Pre-processor 0.0.5

csv and json file preprocessor

Preprocessor

Preprocessor is a python library for preprocessing the csv file and flattening the json file

• Preprocess csv file for missing value handling, missing value replacement
• Preprocess csv file having textual column for text preprocessing and word normalization
• Automatically detects the columns data type for csv file and do the preprocessing
• Flatten any level complex json file .

Documentation

Preprocessor Class :

Pre_processor.preprocessor.Preprocessor(file,filetype=None,encoding=None)

Parameters:

- file : str,csv,dict
        File to be preprocessed
- filetype : str
            Type of the input file.Valid options are either dataframe or json
- encoding : str
            encoding scheme for reading file.Default is ISO-8859-1

Methods :

preprocessor.df_preprocessor(threshold_4_delete_null=0.5,no_null_columns=None, numeric_null_replace=None,textual_column_word_tokenize=False,textual_column_word_normalize=False)

Parameters:

- threshold_4_delete_null : float
                    Ratio of the null values to number of rows for columns to be deleted.
- no_null_columns :list
                    List of columns which must not have any null values
- numeric_null_replace : dict 
                    Logic for replacement of null values in numeric column. When None all
                    numeric column's null value will be replaced by mean. Dict format 
                    should be {"mean":[list of column name],"median":[list of 
                    columname],"mode":[list of column names]}
                    In case of giving input as dict format, users need to provide 
                    exaustivelist of column combining all three keys mean,median and mode.

- textual_column_word_tokenize : Boolean
                    Whether tokenization of word needed in case of textual column
- textual_column_word_normalize : str
                    Type of normalization of words needed in Textual columns.Either stem 
                    or lemma for word stemming and word lemmatization respectively.

preprocessor.json_preprocessor()

parameters

-No parameters needed

Code Samples

csv file preprocessing using file path

from Pre_processor.preprocessor import Preprocessor as pps
p = pps(file="example.csv")
data = p.csv_preprocessor(threshold_4_delete_null=0.7,textual_column_word_tokenize=True)