Merge branch 'main' of https://github.com/oracle/accelerated-data-science into ODSC-50632/improve_progress_bar

Author: lu-ohai

ads/catalog/model.py

@@ -82,11 +82,11 @@
 _WORK_REQUEST_INTERVAL_IN_SEC = 3
 
 
-class ModelWithActiveDeploymentError(Exception):   # pragma: no cover
+class ModelWithActiveDeploymentError(Exception):  # pragma: no cover
     pass
 
 
-class ModelArtifactSizeError(Exception):   # pragma: no cover
+class ModelArtifactSizeError(Exception):  # pragma: no cover
     def __init__(self, max_artifact_size: str):
         super().__init__(
             f"The model artifacts size is greater than `{max_artifact_size}`. "
@@ -391,7 +391,6 @@ def show_in_notebook(self, display_format: str = "dataframe") -> None:
             Nothing.
         """
         if display_format == "dataframe":
-
             from IPython.core.display import display
 
             display(self.to_dataframe())
@@ -1560,9 +1559,9 @@ def _wait_for_work_request(
                 if work_request_logs:
                     new_work_request_logs = work_request_logs[i:]
 
-                for wr_item in new_work_request_logs:
-                    progress.update(wr_item.message)
-                    i += 1
+                    for wr_item in new_work_request_logs:
+                        progress.update(wr_item.message)
+                        i += 1
 
                 if work_request.data.status in STOP_STATE:
                     if work_request.data.status != WorkRequest.STATUS_SUCCEEDED:

ads/common/dsc_file_system.py

@@ -265,8 +265,6 @@ def update_from_dsc_model(cls, dsc_model) -> dict:
 
 class DSCFileSystemManager:
 
-    storage_mount_dest = set()
-
     @classmethod
     def initialize(cls, arguments: dict) -> dict:
         """Initialize and update arguments to dsc model.
@@ -286,12 +284,6 @@ def initialize(cls, arguments: dict) -> dict:
                 "Parameter `dest` is required for mounting file storage system."
             )
 
-        if arguments["dest"] in cls.storage_mount_dest:
-            raise ValueError(
-                "Duplicate `dest` found. Please specify different `dest` for each file system to be mounted."
-            )
-        cls.storage_mount_dest.add(arguments["dest"])
-
         # case oci://bucket@namespace/prefix
         if arguments["src"].startswith("oci://") and "@" in arguments["src"]:
             return OCIObjectStorage(**arguments).update_to_dsc_model()

ads/config.py

@@ -45,6 +45,7 @@
     NB_SESSION_OCID or JOB_RUN_OCID or MD_OCID or PIPELINE_RUN_OCID or DATAFLOW_RUN_OCID
 )
 NO_CONTAINER = os.environ.get("NO_CONTAINER")
+TMPDIR = os.environ.get("TMPDIR")
 
 
 def export(

ads/model/artifact.py

@@ -30,21 +30,21 @@
 ADS_VERSION = __version__
 
 
-class ArtifactNestedFolderError(Exception):   # pragma: no cover
+class ArtifactNestedFolderError(Exception):  # pragma: no cover
     def __init__(self, folder: str):
         self.folder = folder
         super().__init__("The required artifact files placed in a nested folder.")
 
 
-class ArtifactRequiredFilesError(Exception):   # pragma: no cover
+class ArtifactRequiredFilesError(Exception):  # pragma: no cover
     def __init__(self, required_files: Tuple[str]):
         super().__init__(
             "Not all required files presented in artifact folder. "
             f"Required files for conda runtime: {required_files}. If you are using container runtime, set `ignore_conda_error=True`."
         )
 
