Redactor

A Python library for text redaction and anonymization, a wrapper built on top of Microsoft Presidio. Super-useful for both individuals and enterprises building custom prompt anonymization layer for their LLM Inference pipelines.

Features

• 🔒 Automatic detection and redaction of sensitive information
• 🔄 Reversible redaction with mapping preservation
• 🎯 Custom entity recognition
• 🛠 Configurable fuzzy matching
• 📝 Support for custom word lists
• 🔧 Extensible recognizer framework

Installation

pip install redactor

After installation, you'll need to download the required spaCy model:

python -m spacy download en_core_web_sm

Quick Start

from redactor import Redactor

# Initialize redactor
redactor = Redactor()

# Redact text
text = "My name is John Doe and my email is john.doe@example.com"
redacted_text, mappings = redactor.redact(text)
print(redacted_text)
# Output: "My name is PERSON_1 and my email is EMAIL_ADDRESS_1"

# Restore original text
original_text = redactor.restore(redacted_text, mappings)
print(original_text)
# Output: "My name is John Doe and my email is john.doe@example.com"

Advanced Usage

Custom Word Lists

# Initialize with custom words to redact
redactor = Redactor(custom_words=["PROJECT-X", "OPERATION-Y"])

text = "Discussing PROJECT-X details"
redacted_text, mappings = redactor.redact(text)
print(redacted_text)
# Output: "Discussing CUSTOM_1 details"

Custom Entity Recognition

from redactor import RecognizerBuilder

# Create custom recognizer for product codes
product_recognizer = (RecognizerBuilder("PRODUCT")
                     .with_pattern(r"Product-[A-Z]+")
                     .with_context(["released", "shipment"])
                     .build())

redactor = Redactor()
redactor.add_recognizer(product_recognizer)

text = "The new Product-ALPHA will be released next month"
redacted_text, mappings = redactor.redact(text)
print(redacted_text)
# Output: "The new PRODUCT_1 will be released next month"

Fuzzy Matching

# Enable fuzzy matching
redactor = Redactor(fuzzy_mapping=1)

text = """
John Smith is the CEO.
Jon Smith signed the document.
"""
redacted_text, mappings = redactor.redact(text)
# Similar names will use the same replacement

Supported Entity Types

Default entities include:

• PERSON
• EMAIL_ADDRESS
• PHONE_NUMBER
• CREDIT_CARD
• DATE_TIME
• LOCATION
• ORGANIZATION

Development

Setup Development Environment

# Clone the repository
git clone https://github.com/xaisr/redactor.git
cd redactor

# Create virtual environment
python -m venv venv
source venv/bin/activate  # On Windows: venv\Scripts\activate

# Install development dependencies
pip install -r requirements-dev.txt

# Install package in editable mode
pip install -e .

Run Tests

pytest

Contributing

Contributions are welcome! Please feel free to submit a Pull Request.

1. Fork the repository
2. Create your feature branch (git checkout -b feature/AmazingFeature)
3. Commit your changes (git commit -m 'Add some AmazingFeature')
4. Push to the branch (git push origin feature/AmazingFeature)
5. Open a Pull Request

License

This project is licensed under the MIT License - see the LICENSE file for details.

Acknowledgments

• Built on top of Microsoft Presidio
• Uses spaCy for NLP tasks