How about Xapian, an option that the author totally ignores. https://xapian.org/...

1vuio0pswjnm7 · on Oct 15, 2022

https://xapian.org/users.html

Also, it is used by the the public-inbox Perl scripts. See https://public-inbox.org

That includes the Linux kernel mailing list, Git mailing list and other lists available via lore.kernel.org:

https://lore.kernel.org

For example, to search the BPF mailing list for "bpf" and sort results by relevance:

https://lore.kernel.org/bpf/?q=bpf&r

To view the Xapian query parser operators:

https://lore.kernel.org/bpf/_/text/help

throwaway81523 · on Oct 15, 2022

I think that isn't really built for large datasets. I had a buddy who used it a lot on smaller web sites though. I'm way out of date and this stuff might be horribly old fashioned. Vector search is the new hotness but on the other hand, search systems everywhere are getting worse, so who knows.

rurban · on Oct 15, 2022

No, gmane was big enough. And it's far advanced over PostgresQL FTS or Elasticseatch