teradataml 17.20.0.7__py3-none-any.whl → 20.0.0.0__py3-none-any.whl

teradataml/opensource/sklearn/_sklearn_wrapper.py

@@ -0,0 +1,1668 @@
+# ################################################################## 
+# 
+# Copyright 2023 Teradata. All rights reserved.
+# TERADATA CONFIDENTIAL AND TRADE SECRET
+# 
+# Primary Owner: Adithya Avvaru (adithya.avvaru@teradata.com)
+# Secondary Owner: Pankaj Purandare (pankajvinod.purandare@teradata.com)
+# 
+# Version: 1.0
+# Function Version: 1.0
+#
+# This file contains object wrapper class for opensource packages and child object
+# wrapper classes for each opensource package. Currently, we have child object
+# wrapper class for scikit-learn.
+# 
+# ################################################################## 
+
+from collections import OrderedDict, defaultdict
+from importlib import import_module
+
+import base64
+import functools
+import json
+import numpy
+import os
+import pickle
+import time
+import inspect
+import warnings
+import json
+import random
+import pandas as pd
+from teradatasqlalchemy import BLOB, CLOB, FLOAT, TIMESTAMP, VARCHAR, INTEGER
+import pandas.api.types as pt
+
+from teradataml import _TDML_DIRECTORY, Script, TeradataMlException, Apply
+from teradataml.dataframe.copy_to import _get_sqlalchemy_mapping
+from teradataml.common import pylogger
+from teradataml.common.utils import UtilFuncs
+from teradataml.context.context import _get_current_databasename, get_connection
+from teradataml.dbutils.filemgr import install_file, remove_file
+from teradataml.utils.utils import execute_sql
+from teradataml.options.configure import configure
+from teradataml.opensource.sklearn._wrapper_utils import _validate_fit_run, _generate_new_name,\
+    _validate_opensource_func_args, _derive_df_and_required_columns, _validate_df_query_type
+from teradataml.opensource.sklearn.constants import OpenSourcePackage, _OSML_MODELS_PRIMARY_INDEX,\
+    _OSML_MODELS_TABLE_NAME, _OSML_MODELS_TABLE_COLUMNS_TYPE_DICT, OpensourceModels,\
+    _OSML_ADDITIONAL_COLUMN_TYPES
+from teradataml.common.messagecodes import MessageCodes
+from teradataml.common.messages import Messages
+from teradataml.catalog.byom import save_byom, retrieve_byom, delete_byom
+from teradataml.dbutils.dbutils import _create_table
+from teradataml.utils.validators import _Validators
+from teradataml.dataframe.dataframe import DataFrame
+from teradataml.dataframe.dataframe_utils import DataFrameUtils
+from teradataml.scriptmgmt.lls_utils import create_env, get_env
+from teradataml.common.garbagecollector import GarbageCollector
+from teradataml.common.constants import TeradataConstants
+
+
+logger = pylogger.getLogger()
+
+validator = _Validators()
+
+installed_model_files = defaultdict(int)
+
+class _GenericObjectWrapper:
+    def __init__(self) -> None:
+        self._db_name = _get_current_databasename()
+
+        self._scripts_path = os.path.join(_TDML_DIRECTORY, "data", "scripts", "sklearn")
+
+        # Some random number to be used as partition value if partition_columns is None for fit().
+        self._default_data_partition_value = -1001
+
+        self.modelObj = None
+        self._model_data = None
+
+        self._tdml_tmp_dir = os.path.join(os.path.expanduser("~"), ".teradataml")
+
+        self._env = None
+
+        self._is_lake_system = UtilFuncs._is_lake()
+
+        if self._is_lake_system:
+            if configure.openml_user_env is not None:
+                self._env = configure.openml_user_env
+            else:
+                self._create_or_get_env()
+        else:
+            execute_sql(f"SET SESSION SEARCHUIFDBPATH = {self._db_name};")
+
+    def _create_or_get_env(self):
+        """
+        Internal function to return the env if already exists else
+        creates the environment using template file and return the env.
+        """
+        # Get the template file path.
+        template_dir_path = os.path.join(_TDML_DIRECTORY, "data", "templates",
+                                         "open_source_ml.json")
+
+        # Read template file.
+        with open(template_dir_path, "r") as r_file:
+            data = json.load(r_file)
+
+        # Get env_name.
+        _env_name = data["env_specs"][0]["env_name"]
+
+        try:
+            # Call function to 'openml_env' get env.
+            self._env = get_env(_env_name)
+        except TeradataMlException as tdml_e:
+            # We will get here when error says, env does not exist otherwise raise the exception as is.
+            # Env does not exist so create one.
+
+            exc_msg = "Failed to execute get_env(). User environment '{}' not " \
+                      "found.".format(_env_name)
+            if exc_msg in tdml_e.args[0]:
+                print(f"No OpenAF environment with name '{_env_name}' found. Creating one with "\
+                      "latest supported python and required packages.")
+                _env = create_env(template=template_dir_path)
+            else:
+                raise tdml_e
+        except Exception as exc:
+            raise exc
+
+    def _get_columns_as_list(self, cols):
+        """
+        Internal function to get columns as list of strings.
+        Empty list is returned if cols is None.
+        """
+        if cols is None:
+            return []
+        if not isinstance(cols, list) and not isinstance(cols, tuple):
+            return [cols]
+        return cols
+
+    def _get_data_and_data_partition_columns(self, data, feature_columns, label_columns,
+                                             partition_columns=None, group_columns=[]):
+        """
+        Internal function to generate one new partition column (if not provided) and return
+        data and partition columns (either generated or passed one).
+        """
+        new_partition_columns = self._get_columns_as_list(partition_columns)
+
+        if not partition_columns:
+            # If partition column is not specified, create a partition column and run Script.
+            # This runs the Script in one AMP as we are partitioning data using this column
+            # which contains only one value.
+            new_partition_columns = [_generate_new_name(type="column")]
+            data = data.assign(**{new_partition_columns[0]: self._default_data_partition_value})
+
+        # Filter out partition columns from feature columns and label columns.
+        new_partition_columns_filtered = [col for col in new_partition_columns
+                                          if col not in (feature_columns + label_columns + group_columns)]
+
+        all_columns = feature_columns + label_columns + group_columns + new_partition_columns_filtered
+        return data.select(all_columns), new_partition_columns
+
+    def _run_script(self, data, command, partition_columns, return_types):
+        """
+        Internal function to run Script(), given the argument needed by STO's or
+        Apply's Script.
+        """
+        if isinstance(partition_columns, list) and len(partition_columns) == 0:
+            partition_columns = None
+
+        if self._is_lake_system:
+            obj = Apply(data=data,
+                        returns=OrderedDict(return_types),
+                        apply_command=command,
+                        data_partition_column=partition_columns,
+                        env_name=self._env,
+                        delimiter="\t")
+        else:
+            obj = Script(data=data,
+                         returns=OrderedDict(return_types),
+                         script_command=command,
+                         data_partition_column=partition_columns)
+            obj.check_reserved_keyword = False
+
+        obj.skip_argument_validation = True
+        return obj.execute_script(output_style="TABLE")
+
+    def _install_script_file(self, 
+                             file_identifier=None, 
+                             file_name=None, 
+                             is_binary=False, 
+                             file_location=None):
+        """
+        Internal function to install script file in Vantage.
+        """
+        if file_location is None:
+            file_location = self._scripts_path
+        new_script = os.path.join(file_location, file_name)
+
+        # _env is set while object creation
+        # If not set, it is Vantage Enterprise. Otherwise, it is Vantage Lake.
+
+        if not self._is_lake_system:
+            status = install_file(file_identifier=file_identifier,
+                                  file_path=new_script,
+                                  replace=True,
+                                  suppress_output=True,
+                                  is_binary=is_binary)
+        else:
+            status = self._env.install_file(file_path=new_script,
+                                       replace=True,
+                                       suppress_output=True)
+        if not status:
+            raise TeradataMlException(
+                f"Script file '{file_name}' failed to get installed/replaced in Vantage."
+            )
+
+    def _get_partition_col_indices_and_types(self, data, partition_columns):
+        """
+            partition_columns can be from feature columns and label columns.
+            So, get the indices and types of these columns from the data columns.
+        """
+        partition_indices = []
+        partition_types = []
+        new_partition_columns = []
+        for i, col in enumerate(data.columns):
+            if col in partition_columns:
+                new_partition_columns.append(col)
+                partition_indices.append(i)
+                partition_types.append(data._td_column_names_and_sqlalchemy_types[col.lower()].\
+                                       python_type.__name__)
+        # Converting to string "None" if they are not present as empty string can't be passed
+        # to Script script_commands' command line arguments.
