AI-powered synthetic data generation — structured tables, unstructured documents, multi-provider LLM support, referential integrity, and code-gen mode for millions of rows
Command-line program that scans NMDC database for referential integrity violations
Herramienta de línea de comandos para analizar y medir la calidad de la integridad referencial en bases de datos relacionales.
Structured data-cleaning pipeline built on the Olist e-commerce dataset, focusing on reproducible ingestion, schema profiling, and light standardization using Python and pandas. Emphasizes clear pipeline stages, auditability, and portfolio-ready documentation.
Profile a real warehouse, ship only a 156 KB JSON profile, regenerate it elsewhere. Referentially complete synthetic data from a schema plus a profile, using a Gaussian copula for correlation, labelled defect injection and byte-identical output from a seed. 101 SQL integrity checks, worst per-column KS 0.057, zero exact source matches.
Deterministic, referentially-intact synthetic datasets for demos, tests and ML. Zero dependencies.
Deterministic, referentially-intact synthetic datasets for demos, tests and ML. Zero dependencies.
To associate your repository with the referential-integrity topic, visit your repo's landing page and select "manage topics."