 
-class AritfactFolderStructureError(Exception):   # pragma: no cover
+class AritfactFolderStructureError(Exception):  # pragma: no cover
     def __init__(self, required_files: Tuple[str]):
         super().__init__(
             "The artifact folder has a wrong structure. "
@@ -171,6 +171,7 @@ def __init__(
         self.ignore_conda_error = ignore_conda_error
         self.model = None
         self.auth = auth or authutil.default_signer()
+
         if reload and not ignore_conda_error:
             self.reload()
             # Extracts the model_file_name from the score.py.
@@ -272,8 +273,9 @@ def prepare_runtime_yaml(
             or runtime_info.model_deployment.inference_conda_env.inference_python_version.strip()
             == ""
         ):
-            warnings.warn(
-                "Cannot automatically detect the inference python version. `inference_python_version` must be provided."
+            raise ValueError(
+                "Cannot automatically detect the inference python version. "
+                "`inference_python_version` must be provided."
             )
         runtime_file_path = os.path.join(self.artifact_dir, "runtime.yaml")
         if os.path.exists(runtime_file_path) and not force_overwrite:
@@ -416,6 +418,7 @@ def from_uri(
         force_overwrite: Optional[bool] = False,
         auth: Optional[Dict] = None,
         ignore_conda_error: Optional[bool] = False,
+        reload: Optional[bool] = False,
     ):
         """Constructs a ModelArtifact object from the existing model artifacts.
 
@@ -426,16 +429,20 @@ def from_uri(
             OCI object storage URI.
         artifact_dir: str
             The local artifact folder to store the files needed for deployment.
-        model_file_name: (str, optional). Defaults to `None`
-            The file name of the serialized model.
-        force_overwrite: (bool, optional). Defaults to False.
-            Whether to overwrite existing files or not.
         auth: (Dict, optional). Defaults to None.
             The default authetication is set using `ads.set_auth` API.
             If you need to override the default, use the `ads.common.auth.api_keys`
             or `ads.common.auth.resource_principal` to create appropriate
             authentication signer and kwargs required to instantiate
             IdentityClient object.
+        force_overwrite: (bool, optional). Defaults to False.
+            Whether to overwrite existing files or not.
+        ignore_conda_error: (bool, optional). Defaults to False.
+            Parameter to ignore error when collecting conda information.
+        model_file_name: (str, optional). Defaults to `None`
+            The file name of the serialized model.
+        reload: (bool, optional). Defaults to False.
+            Whether to reload the Model into the environment.
 
         Returns
         -------
@@ -492,6 +499,8 @@ def from_uri(
                     utils.copy_from_uri(
                         uri=temp_dir, to_path=to_path, force_overwrite=True
                     )
+            except ArtifactRequiredFilesError as ex:
+                logger.warning(ex)
 
         if ObjectStorageDetails.is_oci_path(artifact_dir):
             for root, dirs, files in os.walk(to_path):
@@ -507,10 +516,10 @@ def from_uri(
 
         return cls(
             artifact_dir=artifact_dir,
-            model_file_name=model_file_name,
-            reload=True,
             ignore_conda_error=ignore_conda_error,
             local_copy_dir=to_path,
+            model_file_name=model_file_name,
+            reload=reload,
         )
 
     def __getattr__(self, item):

ads/model/artifact_downloader.py

@@ -63,6 +63,7 @@ def download(self):
                     "Set `force_overwrite` to `True` if you wish to overwrite."
                 )
             shutil.rmtree(self.target_dir)
+            os.mkdir(self.target_dir)
         with utils.get_progress_bar(
             ArtifactDownloader.PROGRESS_STEPS_COUNT + self.PROGRESS_STEPS_COUNT
         ) as progress:
@@ -85,15 +86,16 @@ def _download(self):
         """Downloads model artifacts."""
         self.progress.update("Importing model artifacts from catalog")
         zip_content = self.dsc_model.get_model_artifact_content()
-        with tempfile.TemporaryDirectory() as temp_dir:
-            self.progress.update("Copying model artifacts to the artifact directory")
-            zip_file_path = os.path.join(temp_dir, f"{str(uuid.uuid4())}.zip")
-            with open(zip_file_path, "wb") as zip_file:
-                zip_file.write(zip_content)
-            self.progress.update("Extracting model artifacts")
-            with ZipFile(zip_file_path) as zip_file:
-                zip_file.extractall(self.target_dir)
-
+        self.progress.update("Copying model artifacts to the artifact directory")
+        
+        zip_file_path = os.path.join(self.target_dir, f"{str(uuid.uuid4())}.zip")
+        with open(zip_file_path, "wb") as zip_file:
+            zip_file.write(zip_content)
+        self.progress.update("Extracting model artifacts")
+        with ZipFile(zip_file_path) as zip_file:
+            zip_file.extractall(self.target_dir)
+        
+        utils.remove_file(zip_file_path)
 
