Site search middleware about caddy

[pedronasser]

Could you give some more information you have in mind for this idea? An example?
How would this search work? How would the result be printed?

[mholt]

Here's what I was thinking:

• At startup, the middleware launches a service goroutine that runs for the lifetime of the server. Initially it creates an index of the site... presumably in memory, although there are good packages for lightweight persistent data stores available too.
• Periodically the service function re-indexes.
• An endpoint is "exposed" (meaning that a request to a certain URL would be handled by the search middleware and not passed onto other handlers) such as /search (customizable). A GET request here with a "q" parameter would perform a search.
• The search results would be rendered by a template specified by the user (or perhaps a default template), similar to the browse middleware.
• The search endpoint could also be exposed as a JSON API

[pedronasser]

Ok, got everything.
I'll try work something out.

[mholt]

Awesome! You got it. This is something that would make a great addition to any site.

There is a pretty popular indexing library called bleve that would be worth looking into.

Also, for the record, it will be important to be able to specify a base path to index, but also make it possible to exclude paths from being indexed, for example, if part of the site is a dynamic PHP site or something - no point in indexing PHP source files.

[mholt]

Yay, this is done! \o/ Thanks @pedronasser

[blankoworld]

@mholt what's about this functionnality? Which directive should I use to have a search bar in my website and a /search endpoint?

[francislavoie]

@blankoworld I think it was https://github.com/pedronasser/caddy-search, but it doesn't seem supported anymore as far as I can tell. It's not in the official list of plugins anymore