+        # Otherwise, pass the values as comma separated string.
+        partition_indices = ",".join([str(x) for x in partition_indices])\
+            if partition_indices else "None"
+        partition_types = ",".join([x for x in partition_types]) if partition_types else "None"
+        return partition_indices, partition_types, new_partition_columns
+
+    def _get_kwargs_str(self, kwargs):
+        """
+        Returns string of kwargs in the format:
+            key1 val1-type1 key2 val2-type2 ...
+        """
+        args_str = ""
+        for key, val in kwargs.items():
+            strr = f"{key} {str(val)}-{type(val).__name__}"
+            if args_str == "":
+                args_str += strr
+            else:
+                args_str += f" {strr}"
+        return args_str
+
+    def extract_sklearn_obj(self, n_unique_partitions = 1, n_partition_cols = 1):
+        """
+        Internal function to extract sklearn object from the model(s) depending on the number of
+        partitions. When it is only one model, it is directly used as sklearn object (modelObj).
+        When it is multiple models, it is converted to pandas DataFrame and stored in sklearn
+        object.
+        """
+        vals = execute_sql("select * from {}".format(self._model_data._table_name)).fetchall()
+
+        # pickle will issue a caution warning, if model pickling was done with
+        # different library version than used here. The following disables any warnings
+        # that might otherwise show in the scriptlog files on the Advanced SQL Engine
+        # nodes in this case. Yet, do keep an eye for incompatible pickle versions.
+        warnings.filterwarnings("ignore")
+
+        model_obj = None
+        # Extract and unpickle last column which is the model object.
+        for i, row in enumerate(vals):
+            if self._is_lake_system:
+                model_obj = pickle.loads(row[n_partition_cols])
+            else:
+                model_obj = pickle.loads(base64.b64decode(row[n_partition_cols].partition("'")[2]))
+            row[n_partition_cols] = model_obj
+            vals[i] = row
+        if n_unique_partitions == 1:
+            self.modelObj = model_obj
+        elif n_unique_partitions > 1:
+            self.modelObj = pd.DataFrame(vals, columns=self._model_data.columns)
+        else:
+            ValueError("Number of partitions should be greater than 0.")
+
+        warnings.filterwarnings("default")
+
+
+class _OpenSourceObjectWrapper(_GenericObjectWrapper):
+    # This has to be set for every package which subclasses this class.
+    OPENSOURCE_PACKAGE_NAME = None
+
+    def __init__(self, model=None, module_name=None, class_name=None, pos_args=None, kwargs=None):
+        if not model and not module_name and not class_name:
+            raise TeradataMlException(Messages.get_message(MessageCodes.EITHER_THIS_OR_THAT_ARGUMENT, "model",
+                                                           "module_name and class_name"),
+                                      MessageCodes.EITHER_THIS_OR_THAT_ARGUMENT)
+
+        validator._validate_mutually_inclusive_arguments(module_name, "module_name",
+                                                         class_name, "class_name")
+
+        super().__init__()
+
+        self.module_name = module_name
+        self.class_name = class_name
+        self.kwargs = kwargs if kwargs is not None else {}
+        self.pos_args = pos_args if pos_args is not None else tuple()
+
+        self._fit_label_columns_types = None
+        self._table_name_prefix = None
+
+        self._is_default_partition_value_fit = True # False when the user provides partition columns.
+        self._fit_partition_colums_non_default = None
+        self._is_default_partition_value_predict = True # False when the user provides partition columns.
+
+    def _validate_equality_of_partition_values(self, fit_values, trans_values):
+        """
+        Internal function to compare the partition values in fit() and predict() are same.
+        """
+        if len(fit_values) != len(trans_values):
+            return False
+
+        for val in fit_values:
+            if not all([val in trans_values]):
+                return False
+
+        return True
+
+    def _validate_unique_partition_values(self, data, partition_columns):
+        """
+        Internal function to validate if the partition values in partition_columns used in fit()
+        and predict() are same.
+        """
+        data._index_label = None
+        unique_values = data.drop_duplicate(partition_columns).get_values()
+
+        trans_unique_values = sorted(unique_values.tolist(), key=lambda x: tuple(x))
+        fit_unique_values = sorted(self._fit_partition_unique_values.tolist() \
+                                    if not isinstance(self._fit_partition_unique_values, list) \
+                                    else self._fit_partition_unique_values, key=lambda x: tuple(x))
+        default_unique_values = [[self._default_data_partition_value]]
+
+        if fit_unique_values == default_unique_values and \
+            trans_unique_values != default_unique_values:
+            error_msg = Messages.get_message(MessageCodes.PARTITION_IN_BOTH_FIT_AND_PREDICT,
+                                             "without", "with")
+            msg_code = MessageCodes.PARTITION_IN_BOTH_FIT_AND_PREDICT
+            raise TeradataMlException(error_msg, msg_code)
+
+        if not self._validate_equality_of_partition_values(fit_unique_values, trans_unique_values):
+            raise TeradataMlException(
+                Messages.get_message(MessageCodes.PARTITION_VALUES_NOT_MATCHING),
+                MessageCodes.PARTITION_VALUES_NOT_MATCHING
+            )
+
+    def fit(self, **kwargs):
+        pass
+
+    def __get_obj_attributes_multi_model(self, name):
+        """
+        Internal function to get attributes of all sklearn model objects when multiple models are
+        generated by fit.
+        """
+        # Wrapper function to invoke dynamic method, using arguments
+        # passed by user, on model in each row.
+        def __sklearn_method_invoker_for_multimodel(*c, **kwargs):
+            multi_models = self.modelObj.copy()
+            for i in range(multi_models.shape[0]):
+                curr_model = multi_models.iloc[i]["model"]
+                multi_models.at[i, "model"] = getattr(curr_model, name)(*c, **kwargs)
+            return multi_models.rename(columns={"model": name})
+
+        # Identify if attribute is callable or not to avoid
+        # this check in loop for every model.
+        is_attr_callable = False
+        # Assuming that self.modelObj will have at least 1 row.
+        is_attr_callable = callable(getattr(self.modelObj.iloc[0]["model"], name))
+
+        # If attribute is callable, it should be applied on model in each row
+        # using passed arguments.
+        if is_attr_callable:
+            return __sklearn_method_invoker_for_multimodel
+
+        output_attributes = self.modelObj.copy()
+        for i in range(output_attributes.shape[0]):
+            model = output_attributes.iloc[i]["model"]
+            output_attributes.at[i, "model"] = getattr(model, name)
+        return output_attributes.rename(columns={"model": name})
+
+    def __getattr__(self, name):
+        # This just run attributes (functions and properties) from sklearn object.
+        def __sklearn_method_invoker(*c, **kwargs):
+            return atrribute_instance(*c, **kwargs)
+        if isinstance(self.modelObj, pd.DataFrame):
+            return self.__get_obj_attributes_multi_model(name)
+
+        atrribute_instance = getattr(self.modelObj, name)
+        if callable(atrribute_instance):
+            return __sklearn_method_invoker
+        return atrribute_instance
+
+    @classmethod
+    def _validate_model_supportability(cls, model):
+        """
+        Internal function to validate if the model provided for deployment is supported by
+        teradataml's opensourceML.
+        """
+        error_msg = Messages.get_message(MessageCodes.MODEL_CATALOGING_OPERATION_FAILED, "validate",
+                                         "The given model is not a supported opensource model.")
+        msg_code = MessageCodes.MODEL_CATALOGING_OPERATION_FAILED
+        try:
+            # For scikit-learn, model.__module__ is similar to 'sklearn.linear_model._base'.
+            # TODO: check for other supported packages.
+            if model.__module__.split(".")[0] not in OpenSourcePackage.values():
+                raise TeradataMlException(error_msg, msg_code)
+        except Exception as ex:
+            # If in case, model.__module__ fails.
+            raise TeradataMlException(error_msg, msg_code) from ex
+
+    def _save_model(self, model_name, replace_if_exists=False):
+        """
+        Internal function to save the model stored in file at location mentioned by class variable
+        "model_file_path_local" to Vantage using BYOM methods save_byom() and delete_byom() based
+        on the value of "replace_if_exists" argument.
+        """
+        # Creating a table, if doesn't exist, in Vantage to store the model info.