 class LargeArtifactDownloader(ArtifactDownloader):
     PROGRESS_STEPS_COUNT = 4
@@ -157,17 +159,19 @@ def _download(self):
 
         self.dsc_model.import_model_artifact(bucket_uri=bucket_uri, region=self.region)
         self.progress.update("Copying model artifacts to the artifact directory")
-        with tempfile.TemporaryDirectory() as temp_dir:
-            zip_file_path = os.path.join(temp_dir, f"{str(uuid.uuid4())}.zip")
-            zip_file_path = utils.copy_file(
-                uri_src=bucket_uri,
-                uri_dst=zip_file_path,
-                auth=self.auth,
-                progressbar_description="Copying model artifacts to the artifact directory",
-            )
-            self.progress.update("Extracting model artifacts")
-            with ZipFile(zip_file_path) as zip_file:
-                zip_file.extractall(self.target_dir)
+        
+        zip_file_path = os.path.join(self.target_dir, f"{str(uuid.uuid4())}.zip")
+        zip_file_path = utils.copy_file(
+            uri_src=bucket_uri,
+            uri_dst=zip_file_path,
+            auth=self.auth,
+            progressbar_description="Copying model artifacts to the artifact directory",
+        )
+        self.progress.update("Extracting model artifacts")
+        with ZipFile(zip_file_path) as zip_file:
+            zip_file.extractall(self.target_dir)
+
+        utils.remove_file(zip_file_path)
 
         if self.remove_existing_artifact:
             self.progress.update(

ads/model/deployment/model_deployment_infrastructure.py

@@ -20,7 +20,6 @@
 MODEL_DEPLOYMENT_INFRASTRUCTURE_KIND = "infrastructure"
 
 DEFAULT_BANDWIDTH_MBPS = 10
-DEFAULT_WEB_CONCURRENCY = 10
 DEFAULT_REPLICA = 1
 DEFAULT_SHAPE_NAME = "VM.Standard.E4.Flex"
 DEFAULT_OCPUS = 1
@@ -219,7 +218,6 @@ def _load_default_properties(self) -> Dict:
             defaults[self.CONST_PROJECT_ID] = PROJECT_OCID
 
         defaults[self.CONST_BANDWIDTH_MBPS] = DEFAULT_BANDWIDTH_MBPS
-        defaults[self.CONST_WEB_CONCURRENCY] = DEFAULT_WEB_CONCURRENCY
         defaults[self.CONST_REPLICA] = DEFAULT_REPLICA
 
         if NB_SESSION_OCID:
@@ -628,7 +626,6 @@ def init(self, **kwargs) -> "ModelDeploymentInfrastructure":
             .with_compartment_id(self.compartment_id or "{Provide a compartment OCID}")
             .with_project_id(self.project_id or "{Provide a project OCID}")
             .with_bandwidth_mbps(self.bandwidth_mbps or DEFAULT_BANDWIDTH_MBPS)
-            .with_web_concurrency(self.web_concurrency or DEFAULT_WEB_CONCURRENCY)
             .with_replica(self.replica or DEFAULT_REPLICA)
             .with_shape_name(self.shape_name or DEFAULT_SHAPE_NAME)
             .with_shape_config_details(

ads/model/generic_model.py

@@ -31,6 +31,7 @@
     NB_SESSION_OCID,
     PIPELINE_RUN_COMPARTMENT_OCID,
     PROJECT_OCID,
+    TMPDIR,
 )
 from ads.evaluations import EvaluatorMixin
 from ads.feature_engineering import ADSImage
@@ -183,7 +184,7 @@ def _prepare_artifact_dir(artifact_dir: str = None) -> str:
     if artifact_dir and isinstance(artifact_dir, str):
         return os.path.abspath(os.path.expanduser(artifact_dir))
 
