snowflake-ml-python 1.1.0__py3-none-any.whl → 1.1.2__py3-none-any.whl

snowflake/ml/modeling/feature_selection/variance_threshold.py

@@ -22,17 +22,19 @@ from sklearn.utils.metaestimators import available_if
 from snowflake.ml.modeling.framework.base import BaseTransformer, _process_cols
 from snowflake.ml._internal import telemetry
 from snowflake.ml._internal.exceptions import error_codes, exceptions, modeling_error_messages
+from snowflake.ml._internal.env_utils import SNOWML_SPROC_ENV
 from snowflake.ml._internal.utils import pkg_version_utils, identifier
-from snowflake.snowpark import DataFrame
+from snowflake.snowpark import DataFrame, Session
 from snowflake.snowpark._internal.type_utils import convert_sp_to_sf_type
 from snowflake.ml.modeling._internal.snowpark_handlers import SnowparkHandlers as HandlersImpl
+from snowflake.ml.modeling._internal.model_trainer_builder import ModelTrainerBuilder
+from snowflake.ml.modeling._internal.model_trainer import ModelTrainer
 from snowflake.ml.modeling._internal.estimator_utils import (
     gather_dependencies,
     original_estimator_has_callable,
     transform_snowml_obj_to_sklearn_obj,
     validate_sklearn_args,
 )
-from snowflake.ml.modeling._internal.snowpark_handlers import SklearnWrapperProvider
 from snowflake.ml.modeling._internal.estimator_protocols import FitPredictHandlers
 
 from snowflake.ml.model.model_signature import (
@@ -52,7 +54,6 @@ _PROJECT = "ModelDevelopment"
 _SUBPROJECT = "".join([s.capitalize() for s in "sklearn.feature_selection".replace("sklearn.", "").split("_")])
 
 
-
 class VarianceThreshold(BaseTransformer):
     r"""Feature selector that removes all low-variance features
     For more details on this class, see [sklearn.feature_selection.VarianceThreshold]
@@ -60,39 +61,53 @@ class VarianceThreshold(BaseTransformer):
 
     Parameters
     ----------
-    threshold: float, default=0
-        Features with a training-set variance lower than this threshold will
-        be removed. The default is to keep all features with non-zero variance,
-        i.e. remove the features that have the same value in all samples.
 
     input_cols: Optional[Union[str, List[str]]]
         A string or list of strings representing column names that contain features.
         If this parameter is not specified, all columns in the input DataFrame except
-        the columns specified by label_cols and sample_weight_col parameters are
-        considered input columns.
-
+        the columns specified by label_cols, sample_weight_col, and passthrough_cols
+        parameters are considered input columns. Input columns can also be set after
+        initialization with the `set_input_cols` method.
+    
     label_cols: Optional[Union[str, List[str]]]
-        A string or list of strings representing column names that contain labels.
-        This is a required param for estimators, as there is no way to infer these
-        columns. If this parameter is not specified, then object is fitted without
-        labels (like a transformer).
-
+        This parameter is optional and will be ignored during fit. It is present here for API consistency by convention.
+                    
     output_cols: Optional[Union[str, List[str]]]
         A string or list of strings representing column names that will store the
         output of predict and transform operations. The length of output_cols must
-        match the expected number of output columns from the specific estimator or
+        match the expected number of output columns from the specific predictor or
         transformer class used.
-        If this parameter is not specified, output column names are derived by
-        adding an OUTPUT_ prefix to the label column names. These inferred output
-        column names work for estimator's predict() method, but output_cols must
-        be set explicitly for transformers.
+        If you omit this parameter, output column names are derived by adding an
+        OUTPUT_ prefix to the label column names for supervised estimators, or
+        OUTPUT_<IDX>for unsupervised estimators. These inferred output column names
+        work for predictors, but output_cols must be set explicitly for transformers.
+        In general, explicitly specifying output column names is clearer, especially
+        if you don’t specify the input column names.
+        To transform in place, pass the same names for input_cols and output_cols.
+        be set explicitly for transformers. Output columns can also be set after
+        initialization with the `set_output_cols` method.
 
