.content import SpeakerVoiceConfig
+from .types.content import SpeechConfig
 from .types.content import UrlContextMetadata
 from .types.content import UrlMetadata
 from .types.content import VideoMetadata
+from .types.content import VoiceConfig
 from .types.content import HarmCategory
 from .types.content import Modality
 from .types.context import Context
@@ -837,6 +843,8 @@
 from .types.prediction_service import DirectPredictResponse
 from .types.prediction_service import DirectRawPredictRequest
 from .types.prediction_service import DirectRawPredictResponse
+from .types.prediction_service import EmbedContentRequest
+from .types.prediction_service import EmbedContentResponse
 from .types.prediction_service import ExplainRequest
 from .types.prediction_service import ExplainResponse
 from .types.prediction_service import GenerateContentRequest
@@ -963,6 +971,9 @@
 from .types.tool import FunctionCallingConfig
 from .types.tool import FunctionDeclaration
 from .types.tool import FunctionResponse
+from .types.tool import FunctionResponseBlob
+from .types.tool import FunctionResponseFileData
+from .types.tool import FunctionResponsePart
 from .types.tool import GoogleMaps
 from .types.tool import GoogleSearchRetrieval
 from .types.tool import RagRetrievalConfig
@@ -980,6 +991,7 @@
 from .types.training_pipeline import StratifiedSplit
 from .types.training_pipeline import TimestampSplit
 from .types.training_pipeline import TrainingPipeline
+from .types.tuning_job import PreTunedModel
 from .types.tuning_job import SupervisedHyperParameters
 from .types.tuning_job import SupervisedTuningDatasetDistribution
 from .types.tuning_job import SupervisedTuningDataStats
@@ -995,6 +1007,7 @@
 from .types.types import StringArray
 from .types.types import Tensor
 from .types.unmanaged_container_model import UnmanagedContainerModel
+from .types.usage_metadata import UsageMetadata
 from .types.user_action_reference import UserActionReference
 from .types.value import Value
 from .types.vertex_rag_data import CorpusStatus
@@ -1348,6 +1361,8 @@
     "DnsPeeringConfig",
     "DoubleArray",
     "DynamicRetrievalConfig",
+    "EmbedContentRequest",
+    "EmbedContentResponse",
     "EncryptionSpec",
     "Endpoint",
     "EndpointServiceClient",
@@ -1438,6 +1453,9 @@
     "FunctionCallingConfig",
     "FunctionDeclaration",
     "FunctionResponse",
+    "FunctionResponseBlob",
+    "FunctionResponseFileData",
+    "FunctionResponsePart",
     "GcsDestination",
     "GcsSource",
     "GenAiAdvancedFeaturesConfig",
@@ -1689,6 +1707,7 @@
     "ModelServiceClient",
     "ModelSourceInfo",
     "ModelVersionCheckpoint",
+    "MultiSpeakerVoiceConfig",
     "MutateDeployedIndexOperationMetadata",
     "MutateDeployedIndexRequest",
     "MutateDeployedIndexResponse",
@@ -1752,6 +1771,8 @@
     "PointwiseMetricSpec",
     "Port",
     "PostStartupScriptConfig",
+    "PreTunedModel",
+    "PrebuiltVoiceConfig",
     "PredefinedSplit",
     "PredictRequest",
     "PredictRequestResponseLoggingConfig",
@@ -1840,6 +1861,7 @@
     "RemoveContextChildrenResponse",
     "RemoveDatapointsRequest",
     "RemoveDatapointsResponse",
+    "ReplicatedVoiceConfig",
     "ReservationAffinity",
     "ResourcePool",
     "ResourceRuntime",
@@ -1893,9 +1915,11 @@
     "ShieldedVmConfig",
     "SlackSource",
     "SmoothGradConfig",
+    "SpeakerVoiceConfig",
     "SpecialistPool",
     "SpecialistPoolServiceClient",
     "SpeculativeDecodingSpec",
+    "SpeechConfig",
     "StartNotebookRuntimeOperationMetadata",
     "StartNotebookRuntimeRequest",
     "StartNotebookRuntimeResponse",
@@ -2061,6 +2085,7 @@
     "UrlContext",
     "UrlContextMetadata",
     "UrlMetadata",
+    "UsageMetadata",
     "UserActionReference",
     "Value",
     "VertexAISearch",
@@ -2070,6 +2095,7 @@
     "VertexRagStore",
     "VideoMetadata",
     "VizierServiceClient",
+    "VoiceConfig",
     "WorkerPoolSpec",
     "WriteFeatureValuesPayload",
     "WriteFeatureValuesRequest",

google/cloud/aiplatform_v1/gapic_metadata.json

@@ -4091,6 +4091,11 @@
                 "direct_raw_predict"
               ]
             },
+            "EmbedContent": {
+              "methods": [
+                "embed_content"
+              ]
+            },
             "Explain": {
               "methods": [
                 "explain"
@@ -4161,6 +4166,11 @@
                 "direct_raw_predict"
               ]
             },
+            "EmbedContent": {
+              "methods": [
+                "embed_content"
+              ]
+            },
             "Explain": {
               "methods": [
                 "explain"
@@ -4231,6 +4241,11 @@
                 "direct_raw_predict"
               ]
             },
+            "EmbedContent": {
+              "methods": [
+                "embed_content"
+              ]
+            },
             "Explain": {
               "methods": [
                 "explain"

