Wikidata currently has > 1 SPARQL endpoints based on different triple stores (blazegraph, qlever, virtuoso...). QLever is much more performant than the WDQS, but does not include real-time changes (data dump once a week) like the WDQS. In contrast, the WDQS very often has timing errors or rate limiting errors, especially when more complex queries are made. There is currently no ideal solution here and a trade-off must be made between performance and update frequency. This depends on the application scenarios of the users who wish to utilise ToolFindr. The user should therefore be able to configure the desired endpoint for Research Tool Source.
Wikidata currently has > 1 SPARQL endpoints based on different triple stores (blazegraph, qlever, virtuoso...). QLever is much more performant than the WDQS, but does not include real-time changes (data dump once a week) like the WDQS. In contrast, the WDQS very often has timing errors or rate limiting errors, especially when more complex queries are made. There is currently no ideal solution here and a trade-off must be made between performance and update frequency. This depends on the application scenarios of the users who wish to utilise ToolFindr. The user should therefore be able to configure the desired endpoint for Research Tool Source.