Collection of Python scripts to build a Solr index from a subset of the Dutch and English DBpedia dumps for use with our entity linker. The dumps were taken from the 2016-04 Downloads dataset and loaded into a local Virtuoso triple store.
Building the Solr index from the Virtuoso graph is a two-step process:
-
Generating lists of the Dutch and English resource URIs that are to be indexed:
./get_uris.py -
Retreiving the data for the URIs on the lists and sending it to Solr:
./index.py, where the extraction of the relevant fields from the Virtuoso response(s) for each resource takes place inrecord.py.