whoosh VS lunr.js

The only other embeddable search engine I'm aware off, Whoosh, is brilliant but building the index was quite slow, and search performance degraded quite a lot as number of documents increase (performance is strictly a non-goal). Meilisearch was comparatively faster, I didn't like managing a server to get "just search" in my scripts and applications. However, their underlying engine Milli solves both issues I had, and all that was needed creating bindings for it.

Is it really "just a single statically linked binary"?

I'd love to use Meilisearch as you describe, but their so-called SDKs are just about for the search client, you still need the HTTP server listening on localhost.

I would love to see something like SQLite based off Meilisearch (i.e. a fully selfcontained library like https://github.com/mchaput/whoosh). Do you know if such a thing exists?

For our full text search, we used whoosh, which works pretty well for moderately big amount of data.

Nearly a decade ago (oh god) I converted some overdesigned five node ES mess to https://github.com/mchaput/whoosh. It's (obviously) not the fastest or anything, but it was more than good enough for low-dozens of GBs of mostly static data.

Your best bet is to start using a proper search library rather than the simple loop with 'in' checks that you have now. A search lib will handle things like Unicode/ASCII similarities, removal of stop words, stemming, TF-IDF (and other) weighting, etc. and will be massively faster as well. Quite a few pages come up if you Google "python search engine", also Whoosh looks promising.

If your content is mostly static, you might want to consider pre-building an index and shipping it as a whole. You could look into something like

* https://stork-search.net/ (Rust/WASM)

* tinysearch: https://github.com/tinysearch/tinysearch (Rust/WASM)

* https://lunrjs.com/ (JS, simple, stable)

* http://elasticlunr.com/ - based on the former, slightly more sophisticated tuning options

Also search results can be pre-indexed and stored in a Json file. Just as an example. https://lunrjs.com/

Would anyone be willing to help make this more accessible and clean? I have some front-end dev experience, but it would be cool to work together with people to make sure we have something that makes sense and looks nicer than what I could do myself. As for functionality, searching on GitHub directly seems to work pretty well, but it might be better to have a page and a search feature maybe using something like Lunr. I would also like to create some sort of easy "API" in case Matt wants to embed some transcripts on his website. It would be cool if it would be as easy as just adding a blank div with a special id and a data attribute with the episode number on the Squarespace page.

which is open source, appears to be free, and claims that it can run in the browser

Is there a way to run it in WASM, to get something like Lunr[1]? We prefer to do our (small-index, <2MB) search client-side for a bunch of reasons, currently using Lunr.js, but it's a bit annoying and the typeahead search is something I improvised and not really official.

[1] https://lunrjs.com/

To ensure cross-platform compatibility, Bitwarden uses Lunr.js for searching. This search engine is a bit quirky, and difficult to get used to.

Slightly more js work required, but this should a more customisable solution: https://lunrjs.com/

OWDX runs entirely in the browser; as such, it cannot offer cutting-edge search functionality of the sort you'd find in a search engine or an expensive piece of enterprise software. The search library I'm using — lunr.js, for anyone who's interested — does, however, offer a nice set of core functionality and a modest but handy query language.