WOW! This is incredible! I'll raise this tomorrow during the Scala meeting to take a vote, but it looks amazing. The demo you link to is really great.

The only issue I foresee is that currently scala.github.com runs on Github Pages which means that Jekyll plugins are forbidden, regardless of the version of Jekyll we run. Therefore, just to re-verify, everything is fine with this PR in the absence of Jekyll plugins? Algolia will build an index from our documentation sources on github? Could you explain a bit more about how that's done? I imagine that if we accept this PR, and then reorganize the site, it could have some sort of implications for your indexing, no?

Hi @heathermiller,

Yes you can't use a Jekyll plugin with Github Pages. But, you can run it in travis like here.

Yes, the PR is complete as the moment (just need to update the Algolia API keys to give you "production" ones.
The index is build from scrapping the web pages of http://docs.scala-lang.org with a internal tool, but we intent to open-source it eventually. For the moment, is it running on our servers.

Yes, if the website is reorganized, it will have some implications on the indexing, because we use CSS selectors to find the data on the pages (eg: all h1 inside .container).

Thanks,

ps: Thanks to @asurak, @Nagriar & @maxiloc for the scrapper

Cool! We have a documentation server inside ENSIME for the user's project's dependency's docs. Would it be possible for us to use your search embedded?

We're also really keen on creating a Hoogol style type search customised to a project. The tough bit is defining the search DSL . is that something you've thought about?

Oh we'd need to wait for you to release on a GPL compatible license.

@fommil: You can use Algolia wherever you want, it's a SaaS product, with a free hacker plan.
I do not understand what you mean by a Hoogol style type search, but we have an "advanced mode"

The dropdown menu is made with autocomplete.js which has a MIT licence.

The scrapping is made by an internal tool, not yet available as OpenSource, but we have a docker image with the python code in it, so you can check it out.

@ElPicador not interested in SaaS, we'd need it integrated, sorry.

Hoogol style search wouldn't be captured by "advanced search". It's really very scala specific.

Hi @ElPicador, thanks again for this PR! We discussed it at a group meeting, and everyone seems to like it :)

Two things:

1. We might reorganize the site soon (update the look & feel, as it's quite out-dated). That'd require some work on your scraping tool, I imagine. Probably only a few minutes worth of work, as most of the DOM would remain mostly the same, but an update nonetheless. If we do that in the next month or two, would you guys be up for updating the scrape?
2. Would you update this PR with an appropriate API key?

Thanks again. This is very cool :)

@heathermiller

1. Yeah sure, we'll do it. Just keep us informed
2. I'll update it before monday

@heathermiller I had time today, so the API keys are updated

@heathermiller: Could you wait for tomorrow? We have so UI improvement on the way (like http://eslint.org/)

Sure thing :) Thanks @ElPicador

@heathermiller I tried to improve the UI a bit:

You can also find here the docker image of the scrapper. For the moment it's running in our infra.

If you have any issues please contact [email protected]

Awesome, thanks @ElPicador! LGTM

Thanks @heathermiller 😄

@ElPicador : Just a suggestion, but it'd be awesome if we could do something similar for the Scala specification which has no search functionality at all currently: http://www.scala-lang.org/files/archive/spec/2.11/ .