teradataml 17.20.0.6__py3-none-any.whl → 20.0.0.0__py3-none-any.whl

teradataml/analytics/mle/NERTrainer.py

@@ -1,458 +0,0 @@
-#!/usr/bin/python
-# ##################################################################
-#
-# Copyright 2018 Teradata. All rights reserved.
-# TERADATA CONFIDENTIAL AND TRADE SECRET
-#
-# Primary Owner: Rohit Agrawal (rohit.agrawal@teradata.com)
-# Secondary Owner: Pankaj Purandare (pankajvinod.purandare@teradata.com)
-#
-# Version: 1.2
-# Function Version: 1.7
-#
-# ##################################################################
-
-import inspect
-import time
-from teradataml.common.wrapper_utils import AnalyticsWrapperUtils
-from teradataml.common.utils import UtilFuncs
-from teradataml.context.context import *
-from teradataml.dataframe.dataframe import DataFrame
-from teradataml.common.aed_utils import AedUtils
-from teradataml.analytics.analytic_query_generator import AnalyticQueryGenerator
-from teradataml.common.exceptions import TeradataMlException
-from teradataml.common.messages import Messages
-from teradataml.common.messagecodes import MessageCodes
-from teradataml.common.constants import TeradataConstants
-from teradataml.dataframe.dataframe_utils import DataFrameUtils as df_utils
-from teradataml.options.display import display
-
-class NERTrainer:
-    
-    def __init__(self,
-                 data=None,
-                 text_coloumn=None,
-                 extractor_jar=None,
-                 feature_template=None,
-                 model_file=None,
-                 language="en",
-                 max_iter_num=1000,
-                 eta=1.0E-4,
-                 min_occur_num=0,
-                 data_sequence_column=None):
-        """
-        DESCRIPTION:
-            The NERTrainer function takes training data and outputs a CRF model
-            (a binary file) that can be specified in the function NERExtractor
-            and NEREvaluator.
-         
-         
-        PARAMETERS:
-            data:
-                Required Argument.
-                Specifies an input teradataml DataFrame containing training data.
-         
-            text_coloumn:
-                Required Argument.
-                Specifies the name of the input teradataml DataFrame column that
-                contains the text to analyze.
-                Types: str
-         
-            extractor_jar:
-                Optional Argument.
-                Specifies the name of the JAR file that contains the Java classes
-                that extract features. The function includes the predefined extractor
-                classes described in the file provided in the argument feature_template.
-                Note:
-                    1. The name of the JAR file is case-sensitive.
-                    2. The ML Engine does not support the creation of new extractor classes.
-                       However, it does support existing JAR files—for installation instructions,
-                       see Teradata Vantage User Guide.
-                Types: str
-         
-            feature_template:
-                Required Argument.
-                Specifies the name of the file that specifies how to generate
-                features when training the model. This file is pre-installed
-                in ML Engine under the name "template_1.txt".
-                Types: str
-         
-            model_file:
-                Required Argument.
-                Specifies the name of the model file that is generated and installed
-                in the ML Engine by the function.
-                Types: str
-         
-            language:
-                Optional Argument.
-                Specifies the language of the input text:
-                    * en - English
-                    * zh_CN - Simplified Chinese
-                    * zh_TW - Traditional Chinese
-                Default Value: "en"
-                Permitted Values: en, zh_CN, zh_TW
-                Types: str
-         
-            max_iter_num:
-                Optional Argument.
-                Specifies the maximum number of iterations.
-                Types: int
-         
-            eta:
-                Optional Argument.
-                Specifies the tolerance of the termination criterion. Defines the
-                differences of the values of the loss function between two sequential
-                epochs. When training a model, the function performs n-times
-                iterations. At the end of each epoch, the function calculates the
-                loss or cost function on the training samples. If the loss function
-                value change is very small between two sequential epochs, the
-                function considers the training process to have converged.
-                The function defines eta as:
-                    Eta=(f(n)-f(n-1))/f(n-1), where f(n) is the loss function value of the nth epoch.
-                Default Value: 1.0E-4
-                Types: float
-         
-            min_occur_num:
-                Optional Argument.
-                Specifies the minimum number of times that a feature must occur in the
-                input text before the function uses the feature to construct the
-                model.
-                Default Value: 0
-                Types: int
-         
-            data_sequence_column:
-                Optional Argument.
-                Specifies the list of column(s) that uniquely identifies each row of
-                the input argument "data". The argument is used to ensure
-                deterministic results for functions which produce results that vary
-                from run to run.
-                Types: str OR list of Strings (str)
-         
-        RETURNS:
-            Instance of NERTrainer.
-            Output teradataml DataFrames can be accessed using attribute
-            references, such as NERTrainerObj.<attribute_name>.
