flexsearch VS lyra

Full-text search is powered by FlexSearch and Nextra will index all of your pages at build time ⚡.

I have a web page that reads in various .md files and displays them as HTML. The app uses the marked library to convert the markdown into HTML for display. I create a flexsearch search index out of the raw text values from the documents (raw text is gathered using DOMParser over all HTML elements) so that user can search for keywords in the docs and get back a table of results. The order of operations and search index code looks like:

I am trying to get flexsearch (a lib written in js) up and running in a TS project and found a working example here. I downloaded the project and ran 'yarn add flexsearch' and also 'yarn add @/types/flexsearch' since I know that you need a special index.d.ts file to convert the JS to TS properly.... however the code errors out during the Index object creation with the message.

Try using a proper browser search like Lunr or Flexsearch

Current version of FlexSearch (0.7.2) is not typo tolerant, see https://github.com/nextapps-de/flexsearch/issues/118

Very nice! Seems to perform very well. I'm curious, have you compared Fuse with other search engines? Like flex search or elasticlunr? Why did you choose fuse ?

I actually looked into this term before localeCompare(): Full Text Search. It's pretty heavy duty. In JavaScript, this can come in the form of a library dependency like FlexSearch. Far too bulky for the humble sorting task I have at hand.

Javascript seems to have a comprehensive in memory solution https://github.com/nextapps-de/flexsearch

Another option that was often suggested to me was building the search index with a library such as https://github.com/nextapps-de/flexsearch and distribute the index than to the client and handle it one the client. But yeah sounds like a lot of overhead and I haven't tried it.

Since Ballard's codec is "AI" based, can you add google's lyrav2 ( https://github.com/google/lyra ) and Facebook's/meta EnCodec ( https://github.com/facebookresearch/encodec ).

Also I don't seem to be able to access your page, so there might be error.

Finally, when doing opus comparison it's good now to denote if it is using Lace or NoLace decoder post processing filters that became available in opus 1.5 (note, this feature need to be enabled at compile time, and defying decode a new API call needs to be made to force higher complexity decoder) . See https://opus-codec.org/demo/opus-1.5/

I've thought about doing something similar for google's voice compression lyra https://github.com/google/lyra

edit: It seems Lyra is opensource https://github.com/google/lyra

1) https://github.com/google/lyra/releases/tag/v1.3.0

Very impressive.

It'd be interesting to see what the lift would be to get encoding & decoding running in webassembly/wasm. Further, it'd be really neat to try to take something like the tflife_model_wrapper[1] and to get it backed by something like tsjs-tflite[2] perhaps even atop for example tfjs-backend-webgpu[3].

Longer run, the web-nn[4] spec should hopefully simplify/bake-in some of these libraries to the web platform, make running inference much easier. But there's still an interesting challenge & question, that I'm not sure how to tackle; how to take native code, compile it to wasm, but to have some of the implementation provided else-where.

[1] https://github.com/google/lyra/pull/89/files#diff-ed2f131a63...

[2] https://www.npmjs.com/package/@tensorflow/tfjs-tflite

[3] https://www.npmjs.com/package/@tensorflow/tfjs-backend-webgp...

[4] https://www.w3.org/TR/webnn/

You can find an Android, Linux and macOS app here: https://github.com/google/lyra/actions/runs/3156735950

Here's the repository: https://github.com/google/lyra and it's licensed under Apache.