oracle-ads 2.12.2__py3-none-any.whl → 2.12.4__py3-none-any.whl

ads/aqua/common/enums.py

@@ -52,6 +52,7 @@ class InferenceContainerTypeFamily(str, metaclass=ExtendedEnumMeta):
     AQUA_VLLM_CONTAINER_FAMILY = "odsc-vllm-serving"
     AQUA_TGI_CONTAINER_FAMILY = "odsc-tgi-serving"
     AQUA_LLAMA_CPP_CONTAINER_FAMILY = "odsc-llama-cpp-serving"
+    AQUA_TEI_CONTAINER_FAMILY = "odsc-tei-serving"
 
 
 class InferenceContainerParamType(str, metaclass=ExtendedEnumMeta):
@@ -80,3 +81,11 @@ class RqsAdditionalDetails(str, metaclass=ExtendedEnumMeta):
     MODEL_VERSION_SET_NAME = "modelVersionSetName"
     PROJECT_ID = "projectId"
     VERSION_LABEL = "versionLabel"
+
+
+class TextEmbeddingInferenceContainerParams(str, metaclass=ExtendedEnumMeta):
+    """Contains a subset of params that are required for enabling model deployment in OCI Data Science. More options
+    are available at https://huggingface.co/docs/text-embeddings-inference/en/cli_arguments"""
+
+    MODEL_ID = "model-id"
+    PORT = "port"

ads/aqua/common/utils.py

@@ -35,6 +35,7 @@ from ads.aqua.common.enums import (
     InferenceContainerParamType,
     InferenceContainerType,
     RqsAdditionalDetails,
+    TextEmbeddingInferenceContainerParams,
 )
 from ads.aqua.common.errors import (
     AquaFileNotFoundError,
@@ -51,6 +52,7 @@ from ads.aqua.constants import (
     MODEL_BY_REFERENCE_OSS_PATH_KEY,
     SERVICE_MANAGED_CONTAINER_URI_SCHEME,
     SUPPORTED_FILE_FORMATS,
+    TEI_CONTAINER_DEFAULT_HOST,
     TGI_INFERENCE_RESTRICTED_PARAMS,
     UNKNOWN,
     UNKNOWN_JSON_STR,
@@ -63,7 +65,12 @@ from ads.common.extended_enum import ExtendedEnumMeta
 from ads.common.object_storage_details import ObjectStorageDetails
 from ads.common.oci_resource import SEARCH_TYPE, OCIResource
 from ads.common.utils import copy_file, get_console_link, upload_to_os
-from ads.config import AQUA_SERVICE_MODELS_BUCKET, CONDA_BUCKET_NS, TENANCY_OCID
+from ads.config import (
+    AQUA_MODEL_DEPLOYMENT_FOLDER,
+    AQUA_SERVICE_MODELS_BUCKET,
+    CONDA_BUCKET_NS,
+    TENANCY_OCID,
+)
 from ads.model import DataScienceModel, ModelVersionSet
 
 logger = logging.getLogger("ads.aqua")
@@ -569,15 +576,13 @@ def get_container_image(
         A dict of allowed configs.
     """
 
+    container_image = UNKNOWN
     config = config_file_name or get_container_config()
     config_file_name = service_config_path()
 
     if container_type not in config:
-        raise AquaValueError(
-            f"{config_file_name} does not have config details for model: {container_type}"
-        )
+        return UNKNOWN
 
