|
| 1 | +''' |
| 2 | +Filename: excel_anonymizer.py |
| 3 | +Author: Siddharth Bhatia |
| 4 | +''' |
| 5 | + |
| 6 | +import argparse |
| 7 | +import logging |
| 8 | +import logging.config |
| 9 | + |
| 10 | +import pandas as pd |
| 11 | +from presidio_analyzer import AnalyzerEngine |
| 12 | +from presidio_anonymizer import AnonymizerEngine |
| 13 | +from presidio_anonymizer.entities.engine import OperatorConfig |
| 14 | +from faker import Faker |
| 15 | + |
| 16 | +def main(): |
| 17 | + """Just a main function needed to publish this to PyPI""" |
| 18 | + |
| 19 | + # Disable loggers from all imported modules |
| 20 | + logging.config.dictConfig({ |
| 21 | + 'version': 1, |
| 22 | + 'disable_existing_loggers': True, |
| 23 | + }) |
| 24 | + |
| 25 | + # Initialize parser |
| 26 | + parser = argparse.ArgumentParser( |
| 27 | + prog='excel_anonymizer.py', |
| 28 | + description='Anonymizes an Excel file and \ |
| 29 | + synthesizes new data in its place.', |
| 30 | + epilog='Made by Siddharth Bhatia') |
| 31 | + |
| 32 | + # Take file as input |
| 33 | + parser.add_argument('filename', help="your excel file here") |
| 34 | + parser.add_argument('-v', '--verbose', |
| 35 | + action='store_true') |
| 36 | + |
| 37 | + # Read arguments from command line |
| 38 | + args = parser.parse_args() |
| 39 | + |
| 40 | + filename = args.filename |
| 41 | + |
| 42 | + if args.verbose is True: |
| 43 | + logging.basicConfig(format="%(message)s", level=logging.INFO) |
| 44 | + logging.info("Verbose output.") |
| 45 | + |
| 46 | + def log(string): |
| 47 | + """Make function for logging.""" |
| 48 | + if args.verbose is True: |
| 49 | + logging.info(string) |
| 50 | + |
| 51 | + df = pd.read_excel(f"{filename}") |
| 52 | + log(df) |
| 53 | + log("") |
| 54 | + |
| 55 | + # Column values to list, which I will use at the end |
| 56 | + columns_ordered_list = df.columns.values.tolist() |
| 57 | + log(f"Columns: {columns_ordered_list}") |
| 58 | + log("") |
| 59 | + |
| 60 | + # Initialize an empty dictionary to store cell locations and values |
| 61 | + cell_data = {} |
| 62 | + |
| 63 | + # Iterate over every cell |
| 64 | + for index, row in df.iterrows(): |
| 65 | + for column in df.columns: |
| 66 | + cell_value = row[column] |
| 67 | + cell_location = (index, column) |
| 68 | + cell_data[cell_location] = cell_value |
| 69 | + |
| 70 | + # log the list of cell values |
| 71 | + log(f"Cell Data: {cell_data}") |
| 72 | + log("") |
| 73 | + log("###") |
| 74 | + |
| 75 | + # Presidio code begins here |
| 76 | + analyzer = AnalyzerEngine() |
| 77 | + anonymizer = AnonymizerEngine() |
| 78 | + |
| 79 | + # Faker code begins here |
| 80 | + fake = Faker() |
| 81 | + |
| 82 | + # Faker Custom Operators |
| 83 | + fake_operators = { |
| 84 | + "PERSON": OperatorConfig("custom", {"lambda": lambda x: fake.name()}), |
| 85 | + "PHONE_NUMBER": OperatorConfig("custom", {"lambda": lambda x: fake.phone_number()}), |
| 86 | + "LOCATION": OperatorConfig("custom", {"lambda": lambda x: str(fake.country())}), |
| 87 | + "EMAIL_ADDRESS": OperatorConfig("custom", {"lambda": lambda x: fake.email()}), |
| 88 | + "DATE_TIME": OperatorConfig("custom", {"lambda": lambda x: str(fake.date_time())}), |
| 89 | + "CREDIT_CARD": OperatorConfig("custom", {"lambda": lambda x: fake.credit_card_number()}), |
| 90 | + "US_BANK_NUMBER": OperatorConfig("custom", {"lambda": lambda x: fake.credit_card_number()}), |
| 91 | + #"DEFAULT": OperatorConfig(operator_name="mask", |
| 92 | + # params={'chars_to_mask': 10, |
| 93 | + # 'masking_char': '*', |
| 94 | + # 'from_end': False}), |
| 95 | + } |
| 96 | + |
| 97 | + fake = Faker(locale="en_IN") |
| 98 | + |
| 99 | + for location, entity in cell_data.items(): |
| 100 | + # log every cell with it's location |
| 101 | + # log(cell, cell_data[cell]) |
| 102 | + log(entity) |
| 103 | + |
| 104 | + # Analyze + anonymize it |
| 105 | + analyzer_results = analyzer.analyze(text=str(entity), language="en") |
| 106 | + log(analyzer_results) |
| 107 | + |
| 108 | + anonymized_results = anonymizer.anonymize( |
| 109 | + text=str(entity), |
| 110 | + analyzer_results=analyzer_results, |
| 111 | + operators=fake_operators, |
| 112 | + ) |
| 113 | + |
| 114 | + log(f"text: {anonymized_results.text}") |
| 115 | + log("") |
| 116 | + # then return it to the dictionary |
| 117 | + cell_data[location] = anonymized_results.text |
| 118 | + log("---") |
| 119 | + |
| 120 | + # log(cell_data) |
| 121 | + # OUTPUT: {(0, 'Name'): '<PERSON>', (0, 'Phone Number'): '<PHONE_NUMBER>', |
| 122 | + # (1, 'Name'): '<PERSON>', (1, 'Phone Number'): '<PHONE_NUMBER>'} |
| 123 | + |
| 124 | + data = {} |
| 125 | + columns = list(set(column for _, column in cell_data)) |
| 126 | + for (index, column), value in cell_data.items(): |
| 127 | + data.setdefault(index, [None] * len(columns)) |
| 128 | + data[index][columns_ordered_list.index(column)] = value |
| 129 | + anonymized_df = pd.DataFrame.from_dict(data, columns=columns_ordered_list, orient="index") |
| 130 | + log(anonymized_df) |
| 131 | + |
| 132 | + filename = filename.rstrip(".xlsx") |
| 133 | + anonymized_df.to_excel( |
| 134 | + f"{filename}-anonymized.xlsx", |
| 135 | + # Don't save the auto-generated numeric index |
| 136 | + index=False |
| 137 | + ) |
| 138 | + |
| 139 | + print(f"Output generated: {filename}-anonymized.xlsx") |
| 140 | + |
| 141 | +if __name__ == "__main__": |
| 142 | + main() |
0 commit comments