snowflake-ml-python 1.3.0__py3-none-any.whl → 1.4.0__py3-none-any.whl

snowflake/ml/registry/model_registry.py

@@ -103,7 +103,10 @@ def _create_registry_database(
 
 
 def _create_registry_schema(
-    session: snowpark.Session, database_name: str, schema_name: str, statement_params: Dict[str, Any]
+    session: snowpark.Session,
+    database_name: str,
+    schema_name: str,
+    statement_params: Dict[str, Any],
 ) -> None:
     """Private helper to create the model registry schema.
 
@@ -161,7 +164,11 @@ def _create_registry_views(
 
     # Create a view on active permanent deployments.
     _create_active_permanent_deployment_view(
-        session, fully_qualified_schema_name, registry_table_name, deployment_table_name, statement_params
+        session,
+        fully_qualified_schema_name,
+        registry_table_name,
+        deployment_table_name,
+        statement_params,
     )
 
     # Create views on most recent metadata items.
@@ -437,10 +444,12 @@ fully integrated into the new registry.
         # Could do a multi-table insert here with some pros and cons:
         # [PRO] Atomic insert across multiple tables.
         # [CON] Code logic becomes messy depending on which fields are set.
-        # [CON] Harder to re-use existing methods like set_model_name.
+        # [CON] Harder to reuse existing methods like set_model_name.
         # Context: https://docs.snowflake.com/en/sql-reference/sql/insert-multi-table.html
         return table_manager.insert_table_entry(
-            self._session, table=self._fully_qualified_registry_table_name(), columns=properties
+            self._session,
+            table=self._fully_qualified_registry_table_name(),
+            columns=properties,
         )
 
     def _insert_metadata_entry(self, *, id: str, attribute: str, value: Any, operation: str) -> List[snowpark.Row]:
@@ -471,7 +480,9 @@ fully integrated into the new registry.
         columns["VALUE"] = value
 
         return table_manager.insert_table_entry(
-            self._session, table=self._fully_qualified_metadata_table_name(), columns=columns
+            self._session,
+            table=self._fully_qualified_metadata_table_name(),
+            columns=columns,
         )
 
     def _insert_deployment_entry(
@@ -484,7 +495,10 @@ fully integrated into the new registry.
         signature: Dict[str, Any],
         target_method: str,
         options: Optional[
-            Union[model_types.WarehouseDeployOptions, model_types.SnowparkContainerServiceDeployOptions]
+            Union[
+                model_types.WarehouseDeployOptions,
+                model_types.SnowparkContainerServiceDeployOptions,
+            ]
         ] = None,
     ) -> List[snowpark.Row]:
         """Insert a new row into the model deployment table.
@@ -521,7 +535,9 @@ fully integrated into the new registry.
         columns["OPTIONS"] = options
 
         return table_manager.insert_table_entry(
-            self._session, table=self._fully_qualified_deployment_table_name(), columns=columns
+            self._session,
+            table=self._fully_qualified_deployment_table_name(),
+            columns=columns,
         )
 
     def _prepare_deployment_stage(self) -> str:
@@ -596,7 +612,11 @@ fully integrated into the new registry.
         return identifier.get_schema_level_object_identifier(db, schema, stage)
 
     def _list_selected_models(
-        self, *, id: Optional[str] = None, model_name: Optional[str] = None, model_version: Optional[str] = None
+        self,
+        *,
+        id: Optional[str] = None,
+        model_name: Optional[str] = None,
+        model_version: Optional[str] = None,
     ) -> snowpark.DataFrame:
         """Retrieve the Snowpark dataframe of models matching the specified ID or (name and version).
 
@@ -724,7 +744,12 @@ fully integrated into the new registry.
         assert id is not None
 
         try:
-            self._insert_metadata_entry(id=id, attribute=attribute, value={attribute: value}, operation=operation)
+            self._insert_metadata_entry(
+                id=id,
+                attribute=attribute,
+                value={attribute: value},
+                operation=operation,
+            )
         except connector.DataError:
             raise connector.DataError(f"Setting {attribute} for mode id {id} failed.")
 
