Allyanonimiser

Added
- Improved package structure for better organization
- Added clearer examples for custom pattern creation
- Enhanced documentation for pattern validation
- Additional utility functions for pattern registry management

Fixed
- Resolved installation issues on certain platforms
- Improved compatibility with different Python versions

Fixed
- Fixed text mangling issue in anonymizer by separating entity collection from replacement
- Improved handling of overlapping entities during anonymization
- Enhanced text replacement algorithm to maintain correct positions

Changed
- Optimized pattern detection performance for long documents
- Enhanced accuracy of Australian TFN and Medicare patterns
- Improved handling of edge cases in insurance pattern detection
- Bug fixes in enhanced_analyzer.py for certain pattern matches

Fixed
- Fixed performance issues when processing large claim notes
- Resolved pattern matching inconsistencies in insurance_patterns.py
- Fixed detection issues with Australian formats in au_patterns.py
- Improved compatibility with different spaCy model versions

Added
- Greatly expanded pattern sets for more comprehensive PII detection
- Enhanced Australian-specific patterns for TFNs, Medicare numbers, phone numbers, addresses
- Added insurance-specific patterns for policy numbers, claim references, and vehicle identifiers
- Improved detection of general patterns such as person names, email addresses, dates, monetary amounts
- Fixed detection rates for common PII in example claim notes

Changed
- Updated README badges to point to correct GitHub repository
- Improved pattern matching precision for Australian contexts
- Enhanced regex patterns for more accurate entity detection

Added
- Functional implementation of EnhancedAnalyzer with regex-based PII detection
- Added RecognizerResult class to represent detected PII entities
- Implemented anonymization functionality with multiple operator types (replace, mask, redact, hash)
- Enhanced pattern files with comprehensive patterns for Australian PII, insurance information, and general entities
- Added text segmentation and PII detection scoring in long_text_processor
- Implemented the process method in Allyanonimiser for one-step analysis and anonymization

Changed
- Updated README with detailed description of new features
- Fixed badge URLs in README to point to correct GitHub repository
- Made minimal processing work without relying on external packages
- Improved documentation with examples of using the new functionality

Added
- Enhanced factory functions to accept custom patterns parameter
- Improved flexibility in function parameters with better default values
- More robust pattern creation from examples

Changed
- Made API more flexible and customizable
- Improved handling of different pattern types
- Further enhanced code structure to avoid import issues