     sample_weight_col: Optional[str]
         A string representing the column name containing the sample weights.
-        This argument is only required when working with weighted datasets.
+        This argument is only required when working with weighted datasets. Sample
+        weight column can also be set after initialization with the
+        `set_sample_weight_col` method.
+
+    passthrough_cols: Optional[Union[str, List[str]]]
+        A string or a list of strings indicating column names to be excluded from any
+        operations (such as train, transform, or inference). These specified column(s)
+        will remain untouched throughout the process. This option is helpful in scenarios
+        requiring automatic input_cols inference, but need to avoid using specific
+        columns, like index columns, during training or inference. Passthrough columns
+        can also be set after initialization with the `set_passthrough_cols` method.
 
     drop_input_cols: Optional[bool], default=False
         If set, the response of predict(), transform() methods will not contain input columns.
+
+    threshold: float, default=0
+        Features with a training-set variance lower than this threshold will
+        be removed. The default is to keep all features with non-zero variance,
+        i.e. remove the features that have the same value in all samples.
     """
 
     def __init__(  # type: ignore[no-untyped-def]
@@ -102,6 +117,7 @@ class VarianceThreshold(BaseTransformer):
         input_cols: Optional[Union[str, Iterable[str]]] = None,
         output_cols: Optional[Union[str, Iterable[str]]] = None,
         label_cols: Optional[Union[str, Iterable[str]]] = None,
+        passthrough_cols: Optional[Union[str, Iterable[str]]] = None,
         drop_input_cols: Optional[bool] = False,
         sample_weight_col: Optional[str] = None,
     ) -> None:
@@ -110,9 +126,10 @@ class VarianceThreshold(BaseTransformer):
         self.set_input_cols(input_cols)
         self.set_output_cols(output_cols)
         self.set_label_cols(label_cols)
+        self.set_passthrough_cols(passthrough_cols)
         self.set_drop_input_cols(drop_input_cols)
         self.set_sample_weight_col(sample_weight_col)        
-        deps = set(SklearnWrapperProvider().dependencies)
+        deps: Set[str] = set([f'numpy=={np.__version__}', f'scikit-learn=={sklearn.__version__}', f'cloudpickle=={cp.__version__}'])
         
         self._deps = list(deps)
         
@@ -121,13 +138,14 @@ class VarianceThreshold(BaseTransformer):
             args=init_args,
             klass=sklearn.feature_selection.VarianceThreshold
         )
-        self._sklearn_object = sklearn.feature_selection.VarianceThreshold(
+        self._sklearn_object: Any = sklearn.feature_selection.VarianceThreshold(
             **cleaned_up_init_args,
         )
         self._model_signature_dict: Optional[Dict[str, ModelSignature]] = None
         # If user used snowpark dataframe during fit, here it stores the snowpark input_cols, otherwise the processed input_cols
         self._snowpark_cols: Optional[List[str]] = self.input_cols
-        self._handlers: FitPredictHandlers = HandlersImpl(class_name=VarianceThreshold.__class__.__name__, subproject=_SUBPROJECT, autogenerated=True, wrapper_provider=SklearnWrapperProvider())
+        self._handlers: FitPredictHandlers = HandlersImpl(class_name=VarianceThreshold.__class__.__name__, subproject=_SUBPROJECT, autogenerated=True)
+        self._autogenerated = True
 
     def _get_rand_id(self) -> str:
         """
@@ -138,24 +156,6 @@ class VarianceThreshold(BaseTransformer):
         """
         return str(uuid4()).replace("-", "_").upper()
 
