Advanced Text Anonymizer (afanon) 🕵️‍♂️📄

A powerful and configurable Python script (afanon.py) for identifying and anonymizing sensitive information within text documents. This tool employs a multi-layered approach, combining Named Entity Recognition (NER), regular expressions, custom wordlists, and fallback heuristics to provide comprehensive PII redaction.

Key Features

• Multi-Layered Detection:
  • Named Entity Recognition (NER): Leverages NLTK to identify entities like PERSON, ORGANIZATION, LOCATION, DATE, MONEY, etc. (configurable).
  • Regular Expression (Regex) Matching: Pre-defined and customizable regex patterns to detect common PII like emails, phone numbers, SSNs, credit card numbers, IP addresses, physical addresses, and postal codes.
  • Custom Wordlist Anonymization: Allows users to supply their own lists of sensitive words/phrases (e.g., passwords, specific object names, dog breeds, project codenames) for targeted redaction.
  • Fallback Capitalized Word Detection: A heuristic to catch potential names that might be missed by other methods.
• Flexible Anonymization Strategies:
  • placeholder: Replaces sensitive data with generic placeholders (e.g., [PERSON], [EMAIL]).
  • counter: Replaces with unique, numbered placeholders (e.g., [PERSON_1], [PERSON_2]).
  • fake: Substitutes PII with realistic-looking but synthetic data generated by the Faker library (e.g., "John Doe" for a name, a fake address for a real one).
• Intelligent Overlap Resolution: When multiple detection methods identify overlapping text, the script prioritizes matches based on specificity (regex/custom lists > NER > fallback) and length to ensure accurate anonymization.
• Highly Configurable:
  • Specify which NER entities to target.
  • Provide custom files for passwords, objects, dog breeds, or a generic wordlist.
  • Choose the anonymization strategy.
  • (Advanced) Modify built-in regex patterns directly in the code.
• Command-Line Interface (CLI): Easy to integrate into workflows for processing text files.
• Detailed Logging: Offers insights into the anonymization process, including found entities and applied replacements (with verbose/debug mode).
• Automatic NLTK Data Check: Ensures necessary NLTK resources are available and attempts to download them if missing.

Getting Started

Prerequisites

• Python 3.7+
• Required Python packages (see requirements.txt):
  • nltk
  • Faker

Installation

1. Clone the repository:

   git clone https://github.com/your-username/advanced-text-anonymizer.git
   cd advanced-text-anonymizer

2. Install dependencies:

   pip install -r requirements.txt

3. NLTK Data: The script (afanon.py) will automatically check and attempt to download required NLTK data packages (punkt, averaged_perceptron_tagger, maxent_ne_chunker, words) on its first run. Ensure you have an internet connection.

Usage

The script is run from the command line:

python afanon.py <input_file> <output_file> [options]