The Database Test Data Privacy Transformer

 Build a developer tool that transforms production-like database records into safer test datasets while preserving useful structural relationships.

Users can import supported CSV or JSON datasets and define entity relationships, unique keys, sensitive fields, and constraints.

The tool should:

1. Detect likely personal or confidential fields and allow users to confirm classifications.

2. Replace sensitive values using configurable masking, tokenization, or synthetic-data generation strategies.

3. Preserve foreign-key relationships and repeated identities across related tables.

4. Maintain important distributions, formats, null rates, and business-rule constraints where feasible.

5. Generate realistic edge cases such as boundary dates, missing optional values, unusual quantities, and conflicting statuses.

6. Produce a transformation manifest documenting which fields changed and how.

7. Validate that generated records satisfy declared constraints and do not accidentally retain original sensitive values.

Include a before-and-after preview, relationship graph, configurable transformation rules, reproducible generation seed, and downloadable outputs.

Never upload source records to external services by default. Do not claim that masking guarantees anonymization. Warn about re-identification risks, avoid preserving sensitive free-text content without explicit review, and provide tests for referential integrity, uniqueness, and accidental value leakage.