+        conn = get_connection()
+        osml_models_table_exists = conn.dialect.has_table(conn,
+                                                          table_name=_OSML_MODELS_TABLE_NAME,
+                                                          schema=self._db_name)
+        if not osml_models_table_exists:
+            all_columns = _OSML_MODELS_TABLE_COLUMNS_TYPE_DICT.copy()
+            all_columns.update(_OSML_ADDITIONAL_COLUMN_TYPES)
+            _create_table(table_name=_OSML_MODELS_TABLE_NAME, columns=all_columns,
+                          primary_index=_OSML_MODELS_PRIMARY_INDEX, schema_name=self._db_name)
+
+        model_obj = OpensourceModels(is_default_partition_value=self._is_default_partition_value_fit, 
+                                     partition_file_prefix=self._model_file_name_prefix,
+                                     fit_partition_columns_non_default=self._fit_partition_colums_non_default,
+                                     model=self.modelObj,
+                                     pos_args=self.pos_args,
+                                     key_args=self.kwargs)
+
+        # Saved the model object to a file to be used in save_byom() for writing to Vantage table.
+        file_name = os.path.join(self._tdml_tmp_dir, "deployed_file.pickle")
+        with open(file_name, "wb+") as fp:
+            fp.write(pickle.dumps(model_obj))
+
+        try:
+            save_byom(model_id=model_name,
+                    model_file=file_name,
+                    table_name=_OSML_MODELS_TABLE_NAME,
+                    additional_columns_types=_OSML_ADDITIONAL_COLUMN_TYPES,
+                    additional_columns={"package": self.OPENSOURCE_PACKAGE_NAME.value})
+        except TeradataMlException as ex:
+            model_exists_msg = Messages.get_message(MessageCodes.MODEL_ALREADY_EXISTS, model_name)
+            if not replace_if_exists and model_exists_msg == str(ex):
+                raise
+            elif replace_if_exists and model_exists_msg == str(ex):
+                # Delete the model from Model table and save again.
+                delete_byom(model_id=model_name, table_name=_OSML_MODELS_TABLE_NAME)
+                save_byom(model_id=model_name,
+                          model_file=file_name,
+                          table_name=_OSML_MODELS_TABLE_NAME,
+                          additional_columns_types=_OSML_ADDITIONAL_COLUMN_TYPES,
+                          additional_columns={"package": self.OPENSOURCE_PACKAGE_NAME.value})
+            else:
+                raise
+        finally:
+            os.remove(file_name)
+
+    @classmethod
+    def _deploy(cls, model_name, model, replace_if_exists=False):
+        """
+        Internal function to create an instance of the class using the model and deploy
+        the model to Vantage.
+        """
+        cls._validate_model_supportability(model=model)
+
+        cls = cls(model=model)
+        # Load the model file into Vantage node as file can be used in
+        # predict or other operations.
+        cls._install_initial_model_file()
+
+        cls._save_model(model_name, replace_if_exists)
+        
+        return cls
+
+    @classmethod
+    def _load(cls, model_name):
+        """
+        Internal function to load model corresponding to the package (like sklearn etc)
+        from Vantage to client using retrieve_byom() and create an instance of the class if
+        the model is from the same package.
+        """
+        try:
+            model = retrieve_byom(model_id=model_name, table_name=_OSML_MODELS_TABLE_NAME,
+                                  return_addition_columns=True)
+        except TeradataMlException as ex:
+            # Not showing table name in error message as it is an internal table.
+            part_msg = f"Model '{model_name}' not found in the table "
+            if part_msg in str(ex):
+                raise TeradataMlException(Messages.get_message(MessageCodes.MODEL_NOT_FOUND, model_name, ""),
+                                          MessageCodes.MODEL_NOT_FOUND)
+            raise
+
+        model_vals_list = model.get_values()[0]
+        # List of 3 elements -
+        #   - model name as index column,
+        #   - 1st contains model object with fields: is_default_partition_value, partition_file_prefix, model. etc
+        #   - 2nd contains package name.
+        model_obj = pickle.loads(model_vals_list[0])
+        model = model_obj.model
+        package = model_vals_list[1]
+
+        if package != cls.OPENSOURCE_PACKAGE_NAME.value:
+            # Raise error if trying to access model of different package.
+            raise TeradataMlException(Messages.get_message(MessageCodes.MODEL_NOT_FOUND, model_name, 
+                                        f". Requested model is from '{package}' package"),
+                                      MessageCodes.MODEL_NOT_FOUND)
+
+        if isinstance(model, pd.DataFrame):
+            # Create a new instance of the class and set the model object to the instance.
+            # Instantiation can take only model, not model object. Hence, passing one of the model
+            # from pandas df. Updating modelObj and other fields later
+            cls = cls(model=model.iloc[1,2])
+            cls.modelObj = model
+            cls._fit_partition_unique_values = [lst[:len(lst)-1] for lst in model.values.tolist()]
+        else:
+            cls = cls(model=model)
+        
+        cls._model_file_name_prefix = model_obj.partition_file_prefix
+        cls._is_default_partition_value_fit = model_obj.is_default_partition_value
+        cls._fit_partition_colums_non_default = model_obj.fit_partition_columns_non_default
+        cls.pos_args = model_obj.pos_args
+        cls.kwargs = model_obj.key_args
+
+        # Load the model file into Vantage node as file can be used in
+        # predict or other operations.
+        cls._install_initial_model_file()
+
+        return cls
+
+    def deploy(self, model_name, replace_if_exists=False):
+        """
+        DESCRIPTION:
+            Deploys the model held by interface object to Vantage.
+
+        PARAMETERS:
+            model_name:
+                Required Argument.
+                Specifies the unique name of the model to be deployed.
+                Types: str
+
+            replace_if_exists:
+                Optional Argument.
+                Specifies whether to replace the model if a model with the same name already
+                exists in Vantage. If this argument is set to False and a model with the same
+                name already exists, then the function raises an exception.
+                Default Value: False
+                Types: bool
+
+        RETURNS:
+            The opensource object wrapper.
+
+        RAISES:
+            TeradataMLException if model with "model_name" already exists and the argument
+            "replace_if_exists" is set to False.
+
+        EXAMPLES:
+            >>> from teradataml import td_sklearn
+            >>> model = td_sklearn.LinearRegression(normalize=True)
+            >>> model
+            LinearRegression(normalize=True)
+
+            # Example 1: Deploy the model held by interface object to Vantage.
+            >>> lin_reg = model.deploy("linreg_model_ver_2")
+            Model is saved.
+            >>> lin_reg
+            LinearRegression(normalize=True)
+
+            # Example 2: Deploy the model held by interface object to Vantage with the name same
+            #            as that of model that already existed in Vantage.
+            >>> lin_reg = model.deploy("linreg_model_ver_2", replace_if_exists=True)
+            Model is deleted.
+            Model is saved.
+            >>> lin_reg
+            LinearRegression(normalize=True)
+        """
+
+        # Install model file into Vantage, if not installed.
+        self._install_initial_model_file()
+
+        self._save_model(model_name, replace_if_exists)
+        return self
+
+
+class _SkLearnObjectWrapper(_OpenSourceObjectWrapper):
+
+    OPENSOURCE_PACKAGE_NAME = OpenSourcePackage.SKLEARN
+
+    def __init__(self, model=None, module_name=None, class_name=None, pos_args=None, kwargs=None):
+        super().__init__(model=model, module_name=module_name, class_name=class_name,
+                         pos_args=pos_args, kwargs=kwargs)
+
+        self._initialize_variables()
+        if model:
+            self.modelObj = model
+            self.module_name = model.__module__.split("._")[0]
+            self.class_name = model.__class__.__name__
+            # __dict__ gets all the arguments as dictionary including default ones and positional
+            # args.
+            self.kwargs = model.__dict__
+            self.pos_args = tuple() # Kept empty as all are moved to kwargs.
+        else:
+            self._initialize_object()
+
+    def __repr__(self):
+        if self._is_default_partition_value_fit:
+            # Single model use case.
+            return self.modelObj.__repr__()
+
+        pd.set_option("display.expand_frame_repr", None)
+        pd.set_option("display.max_colwidth", None)
+        opt = self.modelObj.__repr__()
+        pd.reset_option("display.expand_frame_repr")
+        pd.reset_option("display.max_colwidth")
+        return opt
+
+    def _validate_args_and_get_data(self, X=None, y=None, groups=None, kwargs={},
+                                    skip_either_or_that=False):
+        """
+        Internal function to validate arguments passed to exposed opensource APIs and return
+        parent DataFrame, feature columns, label columns, group columns, data partition columns.
+        """
+        _validate_opensource_func_args(X=X, y=y, groups=groups,
+                                       fit_partition_cols=self._fit_partition_colums_non_default,
+                                       kwargs=kwargs,
+                                       skip_either_or_that=skip_either_or_that)
+        return _derive_df_and_required_columns(X=X, y=y, groups=groups, kwargs=kwargs,
+                                        fit_partition_cols=self._fit_partition_colums_non_default)
+
+    def _initialize_object(self):
+        """
+        Internal function to initialize sklearn object from module name and class name.
