Rust · Open data · PostgreSQL · Structured extraction
Ceres + Ares
From fragmented public sources to a reusable catalog and structured data.
Ceres synchronizes open-data metadata; Ares extracts schema-validated information from web pages. They address different stages of collecting useful data.
Updated from public sources ·
Ceres: harvesting comes first
Ceres now documents ten harvest paths, including CSW, STAC, and SDMX, alongside CKAN and DCAT variants. Recent work adds statistical dataflows and fixes partial CSW harvests and oversized SPARQL distributions. Parquet snapshots carry manifests, checksums, coverage reports, and changelogs; embeddings remain optional.
Ares and the downstream work
Ares supports hosted or local extraction with JSON Schema validation, persistent jobs, and operational controls. Ceres feeds a separate Databricks medallion pipeline and a natural-language discovery project built with Agent Bricks. These companion repositories are grouped here rather than presented as unrelated products.