This project is not covered by Drupal’s security advisory policy.
Metadata quality audits and human-reviewed AI enrichment proposals for DKAN open data catalogs.
Portions of this module were developed with AI assistance. All code has been reviewed and tested by the maintainer before release. See the README for details on the development and review process.
Two subsystems:
Audit (deterministic, no AI required). Attribute-discovered checks score each dataset 0–100 against DCAT-US/POD quality dimensions: findability, accessibility, interoperability, reusability, contextuality. Per-check weights and thresholds are configurable per site.
Enrichment (LLM, human-in-the-loop). Drafts descriptions, keywords, and Frictionless data dictionaries as proposals a reviewer approves before anything is written to the metastore. Built on the AI module's provider abstraction; works with any configured chat provider.
Status: early development. The audit engine and per-dataset Drush audit (drush dkan-mq:audit ) are working; catalog-wide runs, report UI, enrichment, review UI, and an eval harness are in progress.
Requirements: Drupal 10.5+/11, PHP 8.3+, DKAN ^4.1, dkan_query_tools (bundled with DKAN MCP Server). The AI module (^1.3) plus one provider is only needed for enrichment — audits run without any AI configured.
Project information
- Project categories: Administration tools, Artificial Intelligence (AI), Integrations
- Ecosystem: DKAN, DKAN MCP Server
- Created by dcgoodwin on , updated
This project is not covered by the security advisory policy.
Use at your own risk! It may have publicly disclosed vulnerabilities.