-    container_image = None
     mapping = config[container_type]
     versions = [obj["version"] for obj in mapping]
     # assumes numbered versions, update if `latest` is used
@@ -1071,7 +1076,6 @@ def list_hf_models(query: str) -> List[str]:
     try:
         models = HfApi().list_models(
             model_name=query,
-            task="text-generation",
             sort="downloads",
             direction=-1,
             limit=20,
@@ -1079,3 +1083,76 @@ def list_hf_models(query: str) -> List[str]:
         return [model.id for model in models if model.disabled is None]
     except HfHubHTTPError as err:
         raise format_hf_custom_error_message(err) from err
+
+
+def generate_tei_cmd_var(os_path: str) -> List[str]:
+    """This utility functions generates CMD params for Text Embedding Inference container. Only the
+    essential parameters for OCI model deployment are added, defaults are used for the rest.
+    Parameters
+    ----------
+    os_path: str
+        OCI bucket path where the model artifacts are uploaded - oci://bucket@namespace/prefix
+
+    Returns
+    -------
+        cmd_var:
+            List of command line arguments
+    """
+
+    cmd_prefix = "--"
+    cmd_var = [
+        f"{cmd_prefix}{TextEmbeddingInferenceContainerParams.MODEL_ID}",
+        f"{AQUA_MODEL_DEPLOYMENT_FOLDER}{ObjectStorageDetails.from_path(os_path.rstrip('/')).filepath}/",
+        f"{cmd_prefix}{TextEmbeddingInferenceContainerParams.PORT}",
+        TEI_CONTAINER_DEFAULT_HOST,
+    ]
+
+    return cmd_var
+
+
+def parse_cmd_var(cmd_list: List[str]) -> dict:
+    """Helper functions that parses a list into a key-value dictionary. The list contains keys separated by the prefix
+    '--' and the value of the key is the subsequent element.
+    """
+    parsed_cmd = {}
+
+    for i, cmd in enumerate(cmd_list):
+        if cmd.startswith("--"):
+            if i + 1 < len(cmd_list) and not cmd_list[i + 1].startswith("--"):
+                parsed_cmd[cmd] = cmd_list[i + 1]
+                i += 1
+            else:
+                parsed_cmd[cmd] = None
+    return parsed_cmd
+
+
+def validate_cmd_var(cmd_var: List[str], overrides: List[str]) -> List[str]:
+    """This function accepts two lists of parameters and combines them. If the second list shares the common parameter
+    names/keys, then it raises an error.
+    Parameters
+    ----------
+    cmd_var: List[str]
+        Default list of parameters
+    overrides: List[str]
+        List of parameters to override
+    Returns
+    -------
+        List[str] of combined parameters
+    """
+    cmd_var = [str(x) for x in cmd_var]
+    if not overrides:
+        return cmd_var
+    overrides = [str(x) for x in overrides]
+
+    cmd_dict = parse_cmd_var(cmd_var)
+    overrides_dict = parse_cmd_var(overrides)
+
+    # check for conflicts
+    common_keys = set(cmd_dict.keys()) & set(overrides_dict.keys())
+    if common_keys:
+        raise AquaValueError(
+            f"The following CMD input cannot be overridden for model deployment: {', '.join(common_keys)}"
+        )
+
+    combined_cmd_var = cmd_var + overrides
+    return combined_cmd_var

ads/aqua/config/config.py

@@ -29,19 +29,3 @@ def get_evaluation_service_config(
         .get(ContainerSpec.CONTAINER_SPEC, {})
         .get(container, {})
     )
-
-
-# TODO: move this to global config.json in object storage
-def get_finetuning_config_defaults():
-    """Generate and return the fine-tuning default configuration dictionary."""
-    return {
-        "shape": {
-            "VM.GPU.A10.1": {"batch_size": 1, "replica": "1-10"},
-            "VM.GPU.A10.2": {"batch_size": 1, "replica": "1-10"},
-            "BM.GPU.A10.4": {"batch_size": 1, "replica": 1},
-            "BM.GPU4.8": {"batch_size": 4, "replica": 1},
-            "BM.GPU.L40S-NC.4": {"batch_size": 4, "replica": 1},
-            "BM.GPU.A100-v2.8": {"batch_size": 6, "replica": 1},
-            "BM.GPU.H100.8": {"batch_size": 6, "replica": 1},
-        }
-    }

ads/aqua/constants.py

@@ -27,6 +27,7 @@ NB_SESSION_IDENTIFIER = "NB_SESSION_OCID"
 LIFECYCLE_DETAILS_MISSING_JOBRUN = "The associated JobRun resource has been deleted."
 READY_TO_DEPLOY_STATUS = "ACTIVE"
 READY_TO_FINE_TUNE_STATUS = "TRUE"
+PRIVATE_ENDPOINT_TYPE = "MODEL_DEPLOYMENT"
 AQUA_GA_LIST = ["id19sfcrra6z"]
 AQUA_MODEL_TYPE_SERVICE = "service"
 AQUA_MODEL_TYPE_CUSTOM = "custom"
@@ -79,3 +80,4 @@ LLAMA_CPP_INFERENCE_RESTRICTED_PARAMS = {
     "--port",
     "--host",
 }
+TEI_CONTAINER_DEFAULT_HOST = "8080"

ads/aqua/evaluation/entities.py

@@ -9,19 +9,18 @@ aqua.evaluation.entities
 This module contains dataclasses for aqua evaluation.
 """
 