-    def _infer_input_output_cols(self, dataset: Union[DataFrame, pd.DataFrame]) -> None:
-        """
-        Infer `self.input_cols` and `self.output_cols` if they are not explicitly set.
-
-        Args:
-            dataset: Input dataset.
-        """
-        if not self.input_cols:
-            cols = [
-                c for c in dataset.columns 
-                if c not in self.get_label_cols() and c != self.sample_weight_col
-            ]
-            self.set_input_cols(input_cols=cols)
-
-        if not self.output_cols:
-            cols = [identifier.concat_names(ids=['OUTPUT_', c]) for c in self.label_cols]
-            self.set_output_cols(output_cols=cols)
-
     def set_input_cols(self, input_cols: Optional[Union[str, Iterable[str]]]) -> "VarianceThreshold":
         """
         Input columns setter.
@@ -201,54 +201,48 @@ class VarianceThreshold(BaseTransformer):
             self
         """
         self._infer_input_output_cols(dataset)
-        if isinstance(dataset, pd.DataFrame):
-            assert self._sklearn_object is not None  # keep mypy happy
-            self._sklearn_object = self._handlers.fit_pandas(
-                dataset,
-                self._sklearn_object,
-                self.input_cols,
-                self.label_cols,
-                self.sample_weight_col
-            )
-        elif isinstance(dataset, DataFrame):
-            self._fit_snowpark(dataset)
-        else:
-            raise TypeError(
-                f"Unexpected dataset type: {type(dataset)}."
-                "Supported dataset types: snowpark.DataFrame, pandas.DataFrame."
-            )
+        if isinstance(dataset, DataFrame):
+            session = dataset._session
+            assert session is not None  # keep mypy happy
+            # Validate that key package version in user workspace are supported in snowflake conda channel
+            # If customer doesn't have package in conda channel, replace the ones have the closest versions
+            self._deps = pkg_version_utils.get_valid_pkg_versions_supported_in_snowflake_conda_channel(
+                pkg_versions=self._get_dependencies(), session=session, subproject=_SUBPROJECT)
+            
+            # Specify input columns so column pruning will be enforced
+            selected_cols = self._get_active_columns()
+            if len(selected_cols) > 0:
+                dataset = dataset.select(selected_cols)
+
+            self._snowpark_cols = dataset.select(self.input_cols).columns
+
+             # If we are already in a stored procedure, no need to kick off another one.
+            if SNOWML_SPROC_ENV in os.environ:
+                statement_params = telemetry.get_function_usage_statement_params(
+                    project=_PROJECT,
+                    subproject=_SUBPROJECT,
+                    function_name=telemetry.get_statement_params_full_func_name(inspect.currentframe(), VarianceThreshold.__class__.__name__),
+                    api_calls=[Session.call],
+                    custom_tags=dict([("autogen", True)]) if self._autogenerated else None,
+                )
+                pd_df: pd.DataFrame = dataset.to_pandas(statement_params=statement_params)
+                pd_df.columns = dataset.columns
+                dataset = pd_df
+
+        model_trainer = ModelTrainerBuilder.build(
+            estimator=self._sklearn_object,
+            dataset=dataset,
+            input_cols=self.input_cols,
+            label_cols=self.label_cols,
+            sample_weight_col=self.sample_weight_col,
+            autogenerated=self._autogenerated,
+            subproject=_SUBPROJECT
+        )
+        self._sklearn_object = model_trainer.train()
         self._is_fitted = True
         self._get_model_signatures(dataset)
         return self
 
