reftb

Reftb, short for Reference Table Builder, is a software framework and data format designed to create, validate, and publish reference tables used in data integration, analytics, and reporting.

Purpose and design: The project aims to provide a stable, language-agnostic approach to defining identifiers, labels, synonyms, and metadata for tabular reference data. It emphasizes versioning, provenance, and interoperability across systems.

Data model and formats: A reftb table consists of rows with fields such as id, label, type, and optional metadata (source, date, confidence). Tables are described by a separate schema file and can be stored in CSV, JSON, or YAML formats; repositories may host catalogues with search and provenance.

Tooling and workflow: The reference toolchain includes a validator, a catalog generator, and a registry client. Users author schemas, validate tables, publish to a registry, and generate human-readable catalogs.

Impact and usage: Reftb is used in data governance, data integration, and research repositories to ensure consistent naming and stable mappings between systems. It supports reproducibility through versioned tables and traceable provenance, making it suitable for cross-system data reconciliation and catalog maintenance.