google/cloud/aiplatform_v1/services/prediction_service/async_client.py

@@ -50,9 +50,11 @@
 
 from google.api import httpbody_pb2  # type: ignore
 from google.cloud.aiplatform_v1.types import content
+from google.cloud.aiplatform_v1.types import content as gca_content
 from google.cloud.aiplatform_v1.types import explanation
 from google.cloud.aiplatform_v1.types import prediction_service
 from google.cloud.aiplatform_v1.types import types
+from google.cloud.aiplatform_v1.types import usage_metadata
 from google.cloud.location import locations_pb2  # type: ignore
 from google.iam.v1 import iam_policy_pb2  # type: ignore
 from google.iam.v1 import policy_pb2  # type: ignore
@@ -1931,6 +1933,127 @@ async def sample_stream_generate_content():
         # Done; return the response.
         return response
 
+    async def embed_content(
+        self,
+        request: Optional[Union[prediction_service.EmbedContentRequest, dict]] = None,
+        *,
+        model: Optional[str] = None,
+        content: Optional[gca_content.Content] = None,
+        retry: OptionalRetry = gapic_v1.method.DEFAULT,
+        timeout: Union[float, object] = gapic_v1.method.DEFAULT,
+        metadata: Sequence[Tuple[str, Union[str, bytes]]] = (),
+    ) -> prediction_service.EmbedContentResponse:
+        r"""Embed content with multimodal inputs.
+
+        .. code-block:: python
+
+            # This snippet has been automatically generated and should be regarded as a
+            # code template only.
+            # It will require modifications to work:
+            # - It may require correct/in-range values for request initialization.
+            # - It may require specifying regional endpoints when creating the service
+            #   client as shown in:
+            #   https://googleapis.dev/python/google-api-core/latest/client_options.html
+            from google.cloud import aiplatform_v1
+
+            async def sample_embed_content():
+                # Create a client
+                client = aiplatform_v1.PredictionServiceAsyncClient()
+
+                # Initialize request argument(s)
+                request = aiplatform_v1.EmbedContentRequest(
+                )
+
+                # Make the request
+                response = await client.embed_content(request=request)
+
+                # Handle the response
+                print(response)
+
+        Args:
+            request (Optional[Union[google.cloud.aiplatform_v1.types.EmbedContentRequest, dict]]):
+                The request object. Request message for
+                [PredictionService.EmbedContent][google.cloud.aiplatform.v1.PredictionService.EmbedContent].
+            model (:class:`str`):
+                Required. The name of the publisher model requested to
+                serve the prediction. Format:
+                ``projects/{project}/locations/{location}/publishers/*/models/*``
+
+                This corresponds to the ``model`` field
+                on the ``request`` instance; if ``request`` is provided, this
+                should not be set.
+            content (:class:`google.cloud.aiplatform_v1.types.Content`):
+                Required. Input content to be
+                embedded. Required.
+
+                This corresponds to the ``content`` field
+                on the ``request`` instance; if ``request`` is provided, this
+                should not be set.
+            retry (google.api_core.retry_async.AsyncRetry): Designation of what errors, if any,
+                should be retried.
+            timeout (float): The timeout for this request.
+            metadata (Sequence[Tuple[str, Union[str, bytes]]]): Key/value pairs which should be
+                sent along with the request as metadata. Normally, each value must be of type `str`,
+                but for metadata keys ending with the suffix `-bin`, the corresponding values must
+                be of type `bytes`.
+
+        Returns:
+            google.cloud.aiplatform_v1.types.EmbedContentResponse:
+                Response message for
+                   [PredictionService.EmbedContent][google.cloud.aiplatform.v1.PredictionService.EmbedContent].
+
+        """
+        # Create or coerce a protobuf request object.
+        # - Quick check: If we got a request object, we should *not* have
+        #   gotten any keyword arguments that map to the request.
+        flattened_params = [model, content]
+        has_flattened_params = (
+            len([param for param in flattened_params if param is not None]) > 0
+        )
+        if request is not None and has_flattened_params:
+            raise ValueError(
+                "If the `request` argument is set, then none of "
+                "the individual field arguments should be set."
+            )
+
+        # - Use the request object if provided (there's no risk of modifying the input as
+        #   there are no flattened fields), or create one.
+        if not isinstance(request, prediction_service.EmbedContentRequest):
+            request = prediction_service.EmbedContentRequest(request)
+
+        # If we have keyword arguments corresponding to fields on the
+        # request, apply these.
+        if model is not None:
+            request.model = model
+        if content is not None:
+            request.content = content
+
+        # Wrap the RPC method; this adds retry and timeout information,
+        # and friendly error handling.
+        rpc = self._client._transport._wrapped_methods[
+            self._client._transport.embed_content
+        ]
+
+        # Certain fields should be provided within the metadata header;
+        # add these here.
+        metadata = tuple(metadata) + (
+            gapic_v1.routing_header.to_grpc_metadata((("model", request.model),)),
+        )
+
+        # Validate the universe domain.
+        self._client._validate_universe_domain()
+
+        # Send the request.
+        response = await rpc(
+            request,
+            retry=retry,
+            timeout=timeout,
+            metadata=metadata,
+        )
+
+        # Done; return the response.
+        return response
+
     async def list_operations(
         self,
         request: Optional[operations_pb2.ListOperationsRequest] = None,