-            Output teradataml DataFrame attribute names are:
-                result
-         
-         
-        RAISES:
-            TeradataMlException
-         
-         
-        EXAMPLES:
-            # Load the data to run the example
-            load_example_data("nertrainer","ner_sports_train")
-         
-            # Create teradataml DataFrame object.
-            ner_sports_train = DataFrame.from_table("ner_sports_train")
-         
-            # Run the NERTrain function to generated a trained model file which is used in NERExtractor or NEREvaluator
-            nertrainer_train = NERTrainer(data=ner_sports_train,
-                                          text_coloumn='content',
-                                          model_file='ner_model.bin',
-                                          feature_template='template_1.txt',
-                                          language='en',
-                                          eta=0.0001,
-                                          max_iter_num=1000,
-                                          min_occur_num=0,
-                                          extractor_jar=' ')
-         
-            # Print the result DataFrame
-            print(nertrainer_train.result)
-        
-        """
-        
-        # Start the timer to get the build time
-        _start_time = time.time()
-        
-        self.data  = data 
-        self.text_coloumn  = text_coloumn 
-        self.extractor_jar  = extractor_jar 
-        self.feature_template  = feature_template 
-        self.model_file  = model_file 
-        self.language  = language 
-        self.max_iter_num  = max_iter_num 
-        self.eta  = eta 
-        self.min_occur_num  = min_occur_num 
-        self.data_sequence_column  = data_sequence_column 
-        
-        # Create TeradataPyWrapperUtils instance which contains validation functions.
-        self.__awu = AnalyticsWrapperUtils()
-        self.__aed_utils = AedUtils()
-        
-        # Create argument information matrix to do parameter checking
-        self.__arg_info_matrix = []
-        self.__arg_info_matrix.append(["data", self.data, False, (DataFrame)])
-        self.__arg_info_matrix.append(["text_coloumn", self.text_coloumn, False, (str)])
-        self.__arg_info_matrix.append(["extractor_jar", self.extractor_jar, True, (str)])
-        self.__arg_info_matrix.append(["feature_template", self.feature_template, False, (str)])
-        self.__arg_info_matrix.append(["model_file", self.model_file, False, (str)])
-        self.__arg_info_matrix.append(["language", self.language, True, (str)])
-        self.__arg_info_matrix.append(["max_iter_num", self.max_iter_num, True, (int)])
-        self.__arg_info_matrix.append(["eta", self.eta, True, (float)])
-        self.__arg_info_matrix.append(["min_occur_num", self.min_occur_num, True, (int)])
-        self.__arg_info_matrix.append(["data_sequence_column", self.data_sequence_column, True, (str,list)])
-        
-        if inspect.stack()[1][3] != '_from_model_catalog':
-            # Perform the function validations
-            self.__validate()
-            # Generate the ML query
-            self.__form_tdml_query()
-            # Execute ML query
-            self.__execute()
-            # Get the prediction type
-            self._prediction_type = self.__awu._get_function_prediction_type(self)
-        
-        # End the timer to get the build time
-        _end_time = time.time()
-        
-        # Calculate the build time
-        self._build_time = (int)(_end_time - _start_time)
-        
-    def __validate(self):
-        """
-        Function to validate sqlmr function arguments, which verifies missing 
-        arguments, input argument and table types. Also processes the 
-        argument values.
-        """
-        
-        # Make sure that a non-NULL value has been supplied for all mandatory arguments
-        self.__awu._validate_missing_required_arguments(self.__arg_info_matrix)
-        
-        # Make sure that a non-NULL value has been supplied correct type of argument
-        self.__awu._validate_argument_types(self.__arg_info_matrix)
-        
-        # Check to make sure input table types are strings or data frame objects or of valid type.
-        self.__awu._validate_input_table_datatype(self.data, "data", None)
-        
-        # Check for permitted values
-        language_permitted_values = ["EN", "ZH_CN", "ZH_TW"]
-        self.__awu._validate_permitted_values(self.language, language_permitted_values, "language")
-        
-        # Check whether the input columns passed to the argument are not empty.
-        # Also check whether the input columns passed to the argument valid or not.
-        self.__awu._validate_input_columns_not_empty(self.text_coloumn, "text_coloumn")
-        self.__awu._validate_dataframe_has_argument_columns(self.text_coloumn, "text_coloumn", self.data, "data", False)
-        
-        self.__awu._validate_input_columns_not_empty(self.data_sequence_column, "data_sequence_column")
-        self.__awu._validate_dataframe_has_argument_columns(self.data_sequence_column, "data_sequence_column", self.data, "data", False)
-        
-        
-    def __form_tdml_query(self):
-        """
-        Function to generate the analytical function queries. The function defines 
-        variables and list of arguments required to form the query.