@@ -760,7 +785,10 @@ fully integrated into the new registry.
         return str(result)
 
     def _get_model_path(
-        self, id: Optional[str] = None, model_name: Optional[str] = None, model_version: Optional[str] = None
+        self,
+        id: Optional[str] = None,
+        model_name: Optional[str] = None,
+        model_version: Optional[str] = None,
     ) -> str:
         """Get the stage path for the model with the given (model name + model version) or `id` from the registry.
 
@@ -889,10 +917,17 @@ fully integrated into the new registry.
                 value=new_model,
             )
             if description:
-                self.set_model_description(model_name=model_name, model_version=model_version, description=description)
+                self.set_model_description(
+                    model_name=model_name,
+                    model_version=model_version,
+                    description=description,
+                )
             if tags:
                 self._set_metadata_attribute(
-                    _METADATA_ATTRIBUTE_TAGS, value=tags, model_name=model_name, model_version=model_version
+                    _METADATA_ATTRIBUTE_TAGS,
+                    value=tags,
+                    model_name=model_name,
+                    model_version=model_version,
                 )
         else:
             raise connector.DatabaseError("Failed to insert the model properties to the registry table.")
@@ -961,7 +996,10 @@ fully integrated into the new registry.
         model_tags = self.get_tags(model_name=model_name, model_version=model_version)
         model_tags[tag_name] = tag_value
         self._set_metadata_attribute(
-            _METADATA_ATTRIBUTE_TAGS, model_tags, model_name=model_name, model_version=model_version
+            _METADATA_ATTRIBUTE_TAGS,
+            model_tags,
+            model_name=model_name,
+            model_version=model_version,
         )
 
     @telemetry.send_api_usage_telemetry(
@@ -991,7 +1029,10 @@ fully integrated into the new registry.
             )
 
         self._set_metadata_attribute(
-            _METADATA_ATTRIBUTE_TAGS, model_tags, model_name=model_name, model_version=model_version
+            _METADATA_ATTRIBUTE_TAGS,
+            model_tags,
+            model_name=model_name,
+            model_version=model_version,
         )
 
     @telemetry.send_api_usage_telemetry(
@@ -1089,7 +1130,9 @@ fully integrated into the new registry.
             Description of the model or None.
         """
         result = self._get_metadata_attribute(
-            _METADATA_ATTRIBUTE_DESCRIPTION, model_name=model_name, model_version=model_version
+            _METADATA_ATTRIBUTE_DESCRIPTION,
+            model_name=model_name,
+            model_version=model_version,
         )
         return None if result is None else json.loads(result)
 
@@ -1112,7 +1155,10 @@ fully integrated into the new registry.
             description: Desired new model description.
         """
         self._set_metadata_attribute(
-            _METADATA_ATTRIBUTE_DESCRIPTION, description, model_name=model_name, model_version=model_version
+            _METADATA_ATTRIBUTE_DESCRIPTION,
+            description,
+            model_name=model_name,
+            model_version=model_version,
         )
 
     @telemetry.send_api_usage_telemetry(
@@ -1165,7 +1211,10 @@ fully integrated into the new registry.
             snowpark.DataFrame with the history of the model.
         """
         id = self._get_model_id(model_name=model_name, model_version=model_version)
-        return cast(snowpark.DataFrame, self.get_history().filter(snowpark.Column("MODEL_ID") == id))
+        return cast(
+            snowpark.DataFrame,
+            self.get_history().filter(snowpark.Column("MODEL_ID") == id),
+        )
 
     @telemetry.send_api_usage_telemetry(
         project=_TELEMETRY_PROJECT,
@@ -1194,7 +1243,10 @@ fully integrated into the new registry.
         model_metrics = self.get_metrics(model_name=model_name, model_version=model_version)
         model_metrics[metric_name] = metric_value
         self._set_metadata_attribute(
-            _METADATA_ATTRIBUTE_METRICS, model_metrics, model_name=model_name, model_version=model_version
+            _METADATA_ATTRIBUTE_METRICS,
+            model_metrics,
+            model_name=model_name,
+            model_version=model_version,
         )
 