-from dataclasses import dataclass, field
-from typing import List, Optional, Union
+from pydantic import Field
+from typing import Any, Dict, List, Optional, Union
 
 from ads.aqua.data import AquaResourceIdentifier
-from ads.common.serializer import DataClassSerializable
+from ads.aqua.config.utils.serializer import Serializable
 
 
-@dataclass(repr=False)
-class CreateAquaEvaluationDetails(DataClassSerializable):
-    """Dataclass to create aqua model evaluation.
+class CreateAquaEvaluationDetails(Serializable):
+    """Class for creating aqua model evaluation.
 
-    Fields
-    ------
+    Properties
+    ----------
     evaluation_source_id: str
         The evaluation source id. Must be either model or model deployment ocid.
     evaluation_name: str
@@ -83,69 +82,64 @@ class CreateAquaEvaluationDetails(DataClassSerializable):
     ocpus: Optional[float] = None
     log_group_id: Optional[str] = None
     log_id: Optional[str] = None
-    metrics: Optional[List] = None
+    metrics: Optional[List[str]] = None
     force_overwrite: Optional[bool] = False
 
+    class Config:
+        extra = "ignore"
 
-@dataclass(repr=False)
-class AquaEvalReport(DataClassSerializable):
+class AquaEvalReport(Serializable):
     evaluation_id: str = ""
     content: str = ""
 
+    class Config:
+        extra = "ignore"
 
-@dataclass(repr=False)
-class ModelParams(DataClassSerializable):
-    max_tokens: str = ""
-    top_p: str = ""
-    top_k: str = ""
-    temperature: str = ""
-    presence_penalty: Optional[float] = 0.0
-    frequency_penalty: Optional[float] = 0.0
-    stop: Optional[Union[str, List[str]]] = field(default_factory=list)
-    model: Optional[str] = "odsc-llm"
-
-
-@dataclass(repr=False)
-class AquaEvalParams(ModelParams, DataClassSerializable):
+class AquaEvalParams(Serializable):
     shape: str = ""
     dataset_path: str = ""
     report_path: str = ""
 
+    class Config:
+        extra = "allow"
 
-@dataclass(repr=False)
-class AquaEvalMetric(DataClassSerializable):
+class AquaEvalMetric(Serializable):
     key: str
     name: str
     description: str = ""
 
+    class Config:
+        extra = "ignore"
 
-@dataclass(repr=False)
-class AquaEvalMetricSummary(DataClassSerializable):
+class AquaEvalMetricSummary(Serializable):
     metric: str = ""
     score: str = ""
     grade: str = ""
 
+    class Config:
+        extra = "ignore"
 
-@dataclass(repr=False)
-class AquaEvalMetrics(DataClassSerializable):
+class AquaEvalMetrics(Serializable):
     id: str
     report: str
-    metric_results: List[AquaEvalMetric] = field(default_factory=list)
-    metric_summary_result: List[AquaEvalMetricSummary] = field(default_factory=list)
+    metric_results: List[AquaEvalMetric] = Field(default_factory=list)
+    metric_summary_result: List[AquaEvalMetricSummary] = Field(default_factory=list)
 
+    class Config:
+        extra = "ignore"
 
-@dataclass(repr=False)
-class AquaEvaluationCommands(DataClassSerializable):
+class AquaEvaluationCommands(Serializable):
     evaluation_id: str
     evaluation_target_id: str
-    input_data: dict
-    metrics: list
+    input_data: Dict[str, Any]
+    metrics: List[str]
     output_dir: str
-    params: dict
+    params: Dict[str, Any]
 
+    class Config:
+        extra = "ignore"
 
