Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Show HN: Frontend Fuzzy Search

github.com

Post reply on HN
Print view View original

11–20 of 38 posts

Re: Show HN: Frontend Fuzzy Search

#11

Post by RoyalSloth »

Good job, this looks really well done. One thing I am wondering, however, is how easy it would be to integrate a pre-generated inverted index file into this fuzzy searcher?

Context: Say I have a bunch of blog posts from which I create an inverted index at "export to html" time, in order to avoid indexing the blog content at runtime on every page visit. Is there a way to persist the internal state of the fuzzy search across different page requests (e.g, /blog-post-1, /blog-post-2), so it only builds its internal state/index once?

The pre-generated inverted index could be quite large and I would like to avoid parsing it on every page request.

Re: Show HN: Frontend Fuzzy Search

#12

Post by kmschaal »

RoyalSloth wrote: post #11

Good job, this looks really well done. One thing I am wondering, however, is how easy it would be to integrate a pre-generated inverted index file into this fuzzy searcher? Context: Say I have a bunch of blog posts from which I create an inverted index at "export to html" time, in order to avoid indexing the blog content at runtime on every page visit. Is there a way to persist the internal state of the fuzzy search…

Thank you for your kind feedback! That's a great idea. I have implemented a memento object that can be used for transferring the serialized state of the searcher. The intent of the implementation was to transfer the searcher between a web worker and the main thread. You could try to serve the Memento from the server and store it in an index db. You may have a look at the web worker example I provide in the repository.

Re: Show HN: Frontend Fuzzy Search

#13

Post by RoyalSloth »

kmschaal wrote: post #12
RoyalSloth wrote: post #11

Good job, this looks really well done. One thing I am wondering, however, is how easy it would be to integrate a pre-generated inverted index file into this fuzzy searcher? Context: Say I have a bunch of blog posts from which I create an inverted index at "export to html" time, in order to avoid indexing the blog content at runtime on every page visit. Is there a way to persist the internal state of the fuzzy search…

Thank you for your kind feedback! That's a great idea. I have implemented a memento object that can be used for transferring the serialized state of the searcher. The intent of the implementation was to transfer the searcher between a web worker and the main thread. You could try to serve the Memento from the server and store it in an index db. You may have a look at the web worker example I provide in the repository…

> You could try to serve the Memento from the server and store it in an index db

I am not sure what you mean by "store it in an index db", but I was thinking about using the searcher on a static website (no real backend, only a fileserver serving pre-generated html files). So if I understand you correctly, in order for this to work I would have to cache Memento via a local storage and load it on every page load/search request.

Unfortunately the index would change over time, thus one would have to detect this somehow and regenerate Memento as well.

Re: Show HN: Frontend Fuzzy Search

#14

Post by kmschaal »

RoyalSloth wrote: post #13
kmschaal wrote: post #12

Earlier quoted context omitted.

Thank you for your kind feedback! That's a great idea. I have implemented a memento object that can be used for transferring the serialized state of the searcher. The intent of the implementation was to transfer the searcher between a web worker and the main thread. You could try to serve the Memento from the server and store it in an index db. You may have a look at the web worker example I provide in the repository…

> You could try to serve the Memento from the server and store it in an index db I am not sure what you mean by "store it in an index db", but I was thinking about using the searcher on a static website (no real backend, only a fileserver serving pre-generated html files). So if I understand you correctly, in order for this to work I would have to cache Memento via a local storage and load it on every page load/searc…

Sorry for the confusion. I think you could generate the memento each time you compile your blog into HTML. The memento can be stored as a json file and served statically by your fileserver. When a user visits your page, retrieve the memento from the server. Then, initialize the searcher with it. In this way you avoid indexing content at runtime.

As a bonus you could cache the memento in the local storage or session storage.

Re: Show HN: Frontend Fuzzy Search

#15

Post by _heimdall »

Nice to see this as a framework-agnostic library!

I have always been a bit unsure when to reach for client-side code for festures like search, filtering, and pagination though. The features are powerful for sure and if the site is otherwise static I can see some value there, but it means having to ship all searchable data to the browser. It feels like an easy security risk to avoid, and offloading that work to a backend where the state lives feels more straightforward (as long as you have a hosted backend at all).

Re: Show HN: Frontend Fuzzy Search

#16

Post by metayrnc »

Almost all frontend fuzzy search libraries I tried fail in correcting small typos (whoch is pretty common when searching). From the demo it seems this library can handle it. I will definitely be trying it out on my own project. Thanks for sharing it!

Re: Show HN: Frontend Fuzzy Search

#17

Post by trescenzi »

Have you considered building the search piece in wasm for performance? A few years back I was looking for a library like this and couldn’t find one that seemed fast enough for what I wanted. I ended up building one in rust and it’s easily 2x as fast as the js code I started with.

It’s not nearly as good a fuzzy search as this is as I just kinda tinkered with it till it felt ok for the purpose but it could be way better.

Source: https://github.com/trescenzi/brainstorm Site: https://brainstorm.tcrez.xyz/

Re: Show HN: Frontend Fuzzy Search

#18

Post by kmschaal »

trescenzi wrote: post #17

Have you considered building the search piece in wasm for performance? A few years back I was looking for a library like this and couldn’t find one that seemed fast enough for what I wanted. I ended up building one in rust and it’s easily 2x as fast as the js code I started with. It’s not nearly as good a fuzzy search as this is as I just kinda tinkered with it till it felt ok for the purpose but it could be way bett…

Hi, thanks for sharing! No, I haven't taken wasm into consideration. It would be interesting to see the performance gain, in particular for indexing.

Re: Show HN: Frontend Fuzzy Search

#19

Post by kmschaal »

metayrnc wrote: post #16

Almost all frontend fuzzy search libraries I tried fail in correcting small typos (whoch is pretty common when searching). From the demo it seems this library can handle it. I will definitely be trying it out on my own project. Thanks for sharing it!

Thank you, sounds great!

Re: Show HN: Frontend Fuzzy Search

#20

Post by Kalabasa »

RoyalSloth wrote: post #11

Good job, this looks really well done. One thing I am wondering, however, is how easy it would be to integrate a pre-generated inverted index file into this fuzzy searcher? Context: Say I have a bunch of blog posts from which I create an inverted index at "export to html" time, in order to avoid indexing the blog content at runtime on every page visit. Is there a way to persist the internal state of the fuzzy search…

When i did my static site search function some time ago, I used Elasticlunr. I was able to pregenerate the index file as a big json file that is loaded at the client.

http://elasticlunr.com/

Post reply on HN