-        """
-        
-        # Output table arguments list
-        self.__func_output_args_sql_names = []
-        self.__func_output_args = []
-        
-        # Model Cataloging related attributes.
-        self._sql_specific_attributes = {}
-        self._sql_formula_attribute_mapper = {}
-        self._target_column = None
-        self._algorithm_name = None
-        
-        # Generate lists for rest of the function arguments
-        self.__func_other_arg_sql_names = []
-        self.__func_other_args = []
-        self.__func_other_arg_json_datatypes = []
-        
-        self.__func_other_arg_sql_names.append("TextColumn")
-        self.__func_other_args.append(UtilFuncs._teradata_collapse_arglist(UtilFuncs._teradata_quote_arg(self.text_coloumn, "\""), "'"))
-        self.__func_other_arg_json_datatypes.append("COLUMNS")
-
-        self.__func_other_arg_sql_names.append("ModelFileName")
-        self.__func_other_args.append(UtilFuncs._teradata_collapse_arglist(self.model_file, "'"))
-        self.__func_other_arg_json_datatypes.append("STRING")
-        
-        self.__func_other_arg_sql_names.append("FeatureTemplate")
-        self.__func_other_args.append(UtilFuncs._teradata_collapse_arglist(self.feature_template, "'"))
-        self.__func_other_arg_json_datatypes.append("STRING")
-        
-        if self.language is not None and self.language != "en":
-            self.__func_other_arg_sql_names.append("InputLanguage")
-            self.__func_other_args.append(UtilFuncs._teradata_collapse_arglist(self.language, "'"))
-            self.__func_other_arg_json_datatypes.append("STRING")
-        
-        if self.eta is not None and self.eta != 1.0E-4:
-            self.__func_other_arg_sql_names.append("Eta")
-            self.__func_other_args.append(UtilFuncs._teradata_collapse_arglist(self.eta, "'"))
-            self.__func_other_arg_json_datatypes.append("DOUBLE")
-        
-        if self.max_iter_num is not None and self.max_iter_num != 1000:
-            self.__func_other_arg_sql_names.append("MaxIterNum")
-            self.__func_other_args.append(UtilFuncs._teradata_collapse_arglist(self.max_iter_num, "'"))
-            self.__func_other_arg_json_datatypes.append("INTEGER")
-        
-        if self.min_occur_num is not None and self.min_occur_num != 0:
-            self.__func_other_arg_sql_names.append("MinOccurNum")
-            self.__func_other_args.append(UtilFuncs._teradata_collapse_arglist(self.min_occur_num, "'"))
-            self.__func_other_arg_json_datatypes.append("INTEGER")
-        
-        if self.extractor_jar is not None:
-            self.__func_other_arg_sql_names.append("ExtractorJar")
-            self.__func_other_args.append(UtilFuncs._teradata_collapse_arglist(self.extractor_jar, "'"))
-            self.__func_other_arg_json_datatypes.append("STRING")
-        
-        # Generate lists for rest of the function arguments
-        sequence_input_by_list = []
-        if self.data_sequence_column is not None:
-            sequence_input_by_list.append("input:" + UtilFuncs._teradata_collapse_arglist(self.data_sequence_column, ""))
-        
-        if len(sequence_input_by_list) > 0:
-            self.__func_other_arg_sql_names.append("SequenceInputBy")
-            sequence_input_by_arg_value = UtilFuncs._teradata_collapse_arglist(sequence_input_by_list, "'")
-            self.__func_other_args.append(sequence_input_by_arg_value)
-            self.__func_other_arg_json_datatypes.append("STRING")
-            self._sql_specific_attributes["SequenceInputBy"] = sequence_input_by_arg_value
-        
-        
-        # Declare empty lists to hold input table information.
-        self.__func_input_arg_sql_names = []
-        self.__func_input_table_view_query = []
-        self.__func_input_dataframe_type = []
-        self.__func_input_distribution = []
-        self.__func_input_partition_by_cols = []
-        self.__func_input_order_by_cols = []
-        
-        # Process data
-        self.__table_ref = self.__awu._teradata_on_clause_from_dataframe(self.data, False)
-        self.__func_input_distribution.append("FACT")
-        self.__func_input_arg_sql_names.append("input")
-        self.__func_input_table_view_query.append(self.__table_ref["ref"])
-        self.__func_input_dataframe_type.append(self.__table_ref["ref_type"])
-        self.__func_input_partition_by_cols.append("1")
-        self.__func_input_order_by_cols.append("NA_character_")
-        
-        function_name = "NERTrainer"
-        # Create instance to generate SQLMR.
