Search

Let users search your Docsy site with a choice of configurable search options.

Docsy offers multiple options that let your readers search your site content, so you can pick one that suits your needs. You can choose from:

  • Google Custom Search Engine (GCSE), the default option, which uses Google’s index of your public site to generate a search results page.
  • Algolia DocSearch, which uses Algolia’s indexing and search mechanism, and provides an organized dropdown of search results when your readers use the search box. Algolia DocSearch is free for public documentation sites.
  • Local search with Lunr, which uses Javascript to index and search your site without the need to connect to external services. This option doesn’t require your site to be public.

If you enable any of these search options in your project configuration file, a search box displays in the right of your top navigation bar. By default a search box also displays at the top of the section menu in the left navigation pane, which you can disable if you prefer, or if you’re using a search option that only works with the top search box.

Be aware that if you accidentally enable more than one search option in your project configuration file, you may get unexpected results (for example, if you have added the .js for Algolia DocSearch, you’ll get Algolia results if you enable GCSE search but forget to disable Algolia search).

By default, the search box appears in both the top navigation bar and at the top of the sidebar left navigation pane. If you don’t want the sidebar search box, set sidebar_search_disable to true in hugo.toml/hugo.yaml/hugo.json:

sidebar_search_disable = true
sidebar_search_disable: true
"sidebar_search_disable": true

Configure search with a Google Custom Search Engine

By default Docsy uses a Google Custom Search Engine (GCSE) to search your site. To enable this feature, you’ll first need to make sure that you have built and deployed a production version of your site, as otherwise your site won’t be crawled and indexed.

  1. Create a Google Custom Search Engine for your deployed site by clicking New search engine on the Custom Search page and following the instructions. Make a note of the ID for your new search engine.

  2. Add any further configuration you want to your search engine using the Edit search engine options. In particular you may want to do the following:

    • Select Look and feel. Change from the default Overlay layout to Results only, as this option means your search results are embedded in your search page rather than appearing in a separate box. Click Save to save your changes.
    • Edit the default result link behavior so that search results from your site don’t open in a new tab. To do this, select Search Features - Advanced - Websearch Settings. In the Link Target field, type “_parent”. Click Save to save your changes.

Adding the search page

Once you have your search engine set up, you can add the feature to your site:

  1. Ensure you have a Markdown file in content/en/search.md (and one per other languages if needed) to display your search results. It only needs a title and layout: search, as in the following example:

    +++
    title = "Search Results"
    layout = "search"
    +++
    ---
    title: Search Results
    layout: search
    ---
    {
        "title": "Search Results",
        "layout": "search"
    }
  2. Add your Google Custom Search Engine ID to the site params in hugo.toml/hugo.yaml/hugo.json. You can add different values per language if needed.

    gcs_engine_id = "011737558837375720776:fsdu1nryfng"
    gcs_engine_id: '011737558837375720776:fsdu1nryfng'
    {
      "gcs_engine_id": "011737558837375720776:fsdu1nryfng"
    }

If you don’t specify a Google Custom Search Engine ID for your project and haven’t enabled any other search options, the search box won’t appear in your site. If you’re using the default hugo.toml from the example site and want to disable search, just comment out or remove the relevant line.

Configure Algolia DocSearch

As an alternative to GCSE, you can use Algolia DocSearch with this theme. Algolia DocSearch is free for public documentation sites. Docsy supports Algolia DocSearch v3.

Sign up for Algolia DocSearch

Complete the form at https://docsearch.algolia.com/apply/.

If you are accepted to the program, you will receive the code to add to your documentation site from Algolia by email.

Adding Algolia DocSearch

  1. Enable Algolia DocSearch in hugo.toml/hugo.yaml/hugo.json.

    algolia_docsearch = true
    algolia_docsearch: true
    {
      "algolia_docsearch": true
    }
  2. Remove or comment out any GCSE ID in hugo.toml/hugo.yaml/hugo.json and ensure local search is set to false as you can only have one type of search enabled. See Disabling GCSE search.

  3. Disable the sidebar search in hugo.toml/hugo.yaml/hugo.json as this is not currently supported for Algolia DocSearch. See Disabling the sidebar search box.

  4. Add the CSS and JS to use Algolia to the head and body of every page in your site, following the instructions in Add code to head or before body end.

    • In head-end.html add the DocSearch CSS:

      <link rel="stylesheet" href="https://cdn.jsdelivr.net/npm/@docsearch/css@3" />
      
    • In body-end.html add the DocSearch script, replacing the docsearch details with the snippet you get from Algolia (the example below is Algolia’s own site index!). You must provide #docsearch as your container value as that’s the ID of the div we provide in Docsy’s layout:

      <script src="https://cdn.jsdelivr.net/npm/@docsearch/js@3"></script>
      <script type="text/javascript">docsearch({
        container: '#docsearch',
        appId: 'R2IYF7ETH7',
        apiKey: '599cec31baffa4868cae4e79f180729b',
        indexName: 'docsearch',
        });</script>
      

You can find out more about how to configure DocSearch in the Algolia DocSearch V3 Getting started guide.

When you’ve completed these steps, Algolia search should be enabled on your site. Search results are displayed as a pop-up, so you don’t need to add any search results page.

Configure local search with Lunr

Lunr is a Javascript-based search option that lets you index your site and make it searchable without the need for external, server-side search services. This is a good option particularly for smaller or non-public sites.

To add Lunr search to your Docsy site:

  1. Enable local search in hugo.toml/hugo.yaml/hugo.json.

    offlineSearch = true
    offlineSearch: true
    {
      "offlineSearch": true
    }
  2. Remove or comment out any GCSE ID in hugo.toml/hugo.yaml/hugo.json and ensure Algolia DocSearch is set to false, as you can only have one type of search enabled. See Disabling GCSE search.

Once you’ve completed these steps, local search is enabled for your site and results appear in a drop down when you use the search box.

Changing the summary length of the local search results

You can customize the summary length by setting offlineSearchSummaryLength in hugo.toml/hugo.yaml/hugo.json.

#Enable offline search with Lunr.js
offlineSearch = true
offlineSearchSummaryLength = 200
offlineSearch: true
offlineSearchSummaryLength: 200
{
  "offlineSearch": true,
  "offlineSearchSummaryLength": 200
}

You can customize the maximum result count by setting offlineSearchMaxResults in hugo.toml/hugo.yaml/hugo.json.

offlineSearch = true
offlineSearchMaxResults = 25
offlineSearch: true
offlineSearchMaxResults: 25
{
  "offlineSearch": true,
  "offlineSearchMaxResults": 25
}

Changing the width of the local search results popover

The width of the search results popover will automatically widen according to the content.

If you want to limit the width, add the following scss into assets/scss/_variables_project.scss.

.td-offline-search-results {
  max-width: 460px;
}

Excluding pages from local search results

To exclude pages from local search results, add exclude_search: true to the the frontmatter of each page:

+++
title = "Index"
weight = 10
exclude_search = true
+++
---
title: "Index"
weight: 10
exclude_search: true
---
{
  "title": "Index",
  "weight": 10,
  "exclude_search": true
}