Geospatial Bucketizer

The purpose of the geospatial bucketizer is to apply a geospatial fragmentation to TREE members, based on a property path.

An LDES bucketizer adds triples with the ldes bucket predicate (https://w3id.org/ldes#bucket)[https://w3id.org/ldes#bucket] to the array of quads representating an LDES member, indicating the bucket in which the member belongs.

The fragmentation strategy converts the geospatial string to a GeoJson object, which is used to determine in which tile(s) the LDES member is present and adds these values {zoom}/{x}/{y} to the bucket triples. At the moment only wkt literals are supported. The hypermedia controls are stored in a map. There is a root, which contains relation to one or more intermediate root files, the columns (X). Each of these root files, then contains relation to the tiles that contain the actual LDES members

Example

We assume the following LDES member on which we apply a geospatial fragmentation, using the property path <http://www.w3.org/ns/dcat#bbox> and zoom level set to 4.

<http://example.ord/id/123@456> dct:created "2002-08-13T16:33:18+02:00"^^xsd:dateTime ;
          dct:isVersionOf <http://example.org/id/123> ;
          prov:generatedAtTime "2021-09-07T15:44:05.975Z"^^xsd:dateTime ;
          dcat:bbox "Polygon((-63.37716 45.97597,-71.05923 45.97597,-71.05923 41.65926,-63.37716 41.65926,-63.37716 45.97597))"^^gsp:wktLiteral ;
          rdfs:label "Random place" .

After passing through the bucketizer, the LDES member will have one or more triples:

<http://example.ord/id/123@456> dct:created "2002-08-13T16:33:18+02:00"^^xsd:dateTime ;
          dct:isVersionOf <http://example.org/id/123> ;
          prov:generatedAtTime "2021-09-07T15:44:05.975Z"^^xsd:dateTime ;
          dcat:bbox "Polygon((-63.37716 45.97597,-71.05923 45.97597,-71.05923 41.65926,-63.37716 41.65926,-63.37716 45.97597))"^^gsp:wktLiteral ;
          rdfs:label "Random place" ;
          ldes:bucket "4/0/1", "4/0/2" .

Install

> npm install @treecg/geospatial-bucketizer

Usage

A bucketizer should always be used in combination with the LDES client. More information on how to setup an LDES client can be found here. It is important to set the option in the LDES client to receive the LDES member as an array of quads: representation: 'quads'.

The bucketizer expects a valid property path.

import { GeospatialBucketizer } from '@treecg/geospatial-bucketizer';

const run = async (): Promise<void> => {
  const options = {...};
  const url = ...;
  const zoomLevel = ...;

  const bucketizer = await GeospatialBucketizer.build({propertyPath: '<http://www.w3.org/2000/01/rdf-schema#label>'}, zoomLevel);

  const ldes = LDESClient.createReadStream(url, options);
  ldes.on('data', (member) => {
    bucketizer.bucketize(member.quads, member.id)

    // Continue processing the member, but now the array of quads will have extra triples, the bucket triples
  });
}

run().catch(error => console.error(error.stack));