Changelog

All notable changes to the Protegrity AI Developer Edition project will be documented in this file.

[1.2.0] - 2026-06-30

This release reorganizes the repository around self-contained feature directories, adds a new Anonymization feature, refreshes Data Discovery and Semantic Guardrail, and introduces a guided path for graduating your application from Developer Edition to Team Edition.

🎉 Added

Anonymization (new feature)

• New anonymization/ feature directory with its own docker-compose.yml, README, and an interactive anonymization.ipynb notebook.
• Sample healthcare dataset and a privacy/utility trade-off walkthrough to demonstrate field-level anonymization.

Developer Edition → Team Edition migration

• New "Moving from Developer Edition to Team Edition" guide in the README covering the pty-migrate CLI (check / create-policy), Cloud Protect connection (PTY_CP_HOST), supported auth modes, and the optional ~/.protegrity/config.yaml.
• The same Python and Java sample code runs unchanged against Team Edition — only the endpoint, auth mode, and where the policy lives change.

More interactive notebooks

• Data Discovery: new Jupyter notebooks for text classification, tabular classification, and redaction (data-discovery/samples/jupyter/).
• Data Protection: new getting-started protect/unprotect notebook (data-protection/samples/python/sample-app-protect-unprotect/), runnable directly in Binder.
• Semantic Guardrail and Synthetic Data: refreshed sample notebooks with updated branding and images.

🔄 Changed

Feature-based repository layout

• The repository is reorganized so each feature (data-discovery/, data-protection/, semantic-guardrail/, synthetic-data/, anonymization/) is self-contained with its own docker-compose.yml, samples, data, and README.
• The single root docker-compose.yml is replaced by per-feature compose files — start only the services you need (e.g. cd semantic-guardrail && docker compose up -d).
• Cross-feature assets consolidated under shared/ (config.json, requirements.txt, data/).

Data Discovery 2.0

• Updated to the Data Discovery 2.0 service with improved classification and refreshed entity-to-data-element mappings.

Platform support

• Python: minimum version lowered to 3.11 (from 3.12), supporting 3.11 and above.
• Binder runtime and requirements.txt moved into binder/ and trimmed to the notebook stack.

📚 Documentation

• README restructured for the feature-based layout, with per-feature run instructions for both Python and Java samples.
• Added a Service Health badge linking to the Developer Edition status page.
• Corrected the auth-mode guidance to the actually supported modes (cognito, aws_iam, bearer_token, mtls, none).

⚠️ Migration notes

• If you previously ran docker compose up from the repository root, switch to the per-feature directory (for example cd semantic-guardrail && docker compose up -d). There is no longer a root docker-compose.yml.

[1.1.0] - 2025-12-15

🎉 Major New Features

General Enhancements

• README Improvements: Added badges for improved visibility and quick access to key resources
• Repository Restructuring: Reorganized folders for better organization of samples and source code
• Documentation Updates: Comprehensive updates to getting started guides and feature documentation

Data Discovery v1.1.1

• Structured Text Classification: Added support for structured data classification
• Harmonized Classifications: Introduced categorized "harmonized" entity classifications for consistent data element mapping
• Performance Improvements: General enhancements to classification accuracy and speed
• Enhanced Entity Mapping: Updated entity-to-data-element mapping to align with Discover 1.1

Semantic Guardrails v1.1

• Richer Examples: Included more comprehensive examples in sample files for easier understanding
• Vertical-Specific Models: Added pre-trained support for additional industry verticals (Finance and Healthcare)
• Jupyter Notebook Sample: New interactive notebook for seamless evaluation and execution (samples/python/sample-app-semantic-guardrails/)
• Port Updates: Service now runs on port 8581 with updated image paths

Synthetic Data Generation (NEW)

• Synthetic Data Feature: New capability for generating synthetic test data to support testing and experimentation
• Jupyter Notebook Sample: Interactive notebook for synthetic data generation (samples/python/sample-app-synthetic-data/)
• Docker Compose Profile: New synthetic profile for orchestrating Synthetic Data services
• Service Integration: Seamless integration with existing AI Developer Edition infrastructure

Expanded Language & Platform Support

• Java SDK Samples: Complete Java implementation with CLI scripts for all major workflows
  • Data discovery, classification, protection, and redaction
  • Full source code provided for customization and compilation
  • Cross-platform compatibility (Linux, macOS, Windows)
• Python SDK Updates: Enhanced Python samples with better error handling and documentation
• Dual Language Support: Maintained feature parity between Python and Java implementations
• Java 11+ Compatibility: Ensured compatibility with modern Java versions
• Python 3.12+ Support: Updated minimum Python version requirement

🏗️ Architecture & Structure Changes

Repository Structure Enhancements

• New Java Samples Directory: Added samples/java/ with comprehensive sample applications
  • sample-app-find.sh - PII discovery CLI
  • sample-app-find-and-redact.sh - Discovery and redaction workflow
  • sample-app-find-and-protect.sh - Discovery and protection workflow
  • sample-app-find-and-unprotect.sh - Discovery and unprotection workflow
  • sample-app-protection.sh - Direct protection/unprotection CLI
  • Windows .bat equivalents for all scripts
• Enhanced Python Samples: Updated samples/python/ structure
  • New semantic guardrails Jupyter notebook
  • New synthetic data Jupyter notebook
• Sample Data Organization: Improved organization of configuration files and test data
• Cross-Platform Scripts: Ensured all shell scripts work on Linux, macOS, and Windows