-    artifact_dir = tempfile.mkdtemp()
+    artifact_dir = TMPDIR or tempfile.mkdtemp()
     logger.info(
         f"The `artifact_dir` was not provided and "
         f"automatically set to: {artifact_dir}"
@@ -965,17 +966,20 @@ def prepare(
             auth=self.auth,
             local_copy_dir=self.local_copy_dir,
         )
-        self.runtime_info = self.model_artifact.prepare_runtime_yaml(
-            inference_conda_env=self.properties.inference_conda_env,
-            inference_python_version=self.properties.inference_python_version,
-            training_conda_env=self.properties.training_conda_env,
-            training_python_version=self.properties.training_python_version,
-            force_overwrite=force_overwrite,
-            namespace=namespace,
-            bucketname=DEFAULT_CONDA_BUCKET_NAME,
-            auth=self.auth,
-            ignore_conda_error=self.ignore_conda_error,
-        )
+        try:
+            self.runtime_info = self.model_artifact.prepare_runtime_yaml(
+                inference_conda_env=self.properties.inference_conda_env,
+                inference_python_version=self.properties.inference_python_version,
+                training_conda_env=self.properties.training_conda_env,
+                training_python_version=self.properties.training_python_version,
+                force_overwrite=force_overwrite,
+                namespace=namespace,
+                bucketname=DEFAULT_CONDA_BUCKET_NAME,
+                auth=self.auth,
+                ignore_conda_error=self.ignore_conda_error,
+            )
+        except ValueError as e:
+            raise e
 
         self._summary_status.update_status(
             detail="Generated runtime.yaml", status=ModelState.DONE.value
@@ -1348,13 +1352,15 @@ def from_model_artifact(
         properties.with_dict(local_vars)
         auth = auth or authutil.default_signer()
         artifact_dir = _prepare_artifact_dir(artifact_dir)
+        reload = kwargs.pop("reload", False)
         model_artifact = ModelArtifact.from_uri(
             uri=uri,
             artifact_dir=artifact_dir,
-            model_file_name=model_file_name,
-            force_overwrite=force_overwrite,
             auth=auth,
+            force_overwrite=force_overwrite,
             ignore_conda_error=ignore_conda_error,
+            model_file_name=model_file_name,
+            reload=reload,
         )
         model = cls(
             estimator=model_artifact.model,
@@ -1367,22 +1373,33 @@ def from_model_artifact(
         model.local_copy_dir = model_artifact.local_copy_dir
         model.model_artifact = model_artifact
         model.ignore_conda_error = ignore_conda_error
-        model.reload_runtime_info()
+
+        if reload:
+            model.reload_runtime_info()
+            model._summary_status.update_action(
+                detail="Populated metadata(Custom, Taxonomy and Provenance)",
+                action="Call .populate_metadata() to populate metadata.",
+            )
+
         model._summary_status.update_status(
             detail="Generated score.py",
-            status=ModelState.DONE.value,
+            status=ModelState.NOTAPPLICABLE.value,
         )
         model._summary_status.update_status(
             detail="Generated runtime.yaml",
-            status=ModelState.DONE.value,
+            status=ModelState.NOTAPPLICABLE.value,
         )
         model._summary_status.update_status(
-            detail="Serialized model", status=ModelState.DONE.value
+            detail="Serialized model",
+            status=ModelState.NOTAPPLICABLE.value,
         )
-        model._summary_status.update_action(
+        model._summary_status.update_status(
             detail="Populated metadata(Custom, Taxonomy and Provenance)",
-            action=f"Call .populate_metadata() to populate metadata.",
+            status=ModelState.AVAILABLE.value
+            if reload
+            else ModelState.NOTAPPLICABLE.value,
         )
+
         return model
 
     @classmethod
@@ -3033,6 +3050,7 @@ class ModelState(Enum):
     AVAILABLE = "Available"
     NOTAVAILABLE = "Not Available"
     NEEDSACTION = "Needs Action"
+    NOTAPPLICABLE = "Not Applicable"
 
 
 class SummaryStatus: