oracle-ads 2.13.16__tar.gz → 2.13.17rc0__tar.gz

{oracle_ads-2.13.16 → oracle_ads-2.13.17rc0}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: oracle_ads
-Version: 2.13.16
+Version: 2.13.17rc0
 Summary: Oracle Accelerated Data Science SDK
 Keywords: Oracle Cloud Infrastructure,OCI,Machine Learning,ML,Artificial Intelligence,AI,Data Science,Cloud,Oracle,GenAI,Generative AI,Forecast,Anomaly,Document Understanding,Anomaly Detection
 Author: Oracle Data Science

{oracle_ads-2.13.16 → oracle_ads-2.13.17rc0}/ads/aqua/client/client.py

@@ -582,6 +582,19 @@ class Client(BaseClient):
         payload = {**(payload or {}), "input": input}
         return self._request(payload=payload, headers=headers)
 
+    def fetch_data(self) -> Union[Dict[str, Any], Iterator[Mapping[str, Any]]]:
+        """Fetch Data in json format by sending a request to the endpoint.
+
+        Args:
+
+        Returns:
+            Union[Dict[str, Any], Iterator[Mapping[str, Any]]]: The server's response, typically including the data in JSON format.
+        """
+        # headers = {"Content-Type", "application/json"}
+        response = self._client.get(self.endpoint)
+        json_response = response.json()
+        return json_response
+
 
 class AsyncClient(BaseClient):
     """

{oracle_ads-2.13.16 → oracle_ads-2.13.17rc0}/ads/aqua/config/container_config.py

@@ -191,21 +191,28 @@ class AquaContainerConfig(Serializable):
                     additional_configurations.get("modelFormats")
                 )
 
-                # Parse environment variables from `additional_configurations`.
-                # Only keys present in the configuration will be added to the result.
-                config_keys = {
-                    "MODEL_DEPLOY_PREDICT_ENDPOINT": UNKNOWN,
-                    "MODEL_DEPLOY_HEALTH_ENDPOINT": UNKNOWN,
-                    "PORT": UNKNOWN,
-                    "HEALTH_CHECK_PORT": UNKNOWN,
-                    "VLLM_USE_V1": UNKNOWN,
-                }
-
-                env_vars = [
-                    {key: additional_configurations.get(key, default)}
-                    for key, default in config_keys.items()
-                    if key in additional_configurations
-                ]
+                # TODO: Remove the else condition once SMC env variable config is updated everywhere
+                if additional_configurations.get("env_vars", None):
+                    env_vars_dict = json.loads(
+                        additional_configurations.get("env_vars") or "{}"
+                    )
+                    env_vars = [
+                        {key: str(value)} for key, value in env_vars_dict.items()
+                    ]
+                else:
+                    config_keys = {
+                        "MODEL_DEPLOY_PREDICT_ENDPOINT": UNKNOWN,
+                        "MODEL_DEPLOY_HEALTH_ENDPOINT": UNKNOWN,
+                        "PORT": UNKNOWN,
+                        "HEALTH_CHECK_PORT": UNKNOWN,
+                        "VLLM_USE_V1": UNKNOWN,
+                    }
+
+                    env_vars = [
+                        {key: additional_configurations.get(key, default)}
+                        for key, default in config_keys.items()
+                        if key in additional_configurations
+                    ]
 
                 # Build container spec
                 container_item.spec = AquaContainerConfigSpec(

{oracle_ads-2.13.16 → oracle_ads-2.13.17rc0}/ads/aqua/extension/deployment_handler.py

@@ -373,6 +373,37 @@ class AquaDeploymentParamsHandler(AquaAPIhandler):
         )
 
 
+class AquaModelListHandler(AquaAPIhandler):
+    """Handler for Aqua model list params REST APIs.
+
+    Methods
+    -------
+    get(self, *args, **kwargs)
+        Validates parameters for the given model id.
+    """
+
+    @handle_exceptions
+    def get(self, model_deployment_id):
+        """
+        Handles get model list for the Active Model Deployment
+        Raises
+        ------
+        HTTPError
+            Raises HTTPError if inputs are missing or are invalid
+        """
+
+        self.set_header("Content-Type", "application/json")
+        endpoint: str = ""
+        model_deployment = AquaDeploymentApp().get(model_deployment_id)
+        endpoint = model_deployment.endpoint.rstrip("/") + "/predict/v1/models"
+        aqua_client = Client(endpoint=endpoint)
+        try:
+            list_model_result = aqua_client.fetch_data()
+            return self.finish(list_model_result)
+        except Exception as ex:
+            raise HTTPError(500, str(ex))
+
+
 __handlers__ = [
     ("deployments/?([^/]*)/params", AquaDeploymentParamsHandler),
     ("deployments/config/?([^/]*)", AquaDeploymentHandler),
@@ -381,4 +412,5 @@ __handlers__ = [
     ("deployments/?([^/]*)/activate", AquaDeploymentHandler),
     ("deployments/?([^/]*)/deactivate", AquaDeploymentHandler),
     ("inference/stream/?([^/]*)", AquaDeploymentStreamingInferenceHandler),
+    ("deployments/models/list/?([^/]*)", AquaModelListHandler),
 ]

{oracle_ads-2.13.16 → oracle_ads-2.13.17rc0}/ads/opctl/operator/lowcode/anomaly/model/automlx.py

@@ -24,8 +24,8 @@ class AutoMLXOperatorModel(AnomalyOperatorBaseModel):
     @runtime_dependency(
         module="automlx",
         err_msg=(
-            "Please run `pip3 install oracle-automlx>=23.4.1` and "
-            "`pip3 install oracle-automlx[classic]>=23.4.1` "
+            "Please run `pip3 install oracle-automlx>=25.3.0` and "
+            "`pip3 install oracle-automlx[classic]>=25.3.0` "
             "to install the required dependencies for automlx."
         ),
     )