+        """
+        # Needed when writing imported modules to generated file. TODO: Remove later.
+        imported_args = {} 
+        # If there are any objects of class `_SkLearnObjectWrapper`, it is modified to
+        # corresponding sklearn object.
+        new_sklearn_pos_args = self.modify_args(None, self.pos_args, imported_args)
+        new_sklearn_kwargs = self.modify_args(None, self.kwargs, imported_args)
+
+        # Create model object from new positional and keyword arguments.
+        class_obj = getattr(import_module(self.module_name), self.class_name)
+        if new_sklearn_pos_args:
+            self.modelObj = class_obj(*new_sklearn_pos_args, **new_sklearn_kwargs)
+        else:
+            self.modelObj = class_obj(**new_sklearn_kwargs)
+
+        # All arguments are moved to kwargs and kept pos_args empty.
+        # Might help in set_params() bug fix.
+        self.pos_args = tuple()
+        _arguments = self.modelObj.__dict__
+
+        if hasattr(self.modelObj, "get_params"):
+            # Update kwargs that are both in modelObj and get_params() as there are
+            # some classes which return other internals variables also. 
+            # Hence, filtering them using get_params().
+            for k, v in _arguments.items():
+                if type(v).__name__ in ["function", "generator"]:
+                    # TODO: ELE-6351: Skipping adding functions and generators to kwargs as these
+                    #       are not supported yet due to pickling issue.
+                    continue
+                if k in self.get_params():
+                    self.kwargs[k] = v
+        else:
+            # Model selection classes will not have `get_params`, in which case modelObj's __dict__
+            # is saved as kwargs.
+            self.kwargs = _arguments
+
+    def _initialize_variables(self):
+        """
+        Internal function to initialize variables used in this class.
+        """
+        self.feature_names_in_ = None
+        self._table_name_prefix = "td_sklearn_"
+        self._model_file_name_prefix = _generate_new_name(type="file")
+        self.model_file_paths_local = set()
+
+        self._fit_execution_time = None
+        self._fit_predict_execution_time = None
+        self._partial_fit_execution_time = None
+        self._predict_execution_time = None
+        self._transform_execution_time = None
+        self._score_execution_time = None
+
+        # Set to partition columns when training is done with partition columns.
+        self._fit_partition_colums_non_default = None
+
+        self._is_model_installed = False
+        self._fit_partition_unique_values = [[self._default_data_partition_value]]
+
+    def modify_args(self, fp1, arg, imported_args):
+        """
+        Internal function to recursively (if "arg" is list/tuple/dict) check if any sklearn object
+        of opensourceML is present in the argument "arg" and modify it to corresponding sklearn
+        object.
+        This function can also be used to write import statements to file (if "fp1" is not
+        None). Update "imported_args" dictionary with imported module and class name to avoid
+        importing same module and class again when writing to file. This is useful when we want to
+        generate script from template file.
+        Pass None to "fp1" if we don't want to write to file and just modify opensourceML sklearn
+        object to corresponding sklearn object.
+        """
+        if isinstance(arg, type(self)):
+            imported_tuple = (arg.module_name, arg.class_name)
+            already_imported = imported_args.get(imported_tuple, False)
+            if not already_imported:
+                imported_args[imported_tuple] = True
+                if fp1:
+                    fp1.write(f"from {arg.module_name} import {arg.class_name}\n")
+                self.modify_args(fp1, arg.pos_args, imported_args)
+                self.modify_args(fp1, arg.kwargs, imported_args)
+            return arg.modelObj
+        elif isinstance(arg, list):
+            return [self.modify_args(fp1, val, imported_args) for val in arg]
+        elif isinstance(arg, tuple):
+            return tuple([self.modify_args(fp1, val, imported_args) for val in arg])
+        elif type(arg).__name__ == "generator":
+            # Raising exception as generator object can't be pickled.
+            # TODO: ELE-6351 - Find ways to pickle generator object later.
+            raise ValueError("Generator type/iterator is not supported for any argument. "\
+                             "Support will be added later.")
+        elif type(arg).__name__ == "function":
+            # Raising exception as functions/lambda functions can't be pickled.
+            # TODO: ELE-6351 - Find ways to pickle functions later.
+            raise ValueError("Functions are not supported for any argument. "\
+                             "Support will be added later.")
+        elif isinstance(arg, dict):
+            return dict(
+                (
+                    self.modify_args(fp1, k, imported_args),
+                    self.modify_args(fp1, v, imported_args),
+                )
+                for k, v in arg.items()
+            )
+        else:
+            return arg
+
+    def _install_initial_model_file(self):
+        """
+        If model file(s) is/are not installed in Vantage, then install it/them.
+        """
+        if isinstance(self.modelObj, pd.DataFrame):
+            # Get list of unique partition values and corresponding model object as dict.
+            partition_values_model_dict = {}
+            obj_list = self.modelObj.values.tolist()
+            for lst in obj_list:
+                partition_values_model_dict[tuple(lst[:len(lst)-1])] = lst[len(lst)-1]
+
+        for partition in self._fit_partition_unique_values:
+            # Create a new file with file name with partition values and
+            # dump sklearn object into it. Finally install the file to Vantage.
+            partition_join = "_".join([str(x) for x in partition])
+            file_name = f"{self._model_file_name_prefix}_{partition_join}"
+            # Replace '-' with '_' as '-' can't be present in file identifier.
+            # Needed this replace because partition_columns can be negative.
+            file_name = file_name.replace("-", "_")
+            full_file_name = os.path.join(self._tdml_tmp_dir, file_name)
+            with open(full_file_name, "wb+") as fp:
+                # Write sklearn object to file.
+                if isinstance(self.modelObj, pd.DataFrame):
+                    # If multiple models, then write the model corresponding to the partition value.
+                    fp.write(pickle.dumps(partition_values_model_dict[tuple(partition)]))
+                else:
+                    fp.write(pickle.dumps(self.modelObj))
+            self.model_file_paths_local.add(file_name)
+
+            self._install_script_file(file_identifier=file_name,
+                                      file_name=file_name,
+                                      is_binary=True,
+                                      file_location=self._tdml_tmp_dir)
+
+            if self._is_lake_system:
+                # Need to pass env_name along with file_name for cleaning up the files in env.
+                obj = f"{self._env.env_name}::{file_name}"
+                if installed_model_files[obj] == 0:
+                    # Add to GC for the first time the model file (along with env name) is encountered.
+                    installed_model_files[obj] = 1
+                    GarbageCollector._add_to_garbagecollector(object_name=obj,
+                                                object_type=TeradataConstants.TERADATA_APPLY)
+            else:
+                if installed_model_files[file_name] == 0:
+                    # Add to GC for the first time the model file is encountered.
+                    installed_model_files[file_name] = 1
+                    GarbageCollector._add_to_garbagecollector(object_name=file_name,
+                                                object_type=TeradataConstants.TERADATA_SCRIPT)
+
+            self._is_model_installed = True
+
+    def _run_fit_related_functions(self,
+                                   data,
+                                   feature_columns,
+                                   label_columns,
+                                   partition_columns,
+                                   func,
+                                   classes=None):
+        """
+        Internal function to run fit() and partial_fit() functions.
+        """
+        label_columns = self._get_columns_as_list(label_columns)
+
+        data, new_partition_columns = self._get_data_and_data_partition_columns(data,
+                                                                                feature_columns,
+                                                                                label_columns,
+                                                                                partition_columns)
+
+        model_type = BLOB() if self._is_lake_system else CLOB()
+        return_types = [(col, data._td_column_names_and_sqlalchemy_types[col.lower()]) 
+                        for col in new_partition_columns] + [("model", model_type)]
+
+        file_name = "sklearn_fit.py"
+        self._install_script_file(file_identifier=file_name.split(".")[0], file_name=file_name)
+
+        if classes:
+            class_type = type(classes[0]).__name__
+            classes = "--".join([str(x) for x in classes])
+        else:
+            classes = str(None)
+            class_type = str(None)
+
+        partition_indices, partition_types, new_partition_columns = \
+            self._get_partition_col_indices_and_types(data, new_partition_columns)
+
+        # db_name is applicable for enterprise system.
+        db_file_name = file_name if self._is_lake_system else f"./{self._db_name}/{file_name}"
+        py_exc = UtilFuncs._get_python_execution_path()
+        script_command = f"{py_exc} {db_file_name} {func} {len(feature_columns)} "\
+            f"{len(label_columns)} {partition_indices} {partition_types} "\
+            f"{self._model_file_name_prefix} {classes} {class_type} {self._is_lake_system}"
+
+        # Get unique values in partitioning columns.