-@dataclass(repr=False)
-class AquaEvaluationSummary(DataClassSerializable):
+class AquaEvaluationSummary(Serializable):
     """Represents a summary of Aqua evalution."""
 
     id: str
@@ -154,17 +148,18 @@ class AquaEvaluationSummary(DataClassSerializable):
     lifecycle_state: str
     lifecycle_details: str
     time_created: str
-    tags: dict
-    experiment: AquaResourceIdentifier = field(default_factory=AquaResourceIdentifier)
-    source: AquaResourceIdentifier = field(default_factory=AquaResourceIdentifier)
-    job: AquaResourceIdentifier = field(default_factory=AquaResourceIdentifier)
-    parameters: AquaEvalParams = field(default_factory=AquaEvalParams)
+    tags: Dict[str, Any]
+    experiment: AquaResourceIdentifier = Field(default_factory=AquaResourceIdentifier)
+    source: AquaResourceIdentifier = Field(default_factory=AquaResourceIdentifier)
+    job: AquaResourceIdentifier = Field(default_factory=AquaResourceIdentifier)
+    parameters: AquaEvalParams = Field(default_factory=AquaEvalParams)
 
+    class Config:
+        extra = "ignore"
 
-@dataclass(repr=False)
-class AquaEvaluationDetail(AquaEvaluationSummary, DataClassSerializable):
+class AquaEvaluationDetail(AquaEvaluationSummary):
     """Represents a details of Aqua evalution."""
 
-    log_group: AquaResourceIdentifier = field(default_factory=AquaResourceIdentifier)
-    log: AquaResourceIdentifier = field(default_factory=AquaResourceIdentifier)
-    introspection: dict = field(default_factory=dict)
+    log_group: AquaResourceIdentifier = Field(default_factory=AquaResourceIdentifier)
+    log: AquaResourceIdentifier = Field(default_factory=AquaResourceIdentifier)
+    introspection: dict = Field(default_factory=dict)

ads/aqua/evaluation/evaluation.py

@@ -7,7 +7,6 @@ import os
 import re
 import tempfile
 from concurrent.futures import ThreadPoolExecutor, as_completed
-from dataclasses import asdict, fields
 from datetime import datetime, timedelta
 from pathlib import Path
 from threading import Lock
@@ -46,7 +45,6 @@ from ads.aqua.common.utils import (
     upload_local_to_os,
 )
 from ads.aqua.config.config import get_evaluation_service_config
-from ads.aqua.config.evaluation.evaluation_service_config import EvaluationServiceConfig
 from ads.aqua.constants import (
     CONSOLE_LINK_RESOURCE_TYPE_MAPPING,
     EVALUATION_REPORT,
@@ -75,7 +73,6 @@ from ads.aqua.evaluation.entities import (
     AquaEvaluationSummary,
     AquaResourceIdentifier,
     CreateAquaEvaluationDetails,
-    ModelParams,
 )
 from ads.aqua.evaluation.errors import EVALUATION_JOB_EXIT_CODE_MESSAGE
 from ads.aqua.ui import AquaContainerConfig
@@ -161,10 +158,11 @@ class AquaEvaluationApp(AquaApp):
             try:
                 create_aqua_evaluation_details = CreateAquaEvaluationDetails(**kwargs)
             except Exception as ex:
+                custom_errors = {
+                    ".".join(map(str, e["loc"])): e["msg"] for e in json.loads(ex.json())
+                }
                 raise AquaValueError(
-                    "Invalid create evaluation parameters. "
-                    "Allowable parameters are: "
-                    f"{', '.join([field.name for field in fields(CreateAquaEvaluationDetails)])}."
+                    f"Invalid create evaluation parameters. Error details: {custom_errors}."
                 ) from ex
 
         if not is_valid_ocid(create_aqua_evaluation_details.evaluation_source_id):
@@ -175,15 +173,7 @@ class AquaEvaluationApp(AquaApp):
 