{oracle_ads-2.13.16 → oracle_ads-2.13.17rc0}/ads/opctl/operator/lowcode/forecast/const.py

@@ -76,7 +76,7 @@ class ForecastOutputColumns(ExtendedEnum):
 
 AUTOMLX_METRIC_MAP = {
     "smape": "neg_sym_mean_abs_percent_error",
-    "mape": "neg_sym_mean_abs_percent_error",
+    "mape": "neg_mean_abs_percent_error",
     "mase": "neg_mean_abs_scaled_error",
     "mae": "neg_mean_absolute_error",
     "mse": "neg_mean_squared_error",

{oracle_ads-2.13.16 → oracle_ads-2.13.17rc0}/ads/opctl/operator/lowcode/forecast/model/automlx.py

@@ -28,7 +28,9 @@ from .forecast_datasets import ForecastDatasets, ForecastOutput
 
 logging.getLogger("report_creator").setLevel(logging.WARNING)
 AUTOMLX_N_ALGOS_TUNED = 4
-AUTOMLX_DEFAULT_SCORE_METRIC = "neg_sym_mean_abs_percent_error"
+AUTOMLX_DEFAULT_SCORE_METRIC = ['neg_sym_mean_abs_percent_error',
+                                'neg_mean_abs_percent_error',
+                                'neg_root_mean_squared_error']
 
 
 class AutoMLXOperatorModel(ForecastOperatorBaseModel):
@@ -45,10 +47,13 @@ class AutoMLXOperatorModel(ForecastOperatorBaseModel):
         model_kwargs_cleaned["n_algos_tuned"] = model_kwargs_cleaned.get(
             "n_algos_tuned", AUTOMLX_N_ALGOS_TUNED
         )
-        model_kwargs_cleaned["score_metric"] = AUTOMLX_METRIC_MAP.get(
-            self.spec.metric,
-            model_kwargs_cleaned.get("score_metric", AUTOMLX_DEFAULT_SCORE_METRIC),
-        )
+        metric_to_optimize = AUTOMLX_METRIC_MAP.get(self.spec.metric)
+        model_kwargs_cleaned["score_metric"] = AUTOMLX_DEFAULT_SCORE_METRIC
+        # The first score metric in the list will be the one for which the pipeline optimizes
+        if metric_to_optimize is not None:
+            model_kwargs_cleaned["score_metric"].remove(metric_to_optimize)
+            model_kwargs_cleaned["score_metric"].insert(0, metric_to_optimize)
+
         model_kwargs_cleaned.pop("task", None)
         time_budget = model_kwargs_cleaned.pop("time_budget", -1)
         model_kwargs_cleaned["preprocessing"] = (
@@ -70,7 +75,7 @@ class AutoMLXOperatorModel(ForecastOperatorBaseModel):
     @runtime_dependency(
         module="automlx",
         err_msg=(
-            "Please run `pip3 install oracle-automlx[forecasting]>=25.1.1` "
+            "Please run `pip3 install oracle-automlx[forecasting]>=25.3.0` "
             "to install the required dependencies for automlx."
         ),
     )
@@ -163,7 +168,7 @@ class AutoMLXOperatorModel(ForecastOperatorBaseModel):
                 self.models[s_id] = {}
                 self.models[s_id]["model"] = model
                 self.models[s_id]["le"] = self.le[s_id]
-                self.models[s_id]["score"] = self.get_validation_score_and_metric(model)
+                self.models[s_id]["score"] = self.get_all_metrics(model)
 
                 # In case of Naive model, model.forecast function call does not return confidence intervals.
                 if f"{target}_ci_upper" not in summary_frame:
@@ -518,26 +523,27 @@ class AutoMLXOperatorModel(ForecastOperatorBaseModel):
                 )
                 logger.debug(f"Full Traceback: {traceback.format_exc()}")
 
-    def get_validation_score_and_metric(self, model):
+    def get_all_metrics(self, model):
         trials = model.completed_trials_summary_
         model_params = model.selected_model_params_
         if len(trials) > 0:
-            score_col = [col for col in trials.columns if "Score" in col][0]
-            validation_score = trials[trials.Hyperparameters == model_params][
-                score_col
+            all_metrics = trials[trials.Hyperparameters == model_params][
+                "All Metrics"
             ].iloc[0]
         else:
-            validation_score = 0
-        return -1 * validation_score
+            all_metrics = {}
+        reverse_map = {v: k for k, v in AUTOMLX_METRIC_MAP.items()}
+        all_metrics = {reverse_map[key]: -1 * value for key, value in all_metrics.items() if key in reverse_map}
+        return all_metrics
 
     def generate_train_metrics(self) -> pd.DataFrame:
         """
-        Generate Training Metrics when fitted data is not available.
+        Generate Training Metrics for Automlx
         """
         total_metrics = pd.DataFrame()
         for s_id in self.forecast_output.list_series_ids():
             try:
-                metrics = {self.spec.metric.upper(): self.models[s_id]["score"]}
+                metrics = self.models[s_id]["score"]
                 metrics_df = pd.DataFrame.from_dict(
                     metrics, orient="index", columns=[s_id]
                 )

{oracle_ads-2.13.16 → oracle_ads-2.13.17rc0}/pyproject.toml

@@ -21,7 +21,7 @@ build-backend = "flit_core.buildapi"
 
 # Required
 name = "oracle_ads" # the install (PyPI) name; name for local build in [tool.flit.module] section below
-version = "2.13.16"
+version = "2.13.17rc0"
 
 # Optional
 description = "Oracle Accelerated Data Science SDK"