acdh-oeaw/arche-openrefine

OpenRefine API implementation on top of the ARCHE Suite

0.5.3 2024-09-20 08:02 UTC

This package is auto-updated.

Last update: 2024-11-20 08:19:47 UTC


README

Implementation of the OpenRefine APIs for the ARCHE Suite.

Allows to use the ARCHE Suite as a reconciliation backend for the OpenRefine.

Supported features

  • Reconcile endpoint based on the full text search over the repository content
  • Preview endpoint implemented as a a redirection to the {repositoryApiBaseURL}/{resourceId}/metadata
  • Suggest endpoint for entities, properties and types

Scoring algorithm

  • As match is made against a metadata property value first a property match score is computed.
    • Property weights are defined in the propertyWeights configuration property (see the config-sample.yaml). If property weight isn't provided in the config, it's assumed to be 1.
    • If match is only partial (query phrase is only a part of the property value), the weight is multiplied by the partialMatchCoefficient configuration property.
  • On a resource level the score is the sum of scores for all of its property matches.

Installation

  • Obtain the composer.
  • composer require acdh-oeaw/arche-openrefine
    ln -s vendor/acdh-oeaw/arche-openrefine/index.php index.php
    cp vendor/acdh-oeaw/arche-openrefine/config-sample.yaml config.yaml
    cp vendor/acdh-oeaw/arche-openrefine/.htaccess .htaccess # apache-only
  • Adjust settings in config.yaml

Sample instance

A sample instance which is running on top of the ARCHE repository has its manifest at https://arche.acdh.oeaw.ac.at/openrefine/reconcile.