         # The model to evaluate
         evaluation_source = None
-        # The evaluation service config
-        evaluation_config: EvaluationServiceConfig = get_evaluation_service_config()
-        # The evaluation inference configuration. The inference configuration will be extracted
-        # based on the inferencing container family.
         eval_inference_configuration: Dict = {}
-        # The evaluation inference model sampling params. The system parameters that will not be
-        # visible for user, but will be applied implicitly for evaluation. The service model params
-        # will be extracted based on the container family and version.
-        eval_inference_service_model_params: Dict = {}
 
         if (
             DataScienceResource.MODEL_DEPLOYMENT
@@ -200,29 +190,14 @@ class AquaEvaluationApp(AquaApp):
                     runtime = ModelDeploymentContainerRuntime.from_dict(
                         evaluation_source.runtime.to_dict()
                     )
-                    container_config = AquaContainerConfig.from_container_index_json(
+                    inference_config = AquaContainerConfig.from_container_index_json(
                         enable_spec=True
-                    )
-                    for (
-                        inference_container_family,
-                        inference_container_info,
-                    ) in container_config.inference.items():
-                        if (
-                            inference_container_info.name
-                            == runtime.image[: runtime.image.rfind(":")]
-                        ):
+                    ).inference
+                    for container in inference_config.values():
+                        if container.name == runtime.image[: runtime.image.rfind(":")]:
                             eval_inference_configuration = (
-                                evaluation_config.get_merged_inference_params(
-                                    inference_container_family
-                                ).to_dict()
-                            )
-                            eval_inference_service_model_params = (
-                                evaluation_config.get_merged_inference_model_params(
-                                    inference_container_family,
-                                    inference_container_info.version,
-                                )
+                                container.spec.evaluation_configuration
                             )
-
             except Exception:
                 logger.debug(
                     f"Could not load inference config details for the evaluation source id: "
@@ -277,19 +252,12 @@ class AquaEvaluationApp(AquaApp):
             )
             evaluation_dataset_path = dst_uri
 
-        evaluation_model_parameters = None
-        try:
-            evaluation_model_parameters = AquaEvalParams(
-                shape=create_aqua_evaluation_details.shape_name,
-                dataset_path=evaluation_dataset_path,
-                report_path=create_aqua_evaluation_details.report_path,
-                **create_aqua_evaluation_details.model_parameters,
-            )
-        except Exception as ex:
-            raise AquaValueError(
-                "Invalid model parameters. Model parameters should "
-                f"be a dictionary with keys: {', '.join(list(ModelParams.__annotations__.keys()))}."
-            ) from ex
+        evaluation_model_parameters = AquaEvalParams(
+            shape=create_aqua_evaluation_details.shape_name,
+            dataset_path=evaluation_dataset_path,
+            report_path=create_aqua_evaluation_details.report_path,
+            **create_aqua_evaluation_details.model_parameters,
+        )
 
         target_compartment = (
             create_aqua_evaluation_details.compartment_id or COMPARTMENT_OCID
@@ -370,7 +338,7 @@ class AquaEvaluationApp(AquaApp):
         evaluation_model_taxonomy_metadata = ModelTaxonomyMetadata()
         evaluation_model_taxonomy_metadata[
             MetadataTaxonomyKeys.HYPERPARAMETERS
-        ].value = {"model_params": dict(asdict(evaluation_model_parameters))}
+        ].value = {"model_params": evaluation_model_parameters.to_dict()}
 
         evaluation_model = (
             DataScienceModel()
@@ -443,7 +411,6 @@ class AquaEvaluationApp(AquaApp):
                 dataset_path=evaluation_dataset_path,
                 report_path=create_aqua_evaluation_details.report_path,
                 model_parameters={
-                    **eval_inference_service_model_params,
                     **create_aqua_evaluation_details.model_parameters,
                 },
                 metrics=create_aqua_evaluation_details.metrics,
@@ -580,16 +547,14 @@ class AquaEvaluationApp(AquaApp):
                 **{
                     "AIP_SMC_EVALUATION_ARGUMENTS": json.dumps(
                         {
-                            **asdict(
-                                self._build_launch_cmd(
-                                    evaluation_id=evaluation_id,
-                                    evaluation_source_id=evaluation_source_id,
-                                    dataset_path=dataset_path,
-                                    report_path=report_path,
-                                    model_parameters=model_parameters,
-                                    metrics=metrics,
-                                ),
-                            ),
+                            **self._build_launch_cmd(
+                                evaluation_id=evaluation_id,
+                                evaluation_source_id=evaluation_source_id,
+                                dataset_path=dataset_path,
+                                report_path=report_path,
+                                model_parameters=model_parameters,
+                                metrics=metrics,
+                            ).to_dict(),
                             **(inference_configuration or {}),
                         },
                     ),
@@ -662,9 +627,9 @@ class AquaEvaluationApp(AquaApp):
                 "format": Path(dataset_path).suffix,
                 "url": dataset_path,
             },
-            metrics=metrics,
+            metrics=metrics or [],
             output_dir=report_path,
-            params=model_parameters,
+            params=model_parameters or {},
         )
 
