ODSC 29065/Model Deployment opctl for local dev (#134)

Author: z7ye

ads/model/deployment/model_deployment.py

@@ -863,9 +863,9 @@ def predict(
             and `json_input` required to be json serializable. If `auto_serialize_data` set
             to True, data will be serialized before sending to model deployment endpoint.
         model_name: str
-            Defaults to None. When the `Inference_server="triton"`, the name of the model to invoke.
+            Defaults to None. When the `inference_server="triton"`, the name of the model to invoke.
         model_version: str
-            Defaults to None. When the `Inference_server="triton"`, the version of the model to invoke.
+            Defaults to None. When the `inference_server="triton"`, the version of the model to invoke.
         kwargs:
             content_type: str
                 Used to indicate the media type of the resource.

ads/opctl/backend/ads_model_deployment.py

@@ -4,11 +4,16 @@
 # Copyright (c) 2022, 2023 Oracle and/or its affiliates.
 # Licensed under the Universal Permissive License v 1.0 as shown at https://oss.oracle.com/licenses/upl/
 
+import json
 from typing import Dict
-from ads.common.auth import create_signer, AuthContext
+
+from oci.data_science.models import ModelDeployment as OCIModelDeployment
+
+import ads
+from ads.common.auth import AuthContext, create_signer
 from ads.common.oci_client import OCIClientFactory
-from ads.opctl.backend.base import Backend
 from ads.model.deployment import ModelDeployment
+from ads.opctl.backend.base import Backend
 
 
 class ModelDeploymentBackend(Backend):
@@ -117,3 +122,16 @@ def watch(self) -> None:
             model_deployment.watch(
                 log_type=log_type, interval=interval, log_filter=log_filter
             )
+
+    def predict(self) -> None:
+        ocid = self.config["execution"].get("ocid")
+        data = self.config["execution"].get("payload")
+        model_name = self.config["execution"].get("model_name")
+        model_version = self.config["execution"].get("model_version")
+        with AuthContext(auth=self.auth_type, profile=self.profile):
+            model_deployment = ModelDeployment.from_id(ocid)
+            try:
+                data = json.loads(data)
+            except:
+                pass
+            print(model_deployment.predict(data=data, model_name=model_name, model_version=model_version))

ads/opctl/backend/base.py

@@ -7,19 +7,19 @@
 from abc import abstractmethod
 from typing import Dict
 
-from ads.common.auth import get_signer
+from ads.common.auth import create_signer
+
 
 
 class Backend:
     """Interface for backend"""
 
     def __init__(self, config: Dict) -> None:
         self.config = config
-        self.oci_auth = get_signer(
-            config["execution"].get("oci_config", None),
-            config["execution"].get("oci_profile", None),
-        )
+        self.auth_type = config["execution"].get("auth")
         self.profile = config["execution"].get("oci_profile", None)
+        self.oci_config = config["execution"].get("oci_config", None)
+
 
     @abstractmethod
     def run(self) -> Dict:
@@ -91,3 +91,13 @@ def run_diagnostics(self):
         """
         Implement Diagnostics check appropriate for the backend
         """
+
+    def predict(self) -> None:
+        """
+        Run model predict.
+
+        Returns
+        -------
+        None
+        """
+        raise NotImplementedError("`predict` has not been implemented yet.")