Fork me on GitHub
#cljdoc
<
2021-08-04
>
martinklepsch13:08:55

in-docset search would indeed be pretty great! I’m already fantasizing a little bit but it would be really amazing if we indexed different object types and surfaced that information in the search: • namespaces • vars • headings in articles • paragraphs in articles (not sure how practical/possible it is to go this granular)

deleted14:08:04

I think we could do that? I was evaluating client-side search libraries last night and trying to figure out what gave the most not-crappy search feel and it seems like it's probably this one but I'm not 100% sure yet https://github.com/nextapps-de/flexsearch

martinklepsch15:08:53

Does this rely on a node.js backend? Looking at the docs I’m not quite sure

martinklepsch15:08:10

yeah, interactive examples have been a long time “aspirational goal”' 🙂

martinklepsch15:08:28

But it’s tricky to set up, often requires a build etc

martinklepsch18:08:44

server has a 50GB disk, with 20GB still free

martinklepsch18:08:29

I think if we could do it all client-side that’d be pretty cool but I’m not quite sure how that’d work 😄 Also we could consider an implementation that removes indexes that haven’t been used for a while I guess?

martinklepsch19:08:15

that sounds pretty clever 👍

seancorfield22:08:21

Very slick! Nice work @deleted-user!

martinklepsch22:08:44

feels really snappy 🙂