Contributions

[ziodave]

Hello @wetneb

I forked the project and added several contributions to the project. I didn't bother you so far as I wanted to understand what improvements I could bring to the project.

I'll try to enumerate them here:

1. In order to avoid highlighting stop words or similar (ref Remove noisy words #29), a Solr schema supporting query optimization for more than 40 languages, removing stop words and/or using OpenNLP to filter the most important words (nouns). In order to select the proper analyzer, a new language parameter may be required (ref Allow language selection to prevent aliases adding noise instead of helping #3).
2. Currently it's not possible to configure the profile with subclassOf higher in the hierarchy because the SPARQL query times out. I provide another implementation which performs several queries in order to retrieve (and cache) the children. As part of this improvement I provide a User Agent as per Wikidata requirements (to avoid rate limitation).
3. Add relevant sameAs properties to the index. For places also add geo coordinates. For events also add start/end times.

As part of 1. I intend also to provide a Solr docker image with the required libraries and models, given that they're compatible with the project license (partially related to #31).

I quite like the results so far and I would be happy to orderly make a PR for you to review them and merge them in upstream.

Let me know what you think ⭐️

[wetneb]

Hi @ziodave, this is absolutely fantastic! Yes I would be very interested in integrating such changes in this repository. If you could make PRs that would be a bit granular (introducing one feature at a time, basically) it would help to review them, but if that is too much effort I am sure we can find other ways.

[ziodave]

Great! In the next days, I'll cleanup things a bit and I'll start making PRs.

[eracle]

Any news on that?