First, add the jekyll-algolia gem to your Gemfile, in the :jekyll_plugins section. If you do not yet have a Gemfile, here is the minimal content to get your started.

source 'https://rubygems.org'

gem 'jekyll', '~> 2.5.3'

group :jekyll_plugins do
  gem 'jekyll-algolia', '~> 0.7.0'
end

Once this is done, download all dependencies with bundle install.

Then, add jekyll-algolia to your _config.yml file, under the plugins section, like this:

plugins:
  - jekyll-algolia

If everything went well, you should be able to run jekyll help and see the algolia subcommand listed.

Configuration

Add information about your Algolia configuration into the _config.yml file, under the algolia section, like this:

algolia:
  application_id: 'your_application_id'
  index_name:     'your_index_name'

Your write api key will be read from the ALGOLIA_API_KEY environment variable. You can define it on the same line as your command, allowing you to type ALGOLIA_API_KEY='your_write_api_key' bundle exec jekyll algolia push.

Note that your API key should have write access to both the index_name and _tmp suffixed version of it (eg. your_index_name and your_index_name_tmp) in the previous example). This is due to the way we do atomic pushes by pushing to a temporary index and then renaming it.

⚠ Other, unsecure, method ⚠

You can also store your write api key in a file named _algolia_api_key, in your source directory. If you do this we very, very, very strongly encourage you to make sure the file is not tracked in your versioning system.

Options

The plugin uses sensible defaults, but you may want to override some of its configuration. Here are the options you can add to your _config.yml file, under the algolia section:

`excluded_files`

Defines which files should not be indexed for search.

algolia:
  excluded_files:
    - index.html
    - 2015-01-01-post.md

`nodes_to_index`

All HTML nodes matching this CSS Selector will be indexed. Default value is p, meaning that all <p> paragraphs will be indexed.

If you would like to also index lists, you could set it like this:

algolia:
  nodes_to_index: 'p,ul'

`lazy_update`

Enabling this option can greatly reduce the number of operations consumed by the plugin but comes with some drawbacks mentioned above:

false: The plugin will push all the records to a temporary index and once everything is pushed will overwrite the current index with this new one. This is the most straightforward way to update records and will ensure that all the changes happen in one move. This is the default value.

true: With lazy_update enabled, the plugin will try to reduce the number of calls done to the API. It will get a list of all the records in your index and all the records ready to be pushed. It will compare both and push the new while deleting the old. In most cases it should consume less operations, but the changes won't be atomic (ie. you might have your index in an hybrid state, with old records not yet removed and/or new records not yet added for a couple of minutes).

`settings`

Here you can pass any specific index settings to your Algolia index. All the settings supported by the API can be passed here.

Examples

If you want to activate distinct and some snippets for example, you would do:

algolia:
  settings:
    attributeForDistinct: 'hierarchy'
    distinct: true
    attributesToSnippet: ['text:20']

If you want to search in other fields than the default ones, you'll have to edit the attributesToIndex (default is %w(title h1 h2 h3 h4 h5 h6 unordered(text) unordered(tags))

algolia:
  settings:
    attributesToIndex:
      - title
      - h1
      - h2
      - h3
      - h4
      - h5
      - h6
      - unordered(text)
      - unordered(tags)
      - your_custom_attribute_1
      - your_custom_attribute_2
      - ...

Hooks

The AlgoliaSearchRecordExtractor contains two methods (custom_hook_each and custom_hook_all) that are here so you can overwrite them to add your custom logic. By default, they do nothing except returning the argument they take as input, and are placeholder for you to override.

The best way to override them is to create a ./_plugins/search.rb file, with the following content:

class AlgoliaSearchRecordExtractor
  # Hook to modify a record after extracting
  def custom_hook_each(item, node)
    # `node` is a Nokogiri HTML node, so you can access its type through `node.name`
    # or its classname through `node.attr('class')` for example

    # Just return `nil` instead of `item` if you want to discard this record
    item
  end

  # Hook to modify all records after extracting
  def custom_hook_all(items)
    items
  end
end

The AlgoliaSearchJekyllPush class also lets user define the custom_hook_excluded_file? method. This method is called on every file that the plugin thinks it should parse and index. If it returns true, the file is not indexed. You can add here your custom logic to exclude some files.