     @telemetry.send_api_usage_telemetry(
@@ -1230,7 +1282,10 @@ fully integrated into the new registry.
             )
 
         self._set_metadata_attribute(
-            _METADATA_ATTRIBUTE_METRICS, model_metrics, model_name=model_name, model_version=model_version
+            _METADATA_ATTRIBUTE_METRICS,
+            model_metrics,
+            model_name=model_name,
+            model_version=model_version,
         )
 
     @telemetry.send_api_usage_telemetry(
@@ -1290,7 +1345,9 @@ fully integrated into the new registry.
         # Snowpark snowpark.dataframe returns dictionary objects as strings. We need to convert it back to a dictionary
         # here.
         result = self._get_metadata_attribute(
-            _METADATA_ATTRIBUTE_METRICS, model_name=model_name, model_version=model_version
+            _METADATA_ATTRIBUTE_METRICS,
+            model_name=model_name,
+            model_version=model_version,
         )
 
         if result:
@@ -1432,7 +1489,7 @@ fully integrated into the new registry.
                 metadata=tags,
                 conda_dependencies=conda_dependencies,
                 pip_requirements=pip_requirements,
-                sample_input=sample_input_data,
+                sample_input_data=sample_input_data,
                 code_paths=code_paths,
                 options=options,
             )
@@ -1507,7 +1564,10 @@ fully integrated into the new registry.
         permanent: bool = False,
         platform: deploy_platforms.TargetPlatform = deploy_platforms.TargetPlatform.WAREHOUSE,
         options: Optional[
-            Union[model_types.WarehouseDeployOptions, model_types.SnowparkContainerServiceDeployOptions]
+            Union[
+                model_types.WarehouseDeployOptions,
+                model_types.SnowparkContainerServiceDeployOptions,
+            ]
         ] = None,
     ) -> model_types.Deployment:
         """Deploy the model with the given deployment name.
@@ -1772,7 +1832,9 @@ fully integrated into the new registry.
 
         """
         deployment = self._get_deployment(
-            model_name=model_name, model_version=model_version, deployment_name=deployment_name
+            model_name=model_name,
+            model_version=model_version,
+            deployment_name=deployment_name,
         )
 
         # TODO(SNOW-759526): The following sequence should be a transaction.
@@ -1845,7 +1907,8 @@ fully integrated into the new registry.
 
         # Step 1/3: Delete the registry entry.
         query_result_checker.SqlResultValidator(
-            self._session, f"DELETE FROM {self._fully_qualified_registry_table_name()} WHERE ID='{id}'"
+            self._session,
+            f"DELETE FROM {self._fully_qualified_registry_table_name()} WHERE ID='{id}'",
         ).deletion_success(expected_num_rows=1).validate()
 
         # Step 2/3: Delete the artifact (if desired).
@@ -1966,7 +2029,11 @@ class ModelReference:
 
             def build_method(m: Callable[..., Any]) -> Callable[..., Any]:
                 return lambda self, *args, **kwargs: m(
-                    self._registry, self._model_name, self._model_version, *args, **kwargs
+                    self._registry,
+                    self._model_name,
+                    self._model_version,
+                    *args,
+                    **kwargs,
                 )
 
             method = build_method(m=obj)
@@ -2027,7 +2094,10 @@ class ModelReference:
 
         if di:
             return model_api.predict(
-                session=self._registry._session, deployment=di, X=data, statement_params=statement_params
+                session=self._registry._session,
+                deployment=di,
+                X=data,
+                statement_params=statement_params,
             )
 
         # Mypy enforce to refer to the registry for calling the function
@@ -2059,7 +2129,10 @@ class ModelReference:
             options=options,
         )
         return model_api.predict(
-            session=self._registry._session, deployment=di, X=data, statement_params=statement_params
+            session=self._registry._session,
+            deployment=di,
+            X=data,
+            statement_params=statement_params,
         )
 