+        self._fit_partition_unique_values = data.drop_duplicate(new_partition_columns).get_values()
+
+        self._install_initial_model_file()
+
+        self._model_data = self._run_script(data, script_command, new_partition_columns,
+                                            return_types)
+
+        # Extract sklearn object(s) from the depending on the number of unique partitioning values.
+        self.extract_sklearn_obj(n_unique_partitions=len(self._fit_partition_unique_values),
+                                 n_partition_cols=len(new_partition_columns))
+
+        # Need this label columns types in prediction.
+        self._fit_label_columns_types = [data._td_column_names_and_sqlalchemy_types[l_c.lower()]
+                                         for l_c in label_columns]
+
+    def partial_fit(self, X=None, y=None, classes=None, **kwargs):
+        """
+        Please check the description in Docs/OpensourceML/sklearn.py.
+        """
+        st_time = time.time()
+
+        # "classes" argument validation.
+        arg_info_matrix = []
+        arg_info_matrix.append(["classes", classes, True, (list)])
+        _Validators._validate_function_arguments(arg_info_matrix)
+
+        self._is_default_partition_value_fit = True # False when the user provides partition columns.
+
+        data, feature_columns, label_columns, _, partition_columns = \
+            self._validate_args_and_get_data(X=X, y=y, groups=None, kwargs=kwargs)
+
+        if partition_columns:
+            self._is_default_partition_value_fit = False
+            self._fit_partition_colums_non_default = partition_columns
+
+        self._run_fit_related_functions(data,
+                                        feature_columns,
+                                        label_columns,
+                                        partition_columns,
+                                        inspect.stack()[0][3],
+                                        classes)
+
+        self._partial_fit_execution_time = time.time() - st_time
+
+        return self
+
+    def fit(self, X=None, y=None, **kwargs):
+        """
+        Please check the description in Docs/OpensourceML/sklearn.py.
+        """
+        st_time = time.time()
+
+        self._is_default_partition_value_fit = True # False when the user provides partition columns.
+
+        data, feature_columns, label_columns, _, partition_columns = \
+            self._validate_args_and_get_data(X=X, y=y, groups=None, kwargs=kwargs)
+
+        if partition_columns:
+            self._is_default_partition_value_fit = False
+            self._fit_partition_colums_non_default = partition_columns
+
+        self._run_fit_related_functions(data,
+                                        feature_columns,
+                                        label_columns,
+                                        partition_columns,
+                                        inspect.stack()[0][3])
+
+        self._fit_execution_time = time.time() - st_time
+
+        return self
+
+    def set_params(self, **params):
+        """
+        Please check the description in Docs/OpensourceML/sklearn.py.
+        """
+        for key, val in params.items():
+            self.kwargs[key] = val
+
+        # Initialize with new arguments and return the class/model object.
+        # set_params takes all keyword arguments and no positional arguments.
+        self.__init__(None, self.module_name, self.class_name, tuple(), self.kwargs)
+        return self
+
+    # get_params() will be executed through __getattr__().
+
+    # @_validate_fit_run
+    def __getattr__(self, name):
+        def __run_transform(*c, **kwargs):
+            kwargs["name"] = name
+            return self._transform(*c, **kwargs)
+
+        def __run_function_needing_all_rows(*c, **kwargs):
+            kwargs["name"] = name
+            return self._run_function_needing_all_rows(*c, **kwargs)
+
+        def __run_kneighbors(*c, **kwargs):
+            kwargs["name"] = name
+            return self._run_neighbors(*c, **kwargs)
+
+        if name in ["score", "aic", "bic", "perplexity"]:
+            # TODO: ELE-6352 - Implement error_norm() function later.
+            return __run_function_needing_all_rows
+
+        if name in ["kneighbors",
+                    "radius_neighbors",
+                    "kneighbors_graph",
+                    "radius_neighbors_graph"]:
+            return __run_kneighbors
+
+        if name in ["predict",
+                    "transform",
+                    "inverse_transform",
+                    "predict_proba",
+                    "predict_log_proba",
+                    "decision_function",
+                    "score_samples",
+                    "decision_path",
+                    "apply",
+                    "cost_complexity_pruning_path",
+                    "gibbs",
+                    "kneighbors_graph",
+                    "radius_neighbors_graph",
+                    "mahalanobis",
+                    "correct_covariance",
+                    "reweight_covariance",
+                    "path"]:
+            return __run_transform
+
+        return super().__getattr__(name)
+
+    def _get_return_columns_for_function_(self,
+                                          data,
+                                          feature_columns,
+                                          label_columns,
+                                          func_name,
+                                          n_partitions,
+                                          kwargs):
+        """
+        Internal function to return list of column names and their sqlalchemy types
+        which should be used in return_types of Script.
+        """
+        if func_name == "fit_predict":
+            """
+            Get return columns using label_columns.
+            """
+            return [(f"{self.class_name.lower()}_{func_name}_{(i + 1)}",
+                     data._td_column_names_and_sqlalchemy_types[col.lower()])
+                    for i, col in enumerate(label_columns)]
+        if func_name == "predict":
+            """
+            Return predict columns using either label_columns (if provided) or 
+            self._fit_label_columns_types (if the function is trained using label columns).
+            Otherwise run predict on ten rows of data to get the number of columns and their types
+            after this if condition.
+            """
+            if label_columns:
+                return [(f"{self.class_name.lower()}_{func_name}_{(i + 1)}",
+                         data._td_column_names_and_sqlalchemy_types[col.lower()])
+                             for i, col in enumerate(label_columns)]
+            if self._fit_label_columns_types:
+                return [(f"{self.class_name.lower()}_{func_name}_{(i + 1)}", col_type)
+                        for i, col_type in enumerate(self._fit_label_columns_types)]
+
+        data = data.select(feature_columns + label_columns)
+
+        ## If function is not `fit_predict`:
+        #   then take one row of transform/other functions to execute in client
+        #   to get number of columns in return clause and their Vantage types.
+        n_f = len(feature_columns)
+        n_c = len(label_columns)
+
+        # For paritioning columns, it will be a dataframe and getattr(modelObj, func_name) fails.
+        # Just for getting the number of columns and their types, using only one model of all.
+        if n_partitions == 1:
+            # Single model case.
+            skl_obj = self.modelObj
+        else:
+            # Multi model case.
+            skl_obj = self.modelObj.iloc[0]["model"]
+
+        ten_row_data = data.head(10).get_values()
+        X = numpy.array(ten_row_data)
+        if label_columns:
+            y = X[:,n_f : n_f + n_c]
+            X = X[:,:n_f]
+            # predict() now takes 'y' also for it to return the labels from script. Skipping 'y'
+            # in local run if passed. Generally, 'y' is passed to return y along with actual output.
+            try:
+                trans_opt = getattr(skl_obj, func_name)(X, y, **kwargs)
+            except TypeError as ex:
+                # Function which does not accept 'y' like predict_proba() raises error like
+                # "predict_proba() takes 2 positional arguments but 3 were given".
+                trans_opt = getattr(skl_obj, func_name)(X, **kwargs)
+        else:
+            trans_opt = getattr(skl_obj, func_name)(X, **kwargs)
+
+        if func_name == "path":
+            raise NotImplementedError(
+                "path() returns tuple of ndarrays of different shapes. Not Implemented yet."
+            )
+
+        # This import is as per scipy version 1.10.x in local machine as teradataml does not
+        # impose restrictions on this package in setup.py. TODO
+        from scipy.sparse import csr_matrix
+
+        if isinstance(trans_opt, csr_matrix):
+            no_of_columns = trans_opt.get_shape()[1]
+            trans_opt = trans_opt.toarray()
+        elif isinstance(trans_opt, dict):
+            raise NotImplementedError(f"Output returns dictionary {trans_opt}. NOT implemented yet.")
+        elif isinstance(trans_opt[0], numpy.ndarray) \
+                or isinstance(trans_opt[0], list) \
+                or isinstance(trans_opt[0], tuple):
+            no_of_columns = len(trans_opt[0])
+        else:
+            no_of_columns = 1
+
+        # Special handling required for cross_decomposition classes's transform function, which
+        # takes label columns also. In this case, output is a tuple of numpy arrays - x_scores and
+        # y_scores. If label columns are not provided, only x_scores are returned.
+        if self.module_name == "sklearn.cross_decomposition" and func_name == "transform":
+            # For cross_decomposition, output is a tuple of arrays when label columns are provided
+            # along with feature columns for transform function. In this case, concatenate the
+            # arrays and return the column names accordingly.
