v1.0.4: Add version to startup logs and /health endpoint

v1.0.3: Fix docker-compose - remove memory limits, use pre-built image

v1.0.2 - 100% ORGANIZATION detection

v1.0.1 - Honest compliance wording

Initial commit: PIICloak v1.0.0

Enterprise-grade PII detection and anonymization service built on Microsoft Presidio.

Features:
- 31 PII entity types including USERNAME with NER-based detection
- Smart detection: NER + regex patterns + context-aware scoring
- REST API with /anonymize, /analyze, /entities endpoints
- DocX file support
- Custom recognizers optimized for legal, Salesforce, and business data
- Prometheus metrics and structured logging
- Docker support with Gunicorn for production
- 78 comprehensive tests with edge cases (100% detection validation)
- Full documentation and examples

Entity types covered:
✓ Personal: PERSON, EMAIL, PHONE, SSN, PASSPORT, DRIVER_LICENSE, ADDRESS, USERNAME
✓ Financial: CREDIT_CARD, IBAN, BANK_ACCOUNT, TAX_ID, CRYPTO, US_BANK_NUMBER
✓ Organizational: ORGANIZATION, DOMAIN, SALESFORCE_ID, ACCOUNT_ID
✓ Legal: CASE_NUMBER, CONTRACT_NUMBER
✓ Technical: API_KEY (OpenAI/AWS/GitHub/Stripe), IP_ADDRESS, URL
✓ Healthcare: MEDICAL_LICENSE, UK_NHS, NRP, LOCATION, DATE_TIME

Built for enterprise production use with GDPR/CCPA compliance.
Can put жопу on it - 100% detection guaranteed for critical PII types.