 

snowflake/ml/registry/registry.py

@@ -77,7 +77,7 @@ class Registry:
         model: model_types.SupportedModelType,
         *,
         model_name: str,
-        version_name: str,
+        version_name: Optional[str] = None,
         comment: Optional[str] = None,
         metrics: Optional[Dict[str, Any]] = None,
         conda_dependencies: Optional[List[str]] = None,
@@ -95,9 +95,10 @@ class Registry:
         Args:
             model: Model object of supported types such as Scikit-learn, XGBoost, Snowpark ML,
                 PyTorch, TorchScript, Tensorflow, Tensorflow Keras, MLFlow, HuggingFace Pipeline,
-                Peft-finetuned LLM, or Custom Model.
+                Sentence Transformers, Peft-finetuned LLM, or Custom Model.
             model_name: Name to identify the model.
             version_name: Version identifier for the model. Combination of model_name and version_name must be unique.
+                If not specified, a random name will be generated.
             comment: Comment associated with the model version. Defaults to None.
             metrics: A JSON serializable dictionary containing metrics linked to the model version. Defaults to None.
             signatures: Model data signatures for inputs and outputs for various target methods. If it is None,
@@ -124,6 +125,8 @@ class Registry:
                     Channel. Otherwise, defaults to False
                 - relax_version: Whether or not relax the version constraints of the dependencies.
                     It detects any ==x.y.z in specifiers and replaced with >=x.y, <(x+1). Defaults to True.
+                - function_type: Set the method function type globally. To set method function types individually see
+                  function_type in model_options.
                 - method_options: Per-method saving options including:
                     - case_sensitive: Indicates whether the method and its signature should be case sensitive.
                         This means when you refer the method in the SQL, you need to double quote it.
@@ -131,6 +134,7 @@ class Registry:
                         non-alphabetic characters in your method or feature name. Defaults to False.
                     - max_batch_size: Maximum batch size that the method could accept in the Snowflake Warehouse.
                         Defaults to None, determined automatically by Snowflake.
+                    - function_type: One of supported model method function types (FUNCTION or TABLE_FUNCTION).
 
         Raises:
             NotImplementedError: `pip_requirements` is not supported.

snowflake/ml/version.py

@@ -1 +1 @@
-VERSION="1.3.0"
+VERSION="1.4.0"