+            if isinstance(trans_opt, tuple): # tuple when label_columns is provided.
+                assert trans_opt[0].shape == trans_opt[1].shape,\
+                    "Output arrays should be of same shape when transform/fit_transform is run "\
+                    "with label columns for cross_decomposition classes.."
+                first_cols = [f"x_scores_{(i + 1)}" for i in range(trans_opt[0].shape[1])]
+                second_cols = [f"y_scores_{(i + 1)}" for i in range(trans_opt[1].shape[1])]
+                no_of_columns = trans_opt[0].shape[1] + trans_opt[1].shape[1]
+                col_names = first_cols + second_cols
+
+                trans_opt = numpy.concatenate(trans_opt, axis=1)
+            else:
+                assert isinstance(trans_opt, numpy.ndarray), "When transform/fit_transform is run "\
+                    "without label columns for cross_decomposition classes, "\
+                    "output should be a numpy array."
+                no_of_columns = trans_opt.shape[1]
+                col_names =[f"x_scores_{(i + 1)}" for i in range(trans_opt.shape[1])]
+        else:
+            # Generate list of new column names.
+            col_names = [f"{self.class_name.lower()}_{func_name}_{(i + 1)}" for i in range(no_of_columns)]
+
+        # Get new column sqlalchemy types for pandas df columns of transform output.
+        opt_pd = pd.DataFrame(trans_opt)
+
+        # If the datatype is not specified then check if the datatype is datetime64 and timezone is present then map it to
+        # TIMESTAMP(timezone=True) else map it according to default value.
+        col_types = [TIMESTAMP(timezone=True)
+                     if pt.is_datetime64_ns_dtype(opt_pd.dtypes[key]) and (opt_pd[col_name].dt.tz is not None)
+                     else _get_sqlalchemy_mapping(str(opt_pd.dtypes[key]))
+                     for key, col_name in enumerate(list(opt_pd.columns))]
+
+        return [(c_name, c_type) for c_name, c_type in zip(col_names, col_types)]
+
+    @_validate_fit_run
+    def _run_function_needing_all_rows(self, X=None, y=None, **kwargs):
+        """
+        Internal function to run functions like score, aic, bic which needs all rows and return
+        one floating number as result.
+        """
+        st_time = time.time()
+
+        assert kwargs["name"], "function name should be passed."
+        func_name = kwargs["name"]
+
+        # Remove 'name' to pass other kwargs to script. TODO: Not passing it now.
+        kwargs.pop("name")
+
+        data, feature_columns, label_columns, _, partition_columns = \
+            self._validate_args_and_get_data(X=X, y=y, groups=None, kwargs=kwargs)
+
+        label_columns = self._get_columns_as_list(label_columns)
+
+        data, new_partition_columns = self._get_data_and_data_partition_columns(data,
+                                                                                feature_columns,
+                                                                                label_columns,
+                                                                                partition_columns)
+
+        file_name = "sklearn_score.py"
+        self._install_script_file(file_identifier=file_name.split(".")[0], file_name=file_name)
+
+        script_file_path = f"{file_name}" if self._is_lake_system \
+            else f"./{self._db_name}/{file_name}"
+
+        partition_indices, partition_types, new_partition_columns = \
+            self._get_partition_col_indices_and_types(data, new_partition_columns)
+
+        self._validate_unique_partition_values(data, new_partition_columns)
+
+        py_exc = UtilFuncs._get_python_execution_path()
+        script_command = f"{py_exc} {script_file_path} {func_name} {len(feature_columns)} "\
+            f"{len(label_columns)} {partition_indices} {partition_types} "\
+            f"{self._model_file_name_prefix} {self._is_lake_system}"
+
+        # score, aic, bic returns float values.
+        return_types = [(col, data._td_column_names_and_sqlalchemy_types[col.lower()])
+                        for col in new_partition_columns] + [(func_name, FLOAT())]
+
+        self._install_initial_model_file()
+
+        opt = self._run_script(data, script_command, new_partition_columns, return_types)
+
+        self._score_execution_time = time.time() - st_time
+
+        if self._is_default_partition_value_fit:
+            # For single model case, partition column is internally generated and
+            # no point in returning it to the user.
+            return opt.select(func_name)
+
+        return opt
+
+    @_validate_fit_run
+    def _transform(self, X=None, y=None, **kwargs):
+        """
+        Internal function to run predict/transform and similar functions, which returns
+        multiple columns. This function will return data row along with the generated
+        columns' row data, unlike sklearn's functions which returns just output data.
+        """
+        st_time = time.time()
+
+        assert kwargs["name"], "function name should be passed."
+        func_name = kwargs["name"]
+
+        # Remove 'name' to pass other kwargs to script. TODO: Not passing it now.
+        kwargs.pop("name")
+
+        data, feature_columns, label_columns, _, partition_columns = \
+            self._validate_args_and_get_data(X=X, y=y, groups=None, kwargs=kwargs)
+
+        data, new_partition_columns = self._get_data_and_data_partition_columns(data,
+                                                                                feature_columns,
+                                                                                label_columns,
+                                                                                partition_columns)
+
+        # Since kwargs are passed to transform, removing additional unrelated arguments from kwargs.
+        if "data" in kwargs:
+            kwargs.pop("data")
+        if "feature_columns" in kwargs:
+            kwargs.pop("feature_columns")
+        if "group_columns" in kwargs:
+            kwargs.pop("group_columns")
+        if "partition_columns" in kwargs:
+            kwargs.pop("partition_columns")
+        if "label_columns" in kwargs:
+            kwargs.pop("label_columns")
+
+        file_name = "sklearn_transform.py"
+        self._install_script_file(file_identifier=file_name.split(".")[0], file_name=file_name)
+
+        script_file_path = f"{file_name}" if self._is_lake_system \
+            else f"./{self._db_name}/{file_name}"
+
+        partition_indices, partition_types, new_partition_columns = \
+            self._get_partition_col_indices_and_types(data, new_partition_columns)
+
+        self._validate_unique_partition_values(data, new_partition_columns)
+
+        py_exc = UtilFuncs._get_python_execution_path()
+        script_command = f"{py_exc} {script_file_path} {func_name} {len(feature_columns)} "\
+            f"{len(label_columns)} {partition_indices} {partition_types} "\
+            f"{self._model_file_name_prefix} {self._is_lake_system}"
+
+        # Returning feature columns also along with transformed columns because we don't know the
+        # mapping of feature columns to the transformed columns.
+        return_types = [(col, data._td_column_names_and_sqlalchemy_types[col.lower()])
+                        for col in (new_partition_columns + feature_columns)]
+        if func_name in ["predict", "decision_function"] and label_columns:
+            return_types += [(col, data._td_column_names_and_sqlalchemy_types[col.lower()])
+                             for col in label_columns]
+        return_types += self._get_return_columns_for_function_(data,
+                                                               feature_columns,
+                                                               label_columns,
+                                                               func_name,
+                                                               len(new_partition_columns),
+                                                               kwargs)
+
+        # Installing model files before running sklearn_transform.py.
+        self._install_initial_model_file()
+
+        opt = self._run_script(data, script_command, new_partition_columns, return_types)
+
+        self._transform_execution_time = time.time() - st_time
+
+        return self._get_returning_df(opt, new_partition_columns, return_types)
+
+    def fit_predict(self, X=None, y=None, **kwargs):
+        """
+        Please check the description in Docs/OpensourceML/sklearn.py.
+        """
+        st_time = time.time()
+
+        self._is_default_partition_value_fit = True # False when the user provides partition columns.
+
+        data, feature_columns, label_columns, _, partition_columns = \
+            self._validate_args_and_get_data(X=X, y=y, groups=None, kwargs=kwargs)
+
+        if partition_columns:
+            self._is_default_partition_value_fit = False
+
+        data, new_partition_columns = self._get_data_and_data_partition_columns(data,
+                                                                                feature_columns,
+                                                                                label_columns,
+                                                                                partition_columns)
+
+        # Return label_columns also if user provides in the function call.
+        return_types = [(col, data._td_column_names_and_sqlalchemy_types[col.lower()])
+                        for col in (new_partition_columns + feature_columns + label_columns)]
+
+        func_name = inspect.stack()[0][3]
+        if label_columns:
+            return_types += self._get_return_columns_for_function_(data,
+                                                                   feature_columns,
+                                                                   label_columns,
+                                                                   func_name,
+                                                                   len(new_partition_columns),
+                                                                   {})
+        else:
+            # If there are no label_columns, we will have only one
+            # predicted column.
