Developing an AutoComplete Plugin for ORR

[lewismc]

With the aim of extending/encouraging use of COR, I’m interested in implementing autocomplete plugins for various frameworks e.g. JQuery (as an example), which would enable easy autocomplete functionality for query bars based upon terms present in COR.

Related code for this can be located here however ideally we would want to make it asynchronous such that every character entry results in a refined list of terms.

[lewismc]

A configured plugin would need to be produced for each deployed instance of ORR.

One place we could prototype this would be in the query bar available at http://cor.esipfed.org/ont#/st/

[carueda]

This has been a very desirable feature for a long time. As you say, realized in the form of client modules that can be used in common frameworks. Of course, besides UX features on the client side, this will also demand improved backend performance especially as the triple store grows.

(btw, although not exactly "autocomplete", but with some similar features, this reminds me of our CF Browser, https://mmisw.org/cfsn, which allows searches by literal input, glob, and regex, see for example, https://mmisw.org/cfsn/#/search/!g/*temperature*cloud*top)

[graybeal]

Question: Are you thinking of autocomplete for a particular ontology, or across all ontologies? The former is extremely valuable; the latter only valuable within the context of the repository (i.e., within a search within COR). I will throw in also that the holy grail is to be able to autocomplete within a branch of an ontology. Fast response in any case ends up being dependent on an indexing of the terms in ElasticSearch or similar. I don't think any triple store will natively be able to autocompletely that quickly (unless the triplestore explicitly implements that index).