-    def _fit_snowpark(self, dataset: DataFrame) -> None:
-        session = dataset._session
-        assert session is not None  # keep mypy happy
-        # Validate that key package version in user workspace are supported in snowflake conda channel
-        # If customer doesn't have package in conda channel, replace the ones have the closest versions
-        self._deps = pkg_version_utils.get_valid_pkg_versions_supported_in_snowflake_conda_channel(
-            pkg_versions=self._get_dependencies(), session=session, subproject=_SUBPROJECT)
-        
-        # Specify input columns so column pruning will be enforced
-        selected_cols = self._get_active_columns()
-        if len(selected_cols) > 0:
-            dataset = dataset.select(selected_cols)
-
-        estimator = self._sklearn_object
-        assert estimator is not None  # Keep mypy happy
-
-        self._snowpark_cols = dataset.select(self.input_cols).columns
-
-        self._sklearn_object = self._handlers.fit_snowpark(
-            dataset,
-            session,
-            estimator,
-            ["snowflake-snowpark-python"] + self._get_dependencies(),
-            self.input_cols,
-            self.label_cols,
-            self.sample_weight_col,
-        )
-
     def _get_pass_through_columns(self, dataset: DataFrame) -> List[str]:
         if self._drop_input_cols:
             return []
@@ -436,11 +430,6 @@ class VarianceThreshold(BaseTransformer):
         subproject=_SUBPROJECT,
         custom_tags=dict([("autogen", True)]),
     )
-    @telemetry.add_stmt_params_to_df(
-        project=_PROJECT,
-        subproject=_SUBPROJECT,
-        custom_tags=dict([("autogen", True)]),
-    )
     def predict(self, dataset: Union[DataFrame, pd.DataFrame]) -> Union[DataFrame, pd.DataFrame]:
         """Method not supported for this class.
 
@@ -492,11 +481,6 @@ class VarianceThreshold(BaseTransformer):
         subproject=_SUBPROJECT,
         custom_tags=dict([("autogen", True)]),
     )
-    @telemetry.add_stmt_params_to_df(
-        project=_PROJECT,
-        subproject=_SUBPROJECT,
-        custom_tags=dict([("autogen", True)]),
-    )
     def transform(self, dataset: Union[DataFrame, pd.DataFrame]) -> Union[DataFrame, pd.DataFrame]:
         """Reduce X to the selected features
         For more details on this function, see [sklearn.feature_selection.VarianceThreshold.transform]
@@ -555,7 +539,8 @@ class VarianceThreshold(BaseTransformer):
         if False:
             self.fit(dataset)
             assert self._sklearn_object is not None
-            return self._sklearn_object.labels_
+            labels : npt.NDArray[Any] = self._sklearn_object.labels_
+            return labels
         else:
             raise NotImplementedError
 
@@ -591,6 +576,7 @@ class VarianceThreshold(BaseTransformer):
             output_cols = []
 
         # Make sure column names are valid snowflake identifiers.
+        assert output_cols is not None  # Make MyPy happy
         rv = [identifier.rename_to_valid_snowflake_identifier(c) for c in output_cols]
 
         return rv
@@ -601,11 +587,6 @@ class VarianceThreshold(BaseTransformer):
         subproject=_SUBPROJECT,
         custom_tags=dict([("autogen", True)]),
     )
-    @telemetry.add_stmt_params_to_df(
-        project=_PROJECT,
-        subproject=_SUBPROJECT,
-        custom_tags=dict([("autogen", True)]),
-    )
     def predict_proba(
         self, dataset: Union[DataFrame, pd.DataFrame], output_cols_prefix: str = "predict_proba_"
     ) -> Union[DataFrame, pd.DataFrame]:
@@ -646,11 +627,6 @@ class VarianceThreshold(BaseTransformer):
         subproject=_SUBPROJECT,
         custom_tags=dict([("autogen", True)]),
     )
-    @telemetry.add_stmt_params_to_df(
-        project=_PROJECT,
-        subproject=_SUBPROJECT,
-        custom_tags=dict([("autogen", True)]),
-    )
     def predict_log_proba(
         self, dataset: Union[DataFrame, pd.DataFrame], output_cols_prefix: str = "predict_log_proba_"
     ) -> Union[DataFrame, pd.DataFrame]:
@@ -687,16 +663,6 @@ class VarianceThreshold(BaseTransformer):
         return output_df
 