+            return_types += [(f"{self.class_name.lower()}_{func_name}_1", FLOAT())]
+
+        file_name = "sklearn_fit_predict.py"
+        self._install_script_file(file_identifier=file_name.split(".")[0], file_name=file_name)
+
+        partition_indices, partition_types, new_partition_columns = \
+            self._get_partition_col_indices_and_types(data, new_partition_columns)
+
+        script_file_name = f"{file_name}" if self._is_lake_system \
+            else f"./{self._db_name}/{file_name}"
+        py_exc = UtilFuncs._get_python_execution_path()
+        script_command = f"{py_exc} {script_file_name} {len(feature_columns)} "\
+            f"{len(label_columns)} {partition_indices} {partition_types} "\
+            f"{self._model_file_name_prefix} {self._is_lake_system}"
+
+        # Get unique values in partitioning columns.
+        self._fit_partition_unique_values = data.drop_duplicate(new_partition_columns).get_values()
+
+        self._install_initial_model_file()
+
+        opt = self._run_script(data, script_command, new_partition_columns, return_types)
+
+        self._fit_predict_execution_time = time.time() - st_time
+
+        if self._is_default_partition_value_fit:
+            # For single model case, partition column is internally generated and no point in
+            # returning it to the user.
+
+            # Extract columns from return types.
+            returning_cols = [col[0] for col in return_types[len(new_partition_columns):]]
+            return opt.select(returning_cols)
+
+        return opt
+
+    def fit_transform(self, X=None, y=None, **kwargs):
+        """
+        Please check the description in Docs/OpensourceML/sklearn.py.
+        """
+        # 'y' is not needed for transform().
+        fit_obj = self.fit(X, y, **kwargs)
+        kwargs["label_columns"] = None
+        return fit_obj.transform(X, None, **kwargs)
+
+    @_validate_fit_run
+    def _run_neighbors(self, X=None, **kwargs):
+        """
+        Internal function to run functions like kneighbors, radius_neighbors, kneighbors_graph,
+        radius_neighbors_graph which returns multiple columns. This function will return data row
+        along with the generated columns' row data, unlike sklearn's functions which returns just
+        output data.
+        """
+        assert kwargs["name"], "function name should be passed."
+        func_name = kwargs["name"]
+        kwargs.pop("name")
+
+        if self.module_name != "sklearn.neighbors":
+            raise AttributeError(f"{self.module_name+'.'+self.class_name} does not have {func_name}() method.")
+
+        data = kwargs.get("data", None)
+        partition_columns = kwargs.get("partition_columns", None)
+
+        if not X and not partition_columns and not data:
+            # If data is not passed, then run from client only.
+            # TODO: decide whether to run from client or from Vantage.
+            opt = super().__getattr__(func_name)(**kwargs)
+            from scipy.sparse.csr import csr_matrix
+            if isinstance(opt, csr_matrix):
+                return opt.toarray()
+            return opt
+
+        self._is_default_partition_value_fit = True # False when the user provides partition columns.
+
+        data, feature_columns, _, _, new_partition_columns = \
+            self._validate_args_and_get_data(X=X, y=None, groups=None, kwargs=kwargs,
+                                             skip_either_or_that=True)
+
+        # Remove the kwargs data.
+        input_data = kwargs.pop("data", None)
+        partition_cols = kwargs.pop("partition_columns", None)
+        feature_cols = kwargs.pop("feature_columns", None)
+        label_cols = kwargs.pop("label_columns", None)
+
+        if partition_columns:
+            # kwargs are passed to kneighbors function. So, removing them from kwargs.
+            kwargs.pop("partition_columns")
+            self._is_default_partition_value_fit = False
+
+        # Generating new partition column name.
+        data, new_partition_columns = self._get_data_and_data_partition_columns(data,
+                                                                                feature_columns,
+                                                                                [],
+                                                                                partition_columns)
+
+        args_str = self._get_kwargs_str(kwargs)
+
+        file_name = "sklearn_neighbors.py"
+        self._install_script_file(file_identifier=file_name.split(".")[0], file_name=file_name)
+
+        script_file_path = f"{file_name}" if self._is_lake_system \
+            else f"./{self._db_name}/{file_name}"
+
+        # Returning feature columns also along with new columns.
+        return_types = [(col, data._td_column_names_and_sqlalchemy_types[col.lower()])
+                        for col in (new_partition_columns + feature_columns)]
+
+        # `return_distance` is needed as the result is a tuple of two arrays when it is True.
+        return_distance = kwargs.get("return_distance", True) # Default value is True.
+
+        # Though new columns return numpy arrays, we are returning them as strings.
+        # TODO: Will update to columns later, if requested later.
+        if func_name in ['kneighbors', 'radius_neighbors']:
+            if return_distance:
+                return_types += [("neigh_dist", VARCHAR())]
+            return_types += [("neigh_ind", VARCHAR())]
+        elif func_name in ['kneighbors_graph', 'radius_neighbors_graph']:
+            return_types += [("A", VARCHAR())]
+        else:
+            return_types += [("output", VARCHAR())]
+
+        partition_indices, partition_types, new_partition_columns = \
+            self._get_partition_col_indices_and_types(data, new_partition_columns)
+
+        py_exc = UtilFuncs._get_python_execution_path()
+        script_command = f"{py_exc} {script_file_path} {func_name} {len(feature_columns)} "\
+            f"{partition_indices} {partition_types} {self._model_file_name_prefix} {self._is_lake_system} "\
+            f"{args_str}"
+
+        # Get unique values in partitioning columns.
+        self._fit_partition_unique_values = data.drop_duplicate(new_partition_columns).get_values()
+
+        self._install_initial_model_file()
+
+        opt = self._run_script(data, script_command, new_partition_columns, return_types)
+
+        return self._get_returning_df(opt, new_partition_columns, return_types)
+
+    def split(self, X=None, y=None, groups=None, **kwargs):
+        """
+        Please check the description in Docs/OpensourceML/sklearn.py.
+        """
+        opt = self._run_model_selection("split", X=X, y=y, groups=groups,
+                                        skip_either_or_that=True, kwargs=kwargs)
+
+        # Get number of splits in the result DataFrame.
+        n_splits = opt.drop_duplicate("split_id").shape[0]
+
+        data = kwargs.get("data", None)
+        feature_columns = kwargs.get("feature_columns", [])
+        label_columns = self._get_columns_as_list(kwargs.get("label_columns", []))
+
+        # If there is not X and y, get feature_columns and label_columns for "data".
+        partition_columns = kwargs.get("partition_columns", [])
+        feature_columns = [col for col in X.columns if col not in partition_columns] \
+            if X and not data and not feature_columns else feature_columns
+        label_columns = y.columns if y and not data and not label_columns else label_columns
+
+        # Return iterator of the train and test dataframes for each split.
+        for i in range(1, n_splits+1):
+            train_df = opt[(opt.split_id == i) & (opt.data_type == "train")]\
+                .select(partition_columns + feature_columns + label_columns)
+            train_df._index_label = None
+            test_df = opt[(opt.split_id == i) & (opt.data_type == "test")]\
+                .select(partition_columns + feature_columns + label_columns)
+            test_df._index_label = None
+
+            yield train_df, test_df
+
+    def get_n_splits(self, X=None, y=None, groups=None, **kwargs):
+        """
+        Please check the description in Docs/OpensourceML/sklearn.py.
+        """
+        return self._run_model_selection("get_n_splits", X=X, y=y, groups=groups,
+                                         skip_either_or_that=True, kwargs=kwargs)
+
+    def _run_model_selection(self,
+                             func_name,
+                             X=None,
+                             y=None,
+                             groups=None,
+                             skip_either_or_that=False,
+                             kwargs={}):
+        """
+        Internal function to run functions like split, get_n_splits of model selection module.
+        - get_n_splits() returns number of splits as value, not as teradataml DataFrame.
+        - split() returns teradataml DataFrame containing train and test data for each split
+          (add partition information if the argument "partition_cols" is provided).
+        """
+        if self.module_name != "sklearn.model_selection":
+            raise AttributeError(f"{self.module_name+'.'+self.class_name} does not "
+                                 f"have {func_name}() method.")
+
+        data = kwargs.get("data", None)
+
+        if not X and not y and not groups and not data:
+            # If data is not passed, then run from client only.
+            # TODO: decide whether to run from client or from Vantage.
+            return super().__getattr__(func_name)()
+
+        self._is_default_partition_value_fit = True # False when the user provides partition columns.
