HIPAA-Compliant Patient Record Anonymization API Middleware

Design a Flask-based microservice that automatically anonymizes patient records using advanced data masking techniques while maintaining referential integrity. Implement robust encryption for PHI (Protected Health Information) using Python's cryptography library, ensuring HIPAA compliance. The middleware should support automatic tokenization of sensitive fields like SSN, names, and medical record numbers, with configurable anonymization levels (full, partial, hashed). Include comprehensive logging, error handling for edge cases, and a secure authentication mechanism using JWT tokens. Provide a clear strategy for handling different data types (string, numeric, datetime) with reversible anonymization options.