@Earl @gary_alderson

That's the incompatibility of internal #solr
Usually, with the major version number increase, the solr is upgraded. And solr is able to upgrade only one version up, not two. So you can upgrade #YaCy, for example, from 1.93 to 1.94, not 1.96.
Exporting and reimporting the whole index is a way how to cross this limitation, but can take time and disk space, depending on the size of index.
See: https://eldar.cz/yacydoc/dev/solr.html#upgrading
and
https://eldar.cz/yacydoc/operation/index-export-import.html

@benjamin_e @orbiterlab

Trying for news search engine as well, using #YaCy and https://eldar.cz/news/ aggregator. Relevancy while search is not great. The pseudo-pagerank ("citation rank") doesn't work that much and is so heavy for computation that I switched that off:
https://community.searchlab.eu/t/how-to-activate-and-rank-by-cr-citation-rank/1733/5

Vector search would certainly be a big help. #solr already have that, but not implemented in YaCy so far.

For distinguishing news sites, I just use "collections" feature. see https://community.searchlab.eu/t/what-became-of-yacys-gsa-interface-collection-feature/621/7

@kajer you need pipeline but also real time sentiment ranking, checksums for all files seen, cve, ioc - all of a sudden you went from stodgy siem to real world noc

i like the topic even though i do look at it sarcastically

i think you want combinatorials of top 10 db and real time #mentions I think you are going to get people in federated enclaves to join together and work on problems but also make a typical image - everybody can optimize and get vbetter by sharing there will be a representative manifest of sw that will vary by sector #thomas register ocr scan and convert into semantic/vectordb/graphs #rss #solr #keywords #page rank...it reminds me of the site that correlates ip to domains and more - great osint info if found to be true - how do you advertise and make the site have rev streams - you cab advertise to people trolling the sector - you need a template bots and spiders, get all the info you can and cache sites and then get it into a db - the real time part is basically a tagcloud #i ching #book of changes #backlinks

@Lilith

#solr und #lucene gibt es schon im wesentlichen seit 2004.