     @telemetry(entry_point="plugin=evaluation&action=get", name="aqua")

ads/aqua/extension/deployment_handler.py

@@ -54,6 +54,33 @@ class AquaDeploymentHandler(AquaAPIhandler):
         else:
             raise HTTPError(400, f"The request {self.request.path} is invalid.")
 
+    @handle_exceptions
+    def delete(self, model_deployment_id):
+        return self.finish(AquaDeploymentApp().delete(model_deployment_id))
+
+    @handle_exceptions
+    def put(self, *args, **kwargs):
+        """
+        Handles put request for the activating and deactivating OCI datascience model deployments
+        Raises
+        ------
+        HTTPError
+            Raises HTTPError if inputs are missing or are invalid
+        """
+        url_parse = urlparse(self.request.path)
+        paths = url_parse.path.strip("/").split("/")
+        if len(paths) != 4 or paths[0] != "aqua" or paths[1] != "deployments":
+            raise HTTPError(400, f"The request {self.request.path} is invalid.")
+
+        model_deployment_id = paths[2]
+        action = paths[3]
+        if action == "activate":
+            return self.finish(AquaDeploymentApp().activate(model_deployment_id))
+        elif action == "deactivate":
+            return self.finish(AquaDeploymentApp().deactivate(model_deployment_id))
+        else:
+            raise HTTPError(400, f"The request {self.request.path} is invalid.")
+
     @handle_exceptions
     def post(self, *args, **kwargs):
         """
@@ -102,6 +129,9 @@ class AquaDeploymentHandler(AquaAPIhandler):
         ocpus = input_data.get("ocpus")
         memory_in_gbs = input_data.get("memory_in_gbs")
         model_file = input_data.get("model_file")
+        private_endpoint_id = input_data.get("private_endpoint_id")
+        container_image_uri = input_data.get("container_image_uri")
+        cmd_var = input_data.get("cmd_var")
 
         self.finish(
             AquaDeploymentApp().create(
@@ -124,6 +154,9 @@ class AquaDeploymentHandler(AquaAPIhandler):
                 ocpus=ocpus,
                 memory_in_gbs=memory_in_gbs,
                 model_file=model_file,
+                private_endpoint_id=private_endpoint_id,
+                container_image_uri=container_image_uri,
+                cmd_var=cmd_var,
             )
         )
 
@@ -264,5 +297,7 @@ __handlers__ = [
     ("deployments/?([^/]*)/params", AquaDeploymentParamsHandler),
     ("deployments/config/?([^/]*)", AquaDeploymentHandler),
     ("deployments/?([^/]*)", AquaDeploymentHandler),
+    ("deployments/?([^/]*)/activate", AquaDeploymentHandler),
+    ("deployments/?([^/]*)/deactivate", AquaDeploymentHandler),
     ("inference", AquaDeploymentInferenceHandler),
 ]

ads/aqua/extension/errors.py

@@ -8,3 +8,4 @@ class Errors(str):
     NO_INPUT_DATA = "No input data provided."
     MISSING_REQUIRED_PARAMETER = "Missing required parameter: '{}'"
     MISSING_ONEOF_REQUIRED_PARAMETER = "Either '{}' or '{}' is required."
+    INVALID_VALUE_OF_PARAMETER = "Invalid value of parameter: '{}'"

ads/aqua/extension/evaluation_handler.py

@@ -12,7 +12,6 @@ from ads.aqua.evaluation import AquaEvaluationApp
 from ads.aqua.evaluation.entities import CreateAquaEvaluationDetails
 from ads.aqua.extension.base_handler import AquaAPIhandler
 from ads.aqua.extension.errors import Errors
-from ads.aqua.extension.utils import validate_function_parameters
 from ads.config import COMPARTMENT_OCID
 