Docker Compose Evolution

• Multi-Profile Support: Enhanced docker-compose.yml with profile-based orchestration
  • Default profile: Classification and Semantic Guardrail services
  • synthetic profile: Adds Synthetic Data generation services
• Service Dependencies: Proper orchestration and startup order management
• Resource Optimization: Improved container download and deployment efficiency

Service Endpoints

• Classification API: http://localhost:8580/pty/data-discovery/v1.1/classify
• Semantic Guardrail API: http://localhost:8581/pty/semantic-guardrail/v1.1/conversations/messages/scan
• Synthetic Data API: New endpoints for synthetic data generation (when using synthetic profile)

🔧 Enhanced Configuration & Service Features

Configuration Updates

• Expanded Entity Mapping: Enhanced config.json with additional entity types
• Simplified Schema: Streamlined configuration keys for easier customization
• Java Configuration Support: Added config.ini format for Java samples

Service Health & Logging

• Improved Health Checks: Enhanced service health verification procedures
• Better Logging: Improved logging options and error messages across all services
• Restart Procedures: Documented comprehensive docker compose management commands

🧑‍💻 Sample Applications Evolution

Java Sample Applications (NEW)

• Complete Java implementation of all Python sample workflows
• Maven-based build system with wrapper scripts
• Fat JAR generation for easy distribution
• Shell and batch scripts for cross-platform execution
• Full source code available for customization

Python Sample Enhancements

• Enhanced semantic guardrails samples with richer examples
• New Jupyter notebooks for interactive exploration
• Improved error handling and user feedback
• Better documentation and inline comments

Jupyter Notebook Integration

• Semantic Guardrails Notebook: Step-by-step guide for conversation scanning and risk assessment
• Synthetic Data Notebook: Interactive guide for generating synthetic test data
• Prerequisites Documentation: Clear instructions for Jupyter Lab setup

🤖 GenAI & AI Integration

Advanced AI Security Features

• Improved Risk Scoring: Enhanced semantic guardrail capabilities for multi-turn conversations
• PII Scanning: Advanced PII detection across conversation history
• Privacy in Conversational AI: Better support for securing LLM interactions
• Prompt Sanitization: Enhanced capabilities for cleaning LLM prompts

📚 Documentation & Developer Experience

Improved Getting Started Guides

• Python Setup: Updated prerequisites and installation instructions
• Java Setup: New comprehensive Java environment setup guide
• Feature Documentation: Detailed documentation for all new features
• Troubleshooting: Enhanced debugging guidance for common issues

Community Support

• Issue Reporting: Clear guidelines for reporting issues with sample scripts
• Log Requirements: Specified log snippet requirements for better issue resolution
• Example Code: More comprehensive code examples across documentation

⚙️ Infrastructure & Operations

Docker Compose Improvements

• Profile-Based Orchestration: Use --profile synthetic to enable synthetic data serv...

[1.1.0] - 2026-01-27

🎉 Major New Features

General Enhancements

• README Improvements: Added badges for improved visibility and quick access to key resources
• Repository Restructuring: Reorganized folders for better organization of samples and source code
• Documentation Updates: Comprehensive updates to getting started guides and feature documentation

Data Discovery v1.1.1

• Structured Text Classification: Added support for structured data classification
• Harmonized Classifications: Introduced categorized "harmonized" entity classifications for consistent data element mapping
• Performance Improvements: General enhancements to classification accuracy and speed
• Enhanced Entity Mapping: Updated entity-to-data-element mapping to align with Discover 1.1

Semantic Guardrails v1.1

• Richer Examples: Included more comprehensive examples in sample files for easier understanding
• Vertical-Specific Models: Added pre-trained support for additional industry verticals (Finance and Healthcare)
• Jupyter Notebook Sample: New interactive notebook for seamless evaluation and execution (samples/python/sample-app-semantic-guardrails/)
• Port Updates: Service now runs on port 8581 with updated image paths

Synthetic Data Generation (NEW)

• Synthetic Data Feature: New capability for generating synthetic test data to support testing and experimentation
• Jupyter Notebook Sample: Interactive notebook for synthetic data generation (samples/python/sample-app-synthetic-data/)
• Docker Compose Profile: New synthetic profile for orchestrating Synthetic Data services
• Service Integration: Seamless integration with existing AI Developer Edition infrastructure

Expanded Language & Platform Support

• Java SDK Samples: Complete Java implementation with CLI scripts for all major workflows
  • Data discovery, classification, protection, and redaction
  • Full source code provided for customization and compilation
  • Cross-platform compatibility (Linux, macOS, Windows)
• Python SDK Updates: Enhanced Python samples with better error handling and documentation
• Dual Language Support: Maintained feature parity between Python and Java implementations
• Java 11+ Compatibility: Ensured compatibility with modern Java versions
• Python 3.12+ Support: Updated minimum Python version requirement

🏗️ Architecture & Structure Changes

Repository Structure Enhancements

• New Java Samples Directory: Added samples/java/ with comprehensive sample applications
  • sample-app-find.sh - PII discovery CLI
  • sample-app-find-and-redact.sh - Discovery and redaction workflow
  • sample-app-find-and-protect.sh - Discovery and protection workflow
  • sample-app-find-and-unprotect.sh - Discovery and unprotection workflow
  • sample-app-protection.sh - Direct protection/unprotection CLI
  • Windows .bat equivalents for all scripts
• Enhanced Python Samples: Updated samples/python/ structure
  • New semantic guardrails Jupyter notebook
  • New synthetic data Jupyter notebook
• Sample Data Organization: Improved organization of configuration files and test data
• Cross-Platform Scripts: Ensured all shell scripts work on Linux, macOS, and Windows

Docker Compose Evolution

• Multi-Profile Support: Enhanced docker-compose.yml with profile-based orchestration
  • Default profile: Classification and Semantic Guardrail services
  • synthetic profile: Adds Synthetic Data generation services
• Service Dependencies: Proper orchestration and startup order management
• Resource Optimization: Improved container download and deployment efficiency

Service Endpoints

• Classification API: http://localhost:8580/pty/data-discovery/v1.1/classify
• Semantic Guardrail API: http://localhost:8581/pty/semantic-guardrail/v1.1/conversations/messages/scan
• Synthetic Data API: New endpoints for synthetic data generation (when using synthetic profile)

🔧 Enhanced Configuration & Service Features

Configuration Updates

• Expanded Entity Mapping: Enhanced config.json with additional entity types
• Simplified Schema: Streamlined configuration keys for easier customization
• Java Configuration Support: Added config.ini format for Java samples

Service Health & Logging

• Improved Health Checks: Enhanced service health verification procedures
• Better Logging: Improved logging options and error messages across all services
• Restart Procedures: Documented comprehensive docker compose management commands

🧑‍💻 Sample Applications Evolution

Java Sample Applications (NEW)

• Complete Java implementation of all Python sample workflows
• Maven-based build system with wrapper scripts
• Fat JAR generation for easy distribution
• Shell and batch scripts for cross-platform execution
• Full source code available for customization

Python Sample Enhancements

• Enhanced semantic guardrails samples with richer examples
• New Jupyter notebooks for interactive exploration
• Improved error handling and user feedback
• Better documentation and inline comments

Jupyter Notebook Integration

• Semantic Guardrails Notebook: Step-by-step guide for conversation scanning and risk assessment
• Synthetic Data Notebook: Interactive guide for generating synthetic test data
• Prerequisites Documentation: Clear instructions for Jupyter Lab setup
• Sample Protection Notebook Documentation: Protect / Unprotect Jupyter Notebook samples

🤖 GenAI & AI Integration

Advanced AI Security Features

• Improved Risk Scoring: Enhanced semantic guardrail capabilities for multi-turn conversations
• PII Scanning: Advanced PII detection across conversation history
• Privacy in Conversational AI: Better support for securing LLM interactions
• Prompt Sanitization: Enhanced capabilities for cleaning LLM prompts

📚 Documentation & Developer Experience

Improved Getting Started Guides

• Python Setup: Updated prerequisites and installation instructions
• Java Setup: New comprehensive Java environment setup guide
• Feature Documentation: Detailed documentation for all new features
• Troubleshooting: Enhanced debugging guidance for common issues

Community Support

• Issue Reporting: Clear guidelines for reporting issues with sample scripts
• Log Requirements: Specified log snippet requirements for better issue resolution
• Example Code: More comprehensive code examples across documentation

⚙️ Infrastructure & Operations

Docker Compose Improvements

• Profile-Based Orchestration: Use --profile synthetic to enable synthetic data services
• Optimized Downloads: Reduced container download times
• Better Resource Management: Improved memory and CPU allocation
• Port Configuration: Flexible port management with environment variable support

🔄 Dependencies

• Updated requirements.txt with latest compatible versions
• Enhanced Maven dependencies for Java samples
• Updated Docker image references to latest stable versions

⚠️ Breaking Changes

None - This release maintains backward compatibility with 1.0.0

Enhanced Data Protection Capabilities

• Protection (Tokenization-like): New protect & unprotect functionality for specific data elements
• Find and Protect: Combined discovery and protection workflow via sample-app-find-and-protect.py
• Direct Protection CLI: New sample-app-protection.py for command-line protect/unprotect operations
• PII Discovery: Enhanced entity enumeration with confidence scores via sample-app-find.py

Semantic Guardrail Integration

• GenAI Security: Message & conversation level risk scoring for AI applications
• Multi-turn Conversation Support: PII scanning across conversation history
• Dual Interface Support: Both cURL and Python examples provided in semantic-guardrail/ folder
• Risk Assessment: Comprehensive risk scoring for GenAI flows