     @available_if(original_estimator_has_callable("decision_function"))  # type: ignore[misc]
-    @telemetry.send_api_usage_telemetry(
-        project=_PROJECT,
-        subproject=_SUBPROJECT,
-        custom_tags=dict([("autogen", True)]),
-    )
-    @telemetry.add_stmt_params_to_df(
-        project=_PROJECT,
-        subproject=_SUBPROJECT,
-        custom_tags=dict([("autogen", True)]),
-    )
     def decision_function(
         self, dataset: Union[DataFrame, pd.DataFrame], output_cols_prefix: str = "decision_function_"
     ) -> Union[DataFrame, pd.DataFrame]:
@@ -795,11 +761,6 @@ class VarianceThreshold(BaseTransformer):
         subproject=_SUBPROJECT,
         custom_tags=dict([("autogen", True)]),
     )
-    @telemetry.add_stmt_params_to_df(
-        project=_PROJECT,
-        subproject=_SUBPROJECT,
-        custom_tags=dict([("autogen", True)]),
-    )
     def kneighbors(
         self,
         dataset: Union[DataFrame, pd.DataFrame],
@@ -859,18 +820,28 @@ class VarianceThreshold(BaseTransformer):
             # For classifier, the type of predict is the same as the type of label
             if self._sklearn_object._estimator_type == 'classifier':
                  # label columns is the desired type for output
-                outputs = _infer_signature(dataset[self.label_cols], "output", use_snowflake_identifiers=True)
+                outputs = list(_infer_signature(dataset[self.label_cols], "output", use_snowflake_identifiers=True))
                 # rename the output columns
-                outputs = model_signature_utils.rename_features(outputs, self.output_cols)
+                outputs = list(model_signature_utils.rename_features(outputs, self.output_cols))
                 self._model_signature_dict["predict"] = ModelSignature(inputs,
                                                                        ([] if self._drop_input_cols else inputs)
                                                                        + outputs)
+            # For mixture models that use the density mixin, `predict` returns the argmax of the log prob.
+            # For outlier models, returns -1 for outliers and 1 for inliers.
+            # Clusterer returns int64 cluster labels. 
+            elif self._sklearn_object._estimator_type in ["DensityEstimator", "clusterer", "outlier_detector"]:
+                outputs = [FeatureSpec(dtype=DataType.INT64, name=c) for c in self.output_cols]
+                self._model_signature_dict["predict"] = ModelSignature(inputs,
+                                                                       ([] if self._drop_input_cols else inputs)
+                                                                       + outputs)
+            
             # For regressor, the type of predict is float64
             elif self._sklearn_object._estimator_type == 'regressor':
                 outputs = [FeatureSpec(dtype=DataType.DOUBLE, name=c) for c in self.output_cols]
                 self._model_signature_dict["predict"] = ModelSignature(inputs,
                                                                        ([] if self._drop_input_cols else inputs)
                                                                        + outputs)
+                
         for prob_func in PROB_FUNCTIONS:
             if hasattr(self, prob_func):
                 output_cols_prefix: str = f"{prob_func}_"

snowflake/ml/modeling/framework/base.py

@@ -16,13 +16,16 @@ from snowflake.ml._internal.exceptions import (
     exceptions,
     modeling_error_messages,
 )
-from snowflake.ml._internal.utils import parallelize
+from snowflake.ml._internal.utils import identifier, parallelize
 from snowflake.ml.modeling.framework import _utils
 from snowflake.snowpark import functions as F
 
 PROJECT = "ModelDevelopment"
 SUBPROJECT = "Preprocessing"
 