-        self.__aqg_obj = AnalyticQueryGenerator(function_name, 
-                self.__func_input_arg_sql_names, 
-                self.__func_input_table_view_query, 
-                self.__func_input_dataframe_type, 
-                self.__func_input_distribution, 
-                self.__func_input_partition_by_cols, 
-                self.__func_input_order_by_cols, 
-                self.__func_other_arg_sql_names, 
-                self.__func_other_args, 
-                self.__func_other_arg_json_datatypes, 
-                self.__func_output_args_sql_names, 
-                self.__func_output_args, 
-                engine="ENGINE_ML")
-        # Invoke call to SQL-MR generation.
-        self.sqlmr_query = self.__aqg_obj._gen_sqlmr_select_stmt_sql()
-        
-        # Print SQL-MR query if requested to do so.
-        if display.print_sqlmr_query:
-            print(self.sqlmr_query)
-        
-        # Set the algorithm name for Model Cataloging.
-        self._algorithm_name = self.__aqg_obj._get_alias_name_for_function(function_name)
-        
-    def __execute(self):
-        """
-        Function to execute SQL-MR queries. 
-        Create DataFrames for the required SQL-MR outputs.
-        """
-        # Generate STDOUT table name and add it to the output table list.
-        sqlmr_stdout_temp_tablename = UtilFuncs._generate_temp_table_name(prefix="td_sqlmr_out_", use_default_database=True, gc_on_quit=True, quote=False, table_type=TeradataConstants.TERADATA_TABLE)
-        try:
-            # Generate the output.
-            UtilFuncs._create_table(sqlmr_stdout_temp_tablename, self.sqlmr_query)
-        except Exception as emsg:
-            raise TeradataMlException(Messages.get_message(MessageCodes.TDMLDF_EXEC_SQL_FAILED, str(emsg)), MessageCodes.TDMLDF_EXEC_SQL_FAILED)
-        
-        # Update output table data frames.
-        self._mlresults = []
-        self.result = self.__awu._create_data_set_object(df_input=UtilFuncs._extract_table_name(sqlmr_stdout_temp_tablename), source_type="table", database_name=UtilFuncs._extract_db_name(sqlmr_stdout_temp_tablename))
-        self._mlresults.append(self.result)
-        
-    def show_query(self):
-        """
-        Function to return the underlying SQL query.
-        When model object is created using retrieve_model(), then None is returned.
-        """
-        return self.sqlmr_query
-        
-    def get_prediction_type(self):
-        """
-        Function to return the Prediction type of the algorithm.
-        When model object is created using retrieve_model(), then the value returned is
-        as saved in the Model Catalog.
-        """
-        return self._prediction_type
-        
-    def get_target_column(self):
-        """
-        Function to return the Target Column of the algorithm.
-        When model object is created using retrieve_model(), then the value returned is
-        as saved in the Model Catalog.
-        """
-        return self._target_column
-        
-    def get_build_time(self):
-        """
-        Function to return the build time of the algorithm in seconds.
-        When model object is created using retrieve_model(), then the value returned is
-        as saved in the Model Catalog.
-        """
-        return self._build_time
-        
-    def _get_algorithm_name(self):
-        """
-        Function to return the name of the algorithm.
-        """
-        return self._algorithm_name
-        
-    def _get_sql_specific_attributes(self):
-        """
-        Function to return the dictionary containing the SQL specific attributes of the algorithm.
-        """
-        return self._sql_specific_attributes
-        
-    @classmethod
-    def _from_model_catalog(cls,
-        result = None,
-        **kwargs):
-        """
-        Classmethod is used by Model Cataloging, to instantiate this wrapper class.
-        """
-        kwargs.pop("result", None)
-        
-        # Model Cataloging related attributes.
-        target_column = kwargs.pop("__target_column", None)
-        prediction_type = kwargs.pop("__prediction_type", None)
-        algorithm_name = kwargs.pop("__algorithm_name", None)
-        build_time = kwargs.pop("__build_time", None)
-        
-        # Let's create an object of this class.
-        obj = cls(**kwargs)
-        obj.result  = result 
-        
-        # Initialize the sqlmr_query class attribute.
-        obj.sqlmr_query = None
-        
-        # Initialize the SQL specific Model Cataloging attributes.
-        obj._sql_specific_attributes = None
-        obj._target_column = target_column
-        obj._prediction_type = prediction_type
-        obj._algorithm_name = algorithm_name
-        obj._build_time = build_time
-        
-        # Update output table data frames.
-        obj._mlresults = []
-        obj.result = obj.__awu._create_data_set_object(df_input=UtilFuncs._extract_table_name(obj.result), source_type="table", database_name=UtilFuncs._extract_db_name(obj.result))
-        obj._mlresults.append(obj.result)
-        return obj
-
-    def __repr__(self):
-        """
-        Returns the string representation for a NERTrainer class instance.
-        """
-        repr_string = "############ STDOUT Output ############"
-        repr_string = "{}\n\n{}".format(repr_string, self.result)
-        return repr_string
-