+
+        data, feature_columns, label_columns, group_columns, partition_columns = \
+            self._validate_args_and_get_data(X=X, y=y, groups=groups, kwargs=kwargs,
+                                             skip_either_or_that=skip_either_or_that)
+
+        if partition_columns:
+            self._is_default_partition_value_fit = False
+
+        data, new_partition_columns = self._get_data_and_data_partition_columns(data,
+                                                                                feature_columns,
+                                                                                label_columns,
+                                                                                partition_columns,
+                                                                                group_columns)
+
+        file_name = "sklearn_model_selection_split.py"
+        self._install_script_file(file_identifier=file_name.split(".")[0], file_name=file_name)
+
+        script_file_path = f"{file_name}" if self._is_lake_system \
+            else f"./{self._db_name}/{file_name}"
+
+        if func_name == "split":
+            # Need to generate data into splits of train and test.
+            #   split_id - the column which will be used to identify the split.
+            #   data_type - the column which will be used to identify whether the row is
+            #               train or test row.
+            return_types = [("split_id", INTEGER()), ("data_type", VARCHAR())]
+            # Returning feature columns and label columns as well.
+            return_types += [(col, data._td_column_names_and_sqlalchemy_types[col.lower()])
+                            for col in (feature_columns + label_columns)]
+        else:
+            # Return Varchar by default.
+            # Returns Varchar even for functions like `get_n_splits` which returns large integer
+            # numbers like `4998813702034726525205100` for `LeavePOut` class (when the argument
+            # `p` is 28 and no of data rows is 100) as Vantage cannot scope it to INTEGER.
+            return_types = [(func_name, VARCHAR())]
+
+        return_types = [(col, data._td_column_names_and_sqlalchemy_types[col.lower()])
+                        for col in new_partition_columns] + return_types
+
+        partition_indices, partition_types, new_partition_columns = \
+            self._get_partition_col_indices_and_types(data, new_partition_columns)
+
+        py_exc = UtilFuncs._get_python_execution_path()
+        script_command = f"{py_exc} {script_file_path} {func_name} {len(feature_columns)} "\
+            f"{len(label_columns)} {len(group_columns)} {partition_indices} {partition_types} "\
+            f"{self._model_file_name_prefix} {self._is_lake_system}"
+
+        # Get unique values in partitioning columns.
+        self._fit_partition_unique_values = data.drop_duplicate(new_partition_columns).get_values()
+
+        self._install_initial_model_file()
+
+        opt = self._run_script(data, script_command, new_partition_columns, return_types)
+
+        if func_name == "get_n_splits" and not partition_columns:
+                # Return number of splits as value, not as dataframe.
+                vals = execute_sql("select {} from {}".format(func_name, opt._table_name))
+                opt = vals.fetchall()[0][0]
+
+                # Varchar is returned by the script. Convert it to int.
+                return int(opt)
+
+        return opt
+
+    def _get_returning_df(self, script_df, partition_column, returns):
+        """
+        Internal function to return the teradataml Dataframe except
+        partition_column.
+        """
+        if self._is_default_partition_value_fit:
+            # For single model case, partition column is internally generated
+            # and no point in returning it to the user.
+
+            # Extract columns from return types.
+            returning_cols = [col[0] for col in returns[len(partition_column):]]
+            return script_df.select(returning_cols)
+        return script_df
+
+
+class _SKLearnFunctionWrapper(_GenericObjectWrapper):
+    def __init__(self, module_name, func_name):
+        super().__init__()
+        self.__module_name = module_name
+        self.__func_name = func_name
+        self.__params = None
+        self.__data_args = OrderedDict()
+        self._model_file_name = _generate_new_name(type="file_function", extension="py")
+
+    def __call__(self, **kwargs):
+        """
+        Run the function with all the arguments passed from `td_sklearn.<function_name>` function.
+        """
+        __data_columns = []
+
+        partition_cols = self._get_columns_as_list(kwargs.get("partition_columns", None))
+        if partition_cols:
+            kwargs.pop("partition_columns")
+
+        # Separate dataframe related arguments and their column names from actual kwargs.
+        for k, v in kwargs.items():
+            if isinstance(v, DataFrame):
+                # All dataframes should be select of parent dataframe.
+                _validate_df_query_type(v, "select", k)
+
+                # Save all columns in dataframe related arguments.
+                __data_columns.extend(v.columns)
+
+                self.__data_args[k] = v
+
+
+        # Get common parent dataframe from all dataframes.
+        self.__tdml_df =  DataFrameUtils()._get_common_parent_df_from_dataframes(list(self.__data_args.values()))
+
+        self._validate_existence_of_partition_columns(partition_cols, self.__tdml_df.columns)
+
+        self.__tdml_df = self.__tdml_df.select(__data_columns + partition_cols)
+
+        self.__tdml_df, partition_cols = self._get_data_and_data_partition_columns(self.__tdml_df,
+                                                                                   __data_columns,
+                                                                                   [],
+                                                                                   partition_cols
+                                                                                   )
+
+        # Prepare string of data arguments with name, indices where columns of that argument resides
+        # and types of each of the column.
+        data_args_str = self._prepare_data_args_string(kwargs)
+
+        self.__params = kwargs
+
+        # Get indices and types of partition_columns.
+        idxs, types, partition_cols = self._get_partition_col_indices_and_types(self.__tdml_df,
+                                                                                partition_cols)
+
+        script_file_path = f"{self._model_file_name}" if self._is_lake_system \
+            else f"./{self._db_name}/{self._model_file_name}"
+        py_exc = UtilFuncs._get_python_execution_path()
+        script_command = (f"{py_exc} {script_file_path} {idxs}"
+                          f" ") + \
+            f"{types} {data_args_str}"
+
+        return_types = [(col, self.__tdml_df._td_column_names_and_sqlalchemy_types[col.lower()]) 
+                        for col in partition_cols] + [(self.__func_name, CLOB())]
+
+        # Generate new file in .teradataml directory and install it to Vantage.
+        self._prepare_and_install_file()
+
+        self._model_data = self._run_script(self.__tdml_df, script_command, partition_cols, return_types)
+        self._model_data._index_label = None
+
+        fit_partition_unique_values = self.__tdml_df.drop_duplicate(partition_cols).get_values()
+
+        self.extract_sklearn_obj(n_unique_partitions=len(fit_partition_unique_values),
+                                 n_partition_cols=len(partition_cols))
+
+        # File cleanup after processing.
+        os.remove(self._model_file_local)
+        remove_file(file_identifier=self._model_file_name.split(".")[0], suppress_output=True,
+                    force_remove=True)
+
+        return self.modelObj
+
+    def _prepare_data_args_string(self, kwargs):
+        data_args_str = []
+        for arg_name in list(self.__data_args.keys()):
+            # Remove DataFrame arguments from kwargs, which will be passed to Script.
+            kwargs.pop(arg_name)
+
+            # Get column indices and their types for each dataframe from parent dataframe.
+            _indices, _types, _ = self._get_partition_col_indices_and_types(self.__tdml_df,
+                                                                self.__data_args[arg_name].columns)
+            
+            # Format "<arg_name>-<comma separated indices>-<comma separated types>"
+            data_args_str.append(f"{arg_name}-{_indices}-{_types}")
+        
+        # Format "{<arg_name>-<comma separated indices>-<comma separated types>}--
+        #    {<arg_name>-<comma separated indices>-<comma separated types>}"
+        return "--".join(data_args_str)
+
+    def _validate_existence_of_partition_columns(self, partition_columns, all_columns):
+        """
+        Validate if columns in "partition_columns" argument are present in any of the given
+        dataframes.
+        """
+        invalid_part_cols = [c for c in partition_columns if c not in all_columns]
+
+        if invalid_part_cols:
+            raise ValueError(Messages.get_message(MessageCodes.INVALID_PARTITIONING_COLS,
+                                                  ", ".join(invalid_part_cols),
+                                                  "', '".join(list(self.__data_args.keys())))
+                                                  )
+
+    def _prepare_and_install_file(self):
+        """
+        Prepare function script file from template file and install it in Vaantage.
+        """
+        with open(os.path.join(self._scripts_path, "sklearn_function.template")) as fp:
+            script_data = fp.read()
+        script_data = script_data.replace("<module_name>",self.__module_name).\
+            replace("<func_name>",self.__func_name).replace("<params>", json.dumps(self.__params))
+
+        self._model_file_local = os.path.join(self._tdml_tmp_dir, self._model_file_name)
+
+        with open(self._model_file_local, "w") as fp:
+            fp.write(script_data)
+        
+        self._install_script_file(file_identifier=self._model_file_name.split(".")[0],
+                                  file_name=self._model_file_name,
+                                  file_location=self._tdml_tmp_dir)
+