Piiranha Server

PII Detection and Masking Service

A FastAPI-based service for detecting and masking Personally Identifiable Information (PII) in text using the PII-Ranha model.

Features

• PII Detection: Check if text contains any PII
• PII Masking: Mask detected PII with either aggregated or detailed redaction
• Health Check: Verify service status
• Token Chunking: Handles long texts by splitting into 256-token chunks

Examples

% curl -s  -X POST "http://localhost:8001/check-pii" \
-H "Content-Type: application/json" \
-d '{"text":"Your text here, my name is Jean-Claude Dusse"}' 
{
  "detail": "PII detected in input"
}
(piiserver)

% curl -s  -X POST "http://localhost:8001/check-pii" \
-H "Content-Type: application/json" \
-d '{"text":"Your text here, Lorem ipsum dolor sit amet, consectetur adipiscing elit"}' 
{
  "status": "OK",
  "message": "No PII detected"
}

API Endpoints

Check PII

POST /check-pii

• Request body: {"text": "your text here"}
• Returns: 200 if no PII detected, 400 if PII detected

Mask PII

POST /mask-pii

• Request body: {"text": "your text here"}
• Optional query param: aggregate_redaction=true|false (default: true)
• Returns: Original and masked text

Health Check

GET /health

• Returns: Service status

Running the Service

1. Install dependencies:

pip install fastapi uvicorn transformers torch

2. Run the server:

python app.py

3. The service will be available at http://localhost:8000

Testing with curl

Check for PII:

curl -X POST "http://localhost:8000/check-pii" -H "Content-Type: application/json" -d '{"text":"Your text here"}'

Mask PII:

curl -X POST "http://localhost:8000/mask-pii" -H "Content-Type: application/json" -d '{"text":"Your text here"}'

Development

The service uses the PII-Ranha model from Hugging Face.

Author & Enterprise Support

Author: Sébastien Campion

Sébastien Campion - sebastien.campion@foss4.eu

Note: This project is under active development. Please report any issues or feature requests through the issue tracker.

License

This project is licensed under the GNU Affero General Public License v3.0 - see the LICENSE file for details.