Skip to content

Latest commit

 

History

History
129 lines (89 loc) · 8.91 KB

search.md

File metadata and controls

129 lines (89 loc) · 8.91 KB

Search

Overview

This site's search functionality.

To see all existing search-related issues and pull requests, visit github.com/github/docs/labels/search.


search-screenshot


How to search

The site search is part of every version of docs.github.com. This endpoint responds in JSON format, and fronts our search querying functionality. We recommend using this endpoint, as the endpoint will be more stable. On any page, you can use the search box to search the documents we've indexed. You can also query our search endpoint directly at: https://docs.github.com/search?version=<VERSION>&language=<LANGUAGE CODE>&filters=topics:<TOPIC>&query=<QUERY>

  • The VERSION can be any numbered GitHub Enterprise Server version (e.g., 2.22, 3.0), GitHub AE (ghae), or the Free pro team plan (dotcom).
  • The LANGUAGE CODE can be one of: zh, es, pt, ru, ja, fr, de, ko
  • TOPIC can be any topics in the allowed list of topics. The values in the topics attribute are not case sensitive, so filtering on GitHub actions or github actions will return the same result. Note: Currently, the topics filter only works for the dotcom version in the English language. We plan to expand this search query to other languages and versions in the future.
  • Any search QUERY you'd like.

For example, to filter on the topic ssh and the query passphrases, you'd use this search query:

https://docs.github.com/search?version=dotcom&language=en&filters=topics:ssh&query=passphrases

To filter for the topics oauth apps and github apps and the query install, you'd use this search query:

https://docs.github.com/search?version=dotcom&language=en&filters=topics:'oauth apps'+AND+topics:'github apps'&query=install

Using the topics search filter

Using the attribute topics in your query will only return results that have the matching topic value. When the topic contains spaces, you must use quotes. Filters with spaces or combining filters requires special syntax.

Production deploys

A GitHub Actions workflow that runs every four hours syncs the search data. This process generates structured data for all pages on the site, compares that data to what's currently on search, then adds, updates, or removes indices based on the diff of the local and remote data, being careful not to create duplicate records and avoiding any unnecessary (and costly) indexing operations.

The Actions workflow progress can be viewed (by GitHub employees) in the Actions tab of the repo.

Manually triggering the search index update workflow

You can manually run the workflow to generate the indexes after you push your changes to main to speed up the indexing when needed. It's recommended to do this for only the free-pro-team@latest version and the en language because running all languages and versions takes about 40 minutes. To run it manually, click "Run workflow" button in the Actions tab. Enter the language and version you'd like to generate the indexes for as inputs to the workflow. By default, all languages and versions are generated.

Build and sync

To build all the indices (this takes about an hour):

npm run sync-search

To build indices for a specific language and/or version and sync them:

VERSION=<PLAN@RELEASE LANGUAGE=<TWO-LETTER CODE> npm run sync-search

You can set VERSION and LANGUAGE individually, too.

Substitute a currently supported version for <PLAN@RELEASE> and a currently supported two-letter language code for <TWO-LETTER-CODE>. Languages and versions are lowercase. The options for version are currently free-pro-team, github-ae, and enterprise-server.

Label-triggered Actions workflow

Docs team members can use an Actions workflow on GHES release PRs by applying a label in this format:

sync-english-index-for-<PLAN@RELEASE>

This label will run a workflow on every push that builds and uploads ONLY the English index for the specified version. This means:

  • The GHES content will be searchable at the same time the release PR is shipped, with no delay.
  • The GHES content will be searchable on staging throughout content creation.
  • No manual steps (unless you want to do a dry run test).

Why do we need this? For our daily shipping needs, it's tolerable that search updates aren't available for up to an hour after the content goes live. But GHES releases are more time-sensitive, and writers have a greater need to preview search data on staging.

Files

Actions workflow files

Code files

Indices

There's a separate search index for each combination of product and language. Some examples:

Index Name Description
github-docs-dotcom-en GitHub.com English
github-docs-dotcom-zh GitHub.com Chinese
github-docs-dotcom-es GitHub.com Spanish
github-docs-2.18-en GitHub Enterprise 2.18 English
github-docs-2.18-zh GitHub Enterprise 2.18 Chinese
github-docs-2.18-es GitHub Enterprise 2.18 Spanish
github-docs-2.17-en GitHub Enterprise 2.17 English
github-docs-2.17-zh GitHub Enterprise 2.17 Chinese
github-docs-2.17-es GitHub Enterprise 2.17 Spanish

Records

Each record represents a section of a page. Sections are derived by splitting up pages by their headings. Each record has a title, intro (if one exists in the frontmatter), body content (in text, not HTML), a url, and a unique objectID that is currently just the permalink of the article. Here's an example:

{
  objectID: '/en/actions/creating-actions/about-actions#about-actions',
  url: 'https://docs.github.com/en/actions/creating-actions/about-actions#about-actions',
  slug: 'about-actions',
  breadcrumbs: 'GitHub Actions / Creating actions / About actions',
  heading: 'About actions',
  title: 'About actions',
  content: "You can create actions by writing custom code that interacts with your repository in any way you'd like..."
}

Notes

  • It's not strictly necessary to set an objectID as the search index will create one automatically, but by creating our own we have a guarantee that subsequent invocations of this upload script will overwrite existing records instead of creating numerous duplicate records with differing IDs.
  • Our search querying has typo tolerance. Try spelling something wrong and see what you get!
  • Our search querying has lots of controls for customizing each index, so we can add weights to certain attributes and create rules like "title is more important than body", etc. But it works pretty well as-is without any configuration.
  • Our search querying has support for "advanced query syntax" for exact matching of quoted expressions and exclusion of words preceded by a - sign. This is off by default but we have it enabled in our browser client. The settings in the web interface can be overridden by the search endpoint. See middleware/search.js.
  • When needed, the Docs Engineering team can commit updates to the search index, as long as the label skip-index-check is applied to the PR.