gpf-schema-store

Work in progress (see ROADMAP)

Non official / experimental implementation of OGC API Features - schema by enriching information from the Geoplateforme WFS to ease data discovery by AI (and humans).

It aims to provide a "semantic layer" for the MCP geocontext, pending an equivalent OGC API Features implementation on the Geoplateforme side.

Installation

npm install @ignfab/gpf-schema-store

Usage

Load the catalog

import { getCollectionCatalog } from "@ignfab/gpf-schema-store";

const catalog = getCollectionCatalog();

List collections

[!NOTE] Not in use the MCP geocontext (using an LLM as a search engine is not efficient)

const collections = catalog.list();
console.log("collections:", collections.length);

Long term : should rely on GET /collections from OGC API Feature implementation on the french Géoplateforme

Get the schema of a collection

[!NOTE] In use in the MCP geocontext to provide the "semantic layer" for the collection

// Note : "batiment" = "buildings" in french
const bdtopoBatiment = catalog.getCollectionSchema("BDTOPO_V3:batiment");
console.log(`BDTOPO_V3:batiment schema : ${JSON.stringify(bdtopoBatiment,null,2)}`);

Long term : should rely on GET /collections/{id}/schema from OGC API Feature - schema implementation on the french Géoplateforme.

Search collections by keywords

[!NOTE] In use in the MCP geocontext to search for available collections

// Note "pont" = "bridge" in french
const matches = catalog.search("pont", { limit: 3 });
console.log("top matches:", matches);

Long term: it should rely on a search service on the French Geoplateforme that uses schema information (especially enum values), for example GET /collections/search?q={query} or an improved version of the current Service Géoplateforme de recherche.

Coding

Maintenance and contributor commands are documented in CODING.md.

License

MIT