class AlgoliaSearchJekyllPush < Jekyll::Command
  class << self
    # Hook to exclude some files from indexing
    def custom_hook_excluded_file?(file)
      return true if filepath =~ %r{^/excluded_dir/}
      false
    end
  end
end

Command line

Here is the list of command line options you can pass to the jekyll algolia push command:

Flag	Description
`--config ./_config.yml`	You can here specify the config file to use. Default is `_config.yml`
`--future`	With this flag, the command will also index posts with a future date
`--limit_posts 10`	Limits the number of posts to parse and index
`--drafts`	Index drafts in the `_drafts` folder as well
`--dry-run` or `-n`	Do a dry run, do not actually push anything to your index
`--verbose`	Display more information about what is going to be indexed

Dependencies

This plugin is compatible with version of Jekyll >= 3.6.2 and version of Ruby >= 2.4.0. Those are the versions deployed on GitHub Pages at the time of writing.

You will also need Bundler to install the gem in your project.

Searching

This plugin will index your data in your Algolia index. Building the front-end search is of the scope of this plugin, but you can follow our tutorials or use our forked version of the popular Hyde theme.

GitHub Pages

The initial goal of the plugin was to allow anyone to have access to great search, even on a static website hosted on GitHub pages.

But GitHub does not allow custom plugins to be run on GitHub Pages. This means that you'll either have to run bundle exec jekyll algolia push manually, or configure a CI environment (like [Travis][12] to do it for you.

[Travis CI][13] is an hosted continuous integration service, and it's free for open-source projects. Properly configured, it can automatically reindex your data whenever you push to gh-pages.

For it to work, you'll have 3 steps to perform.

1. Create a `.travis.yml` file

Create a file named .travis.yml at the root of your project, with the following content:

language: ruby
cache: bundler
branches:
  only:
    - gh-pages
script:
  - bundle exec jekyll algolia push
rvm:
 - 2.2

This file will be read by Travis and instruct it to fetch all dependencies defined in the Gemfile, then run jekyll algolia push. This will be triggered when data is pushed to the gh-pages branch.

2. Update your `_config.yml` file to exclude `vendor`

Travis will download all you Gemfile dependencies into a directory named vendor. You have to tell Jekyll to ignore this directory, otherwise Jekyll will try to parse it (and fail).

Doing so is easy, add the following line to your _config.yml file:

exclude: [vendor]

3. Configure Travis

In order for Travis to be able to push data to your index on your behalf, you have to give it your write API Key. This is achieved by defining an ALGOLIA_API_KEY [environment variable][14] in Travis settings.

You should also uncheck the "Build pull requests" option, otherwise any pull request targeting gh-pages will trigger the reindexing.

![Travis Configuration][15]

Done

Commit all the changes to the files, and then push to gh-pages. Travis will catch the event and trigger your indexing for you. You can follow the Travis job execution directly on [their website][16].

FAQS

How can I exclude some HTML nodes from the indexing

By default, the plugin will index every HTML node that matches the nodes_to_index CSS selector option. The default value is p, meaning that it will index all the paragraphs.

You can use a [negation selector][17] to be even more explicit. For example the value p:not(.do-not-index) will index all <p> paragraphs, except those that have the class do-not-index.

If you need a finer granularity on your indexing that cannot be expressed through CSS selectors, you'll have to use the [hook mechanism][18]. The custom_hook_each method takes a [Nokogiri][19] HTML node as a second argument and should let you write more complex filters.

Thanks

Thanks to [Anatoliy Yastreb][20] for a [great tutorial][21] on creating Jekyll plugins.

[12]: https://travis-ci.org/) [13]: https://travis-ci.org/ [14]: http://docs.travis-ci.com/user/environment-variables/ [15]: /docs/travis-settings.png [16]: https://travis-ci.org [17]: https://developer.mozilla.org/en/docs/Web/CSS/:not [18]: #hooks [19]: http://www.nokogiri.org/ [20]: https://github.com/ayastreb/ [21]: https://ayastreb.me/writing-a-jekyll-plugin/

Languages

Ruby 65.1%

SCSS 14%

JavaScript 8.6%

CSS 7.1%

Pug 3.9%

Other 1.2%

README.md

Jekyll Algolia Plugin

Usage

Installation