 
@@ -47,10 +46,6 @@ class AquaEvaluationHandler(AquaAPIhandler):
         if not input_data:
             raise HTTPError(400, Errors.NO_INPUT_DATA)
 
-        validate_function_parameters(
-            data_class=CreateAquaEvaluationDetails, input_data=input_data
-        )
-
         self.finish(
             # TODO: decide what other kwargs will be needed for create aqua evaluation.
             AquaEvaluationApp().create(

ads/aqua/extension/finetune_handler.py

@@ -1,5 +1,4 @@
 #!/usr/bin/env python
-# -*- coding: utf-8 -*-
 # Copyright (c) 2024 Oracle and/or its affiliates.
 # Licensed under the Universal Permissive License v 1.0 as shown at https://oss.oracle.com/licenses/upl/
 
@@ -9,8 +8,8 @@ from urllib.parse import urlparse
 from tornado.web import HTTPError
 
 from ads.aqua.common.decorator import handle_exceptions
-from ads.aqua.extension.errors import Errors
 from ads.aqua.extension.base_handler import AquaAPIhandler
+from ads.aqua.extension.errors import Errors
 from ads.aqua.extension.utils import validate_function_parameters
 from ads.aqua.finetuning import AquaFineTuningApp
 from ads.aqua.finetuning.entities import CreateFineTuningDetails

ads/aqua/extension/model_handler.py

@@ -9,7 +9,10 @@ from tornado.web import HTTPError
 
 from ads.aqua.common.decorator import handle_exceptions
 from ads.aqua.common.errors import AquaRuntimeError, AquaValueError
-from ads.aqua.common.utils import get_hf_model_info, list_hf_models
+from ads.aqua.common.utils import (
+    get_hf_model_info,
+    list_hf_models,
+)
 from ads.aqua.extension.base_handler import AquaAPIhandler
 from ads.aqua.extension.errors import Errors
 from ads.aqua.model import AquaModelApp
@@ -73,6 +76,8 @@ class AquaModelHandler(AquaAPIhandler):
         paths = url_parse.path.strip("/")
         if paths.startswith("aqua/model/cache"):
             return self.finish(AquaModelApp().clear_model_list_cache())
+        elif id:
+            return self.finish(AquaModelApp().delete_model(id))
         else:
             raise HTTPError(400, f"The request {self.request.path} is invalid.")
 
@@ -123,6 +128,7 @@ class AquaModelHandler(AquaAPIhandler):
         download_from_hf = (
             str(input_data.get("download_from_hf", "false")).lower() == "true"
         )
+        inference_container_uri = input_data.get("inference_container_uri")
 
         return self.finish(
             AquaModelApp().register(
@@ -134,9 +140,40 @@ class AquaModelHandler(AquaAPIhandler):
                 compartment_id=compartment_id,
                 project_id=project_id,
                 model_file=model_file,
+                inference_container_uri=inference_container_uri,
             )
         )
 
+    @handle_exceptions
+    def put(self, id):
+        try:
+            input_data = self.get_json_body()
+        except Exception as ex:
+            raise HTTPError(400, Errors.INVALID_INPUT_DATA_FORMAT) from ex
+
+        if not input_data:
+            raise HTTPError(400, Errors.NO_INPUT_DATA)
+
+        inference_container = input_data.get("inference_container")
+        inference_containers = AquaModelApp.list_valid_inference_containers()
+        if (
+            inference_container is not None
+            and inference_container not in inference_containers
+        ):
+            raise HTTPError(
+                400, Errors.INVALID_VALUE_OF_PARAMETER.format("inference_container")
+            )
+
+        enable_finetuning = input_data.get("enable_finetuning")
+        task = input_data.get("task")
+        app=AquaModelApp()
+        self.finish(
+            app.edit_registered_model(
+                id, inference_container, enable_finetuning, task
+            )
+        )
+        app.clear_model_details_cache(model_id=id)
+
 
 class AquaModelLicenseHandler(AquaAPIhandler):
     """Handler for Aqua Model license REST APIs."""