+SKLEARN_SUPERVISED_ESTIMATORS = ["regressor", "classifier"]
+SKLEARN_SINGLE_OUTPUT_ESTIMATORS = ["DensityEstimator", "clusterer", "outlier_detector"]
+
 
 def _process_cols(cols: Optional[Union[str, Iterable[str]]]) -> List[str]:
     """Convert cols to a list."""
@@ -48,10 +51,13 @@ class Base:
             input_cols: Input columns.
             output_cols: Output columns.
             label_cols: Label column(s).
+            passthrough_cols: List columns not to be used or modified by the estimator/trasformers.
+                These columns will be passed through all the estimator/trasformer operations without any modifications.
         """
         self.input_cols: List[str] = []
         self.output_cols: List[str] = []
         self.label_cols: List[str] = []
+        self.passthrough_cols: List[str] = []
 
     def _create_unfitted_sklearn_object(self) -> Any:
         raise NotImplementedError()
@@ -125,6 +131,29 @@ class Base:
         self.label_cols = _process_cols(label_cols)
         return self
 
+    def get_passthrough_cols(self) -> List[str]:
+        """
+        Passthrough columns getter.
+
+        Returns:
+            Passthrough column(s).
+        """
+        return self.passthrough_cols
+
+    def set_passthrough_cols(self, passthrough_cols: Optional[Union[str, Iterable[str]]]) -> "Base":
+        """
+        Passthrough columns setter.
+
+        Args:
+            passthrough_cols: Column(s) that should not be used or modified by the estimator/transformer.
+                Estimator/Transformer just passthrough these columns without any modifications.
+
+        Returns:
+            self
+        """
+        self.passthrough_cols = _process_cols(passthrough_cols)
+        return self
+
     def _check_input_cols(self) -> None:
         """
         Check if `self.input_cols` is set.
@@ -495,6 +524,59 @@ class BaseTransformer(BaseEstimator):
                 ),
             )
 
+    def _infer_input_output_cols(self, dataset: Union[snowpark.DataFrame, pd.DataFrame]) -> None:
+        """
+        Infer `self.input_cols` and `self.output_cols` if they are not explicitly set.
+
+        Args:
+            dataset: Input dataset.
+
+        Raises:
+            SnowflakeMLException: If unable to infer output columns
+        """
+        if not self.input_cols:
+            cols = [
+                c
+                for c in dataset.columns
+                if (
+                    c not in self.get_label_cols()
+                    and c not in self.get_passthrough_cols()
+                    and c != self.sample_weight_col
+                )
+            ]
+            self.set_input_cols(input_cols=cols)
+
+        if not self.output_cols:
+            # keep mypy happy
+            assert self._sklearn_object is not None
+
+            if hasattr(self._sklearn_object, "_estimator_type"):
+                # For supervised estimators, infer the output columns from the label columns
+                if self._sklearn_object._estimator_type in SKLEARN_SUPERVISED_ESTIMATORS:
+                    cols = [identifier.concat_names(["OUTPUT_", c]) for c in self.label_cols]
+                    self.set_output_cols(output_cols=cols)
+
+                # For density estimators, clusterers, and outlier detectors, there is always exactly one output column.
+                elif self._sklearn_object._estimator_type in SKLEARN_SINGLE_OUTPUT_ESTIMATORS:
+                    self.set_output_cols(output_cols=["OUTPUT_0"])
+
+                else:
+                    raise exceptions.SnowflakeMLException(
+                        error_code=error_codes.INVALID_ARGUMENT,
+                        original_exception=ValueError(
+                            f"Unable to infer output columns for estimator type {self._sklearn_object._estimator_type}."
+                            f"Please include `output_cols` explicitly."
+                        ),
+                    )
+            else:
+                raise exceptions.SnowflakeMLException(
+                    error_code=error_codes.INVALID_ARGUMENT,
+                    original_exception=ValueError(
+                        f"Unable to infer output columns for object {self._sklearn_object}."
+                        f"Please include `output_cols` explicitly."
+                    ),
+                )
+
     def set_drop_input_cols(self, drop_input_cols: Optional[bool] = False) -> None:
         self._drop_input_cols = drop_input_cols