{snowflake_ml_python-1.3.0.dist-info → snowflake_ml_python-1.4.0.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: snowflake-ml-python
-Version: 1.3.0
+Version: 1.4.0
 Summary: The machine learning client library that is used for interacting with Snowflake to build machine learning solutions.
 Author-email: "Snowflake, Inc" <support@snowflake.com>
 License: 
@@ -238,7 +238,7 @@ Requires-Dist: anyio <4,>=3.5.0
 Requires-Dist: cachetools <6,>=3.1.1
 Requires-Dist: cloudpickle >=2.0.0
 Requires-Dist: fsspec[http] <2024,>=2022.11
-Requires-Dist: importlib-resources <6,>=5.1.4
+Requires-Dist: importlib-resources <7,>=6.1.1
 Requires-Dist: numpy <2,>=1.23
 Requires-Dist: packaging <24,>=20.9
 Requires-Dist: pandas <2,>=1.0.0
@@ -249,8 +249,8 @@ Requires-Dist: retrying <2,>=1.3.3
 Requires-Dist: s3fs <2024,>=2022.11
 Requires-Dist: scikit-learn <1.4,>=1.2.1
 Requires-Dist: scipy <2,>=1.9
-Requires-Dist: snowflake-connector-python[pandas] <4,>=3.0.4
-Requires-Dist: snowflake-snowpark-python !=1.12.0,<2,>=1.8.0
+Requires-Dist: snowflake-connector-python[pandas] <4,>=3.5.0
+Requires-Dist: snowflake-snowpark-python !=1.12.0,<2,>=1.11.1
 Requires-Dist: sqlparse <1,>=0.4
 Requires-Dist: typing-extensions <5,>=4.1.0
 Requires-Dist: xgboost <2,>=1.7.3
@@ -258,9 +258,10 @@ Provides-Extra: all
 Requires-Dist: lightgbm ==3.3.5 ; extra == 'all'
 Requires-Dist: mlflow <2.4,>=2.1.0 ; extra == 'all'
 Requires-Dist: peft <1,>=0.5.0 ; extra == 'all'
+Requires-Dist: sentence-transformers <3,>=2.2.2 ; extra == 'all'
 Requires-Dist: sentencepiece <0.2,>=0.1.95 ; extra == 'all'
 Requires-Dist: shap ==0.42.1 ; extra == 'all'
-Requires-Dist: tensorflow !=2.12.0,<3,>=2.9 ; extra == 'all'
+Requires-Dist: tensorflow <3,>=2.10 ; extra == 'all'
 Requires-Dist: tokenizers <1,>=0.10 ; extra == 'all'
 Requires-Dist: torchdata <1,>=0.4 ; extra == 'all'
 Requires-Dist: transformers <5,>=4.32.1 ; extra == 'all'
@@ -274,11 +275,12 @@ Requires-Dist: mlflow <2.4,>=2.1.0 ; extra == 'mlflow'
 Provides-Extra: shap
 Requires-Dist: shap ==0.42.1 ; extra == 'shap'
 Provides-Extra: tensorflow
-Requires-Dist: tensorflow !=2.12.0,<3,>=2.9 ; extra == 'tensorflow'
+Requires-Dist: tensorflow <3,>=2.10 ; extra == 'tensorflow'
 Provides-Extra: torch
 Requires-Dist: torchdata <1,>=0.4 ; extra == 'torch'
 Requires-Dist: torch <2.1.0,>=2.0.1 ; (platform_system == "Windows") and extra == 'torch'
 Provides-Extra: transformers
+Requires-Dist: sentence-transformers <3,>=2.2.2 ; extra == 'transformers'
 Requires-Dist: sentencepiece <0.2,>=0.1.95 ; extra == 'transformers'
 Requires-Dist: tokenizers <1,>=0.10 ; extra == 'transformers'
 Requires-Dist: transformers <5,>=4.32.1 ; extra == 'transformers'
@@ -368,7 +370,41 @@ be compatibility issues. Server-side functionality that `snowflake-ml-python` de
 
 # Release History
 
-## 1.3.0
+## 1.4.0
+
+### Bug Fixes
+
+- Registry: Fix a bug when multiple models are being called from the same query, models other than the first one will
+  have incorrect result. This fix only works for newly logged model.
+- Modeling: When registering a model, only method(s) that is mentioned in `save_model` would be added to model signature
+in SnowML models.
+- Modeling: Fix a bug that when n_jobs is not 1, model cannot execute methods such as
+predict, predict_log_proba, and other batch inference methods. The n_jobs would automatically
+set to 1 because vectorized udf currently doesn't support joblib parallel backend.
+- Modeling: Fix a bug that batch inference methods cannot infer the datatype when the first row of data contains NULL.
+- Modeling: Matches Distributed HPO output column names with the snowflake identifier.
+- Modeling: Relax package versions for all Distributed HPO methods if the installed version
+  is not available in the Snowflake conda channel
+- Modeling: Add sklearn as required dependency for LightGBM package.
+
+### Behavior Changes
+
+- Registry: `apply` method is no longer by default logged when logging a xgboost model. If that is required, it could
+  be specified manually when logging the model by `log_model(..., options={"target_methods": ["apply", ...]})`.
+
+### New Features
+
+- Registry: Add support for `sentence-transformers` model (`sentence_transformers.SentenceTransformer`).
+- Registry: Now version name is no longer required when logging a model. If not provided, a random human readable ID
+  will be generated.
+
+## 1.3.1 (2024-03-21)
+
+### New Features
+
+- FileSet: `snowflake.ml.fileset.sfcfs.SFFileSystem` can now be used in UDFs and stored procedures.
+
+## 1.3.0 (2024-03-12)
 
 ### Bug Fixes