Advanced Privacy Scrubbing for Screen Parsing and Action Models
OpenSanitizer is a dedicated module designed to detect and scrub PII/PHI data from screen captures and action logs, enhancing privacy across applications like OpenAdapt and OpenAdapter. This tool integrates with deployment frameworks to ensure data protection in automation and screen parsing workflows.
- PII/PHI Detection and Scrubbing: Identify and remove sensitive data from screen and action inputs using state-of-the-art privacy tools.
- Powerful Privacy Stack: OpenSanitizer includes integrations with Microsoft Presidio for identifying and redacting PII entities, Private AI for customizable and GDPR-compliant scrubbing, and AWS Comprehend for natural language processing to enhance data protection.
- Flexible Integration: Supports OpenAdapt and OpenAdapter for end-to-end privacy in screen-based automation.
- Efficient Processing: Streamlined for low latency in high-frequency applications.
- Python 3.10+
- Environment variables for PII/PHI detection settings (e.g.,
DETECTION_MODE,LOG_LEVEL).
- Clone this repository:
git clone https://github.com/OpenAdaptAI/OpenSanitizer.git cd OpenSanitizer - Set up virtual environment and install dependencies:
python3 -m venv venv && source venv/bin/activate pip install -r requirements.txt
Use OpenSanitizer to scrub sensitive data from screen logs before processing:
from opensanitizer import ScreenScrubber
scrubber = ScreenScrubber()
cleaned_data = scrubber.scrub(screenshot_data, action_data)OpenSanitizer complies with privacy standards for PII/PHI handling and is optimized for healthcare and enterprise deployments.
- Expanded Scrubbing Patterns: Add support for custom scrubbing patterns.
- Cloud Integration: Automate PII/PHI scrubbing in cloud deployments for EC2 and containerized models.
- Modular Expansion: Extend compatibility with additional OpenAdapt tools.
OpenSanitizer is licensed under the MIT License.