Remove cheerio from Page#render (github#17566)

* Write our plugin

* Include it

* Move the RegEx

* Don't rewriteLocalLinks with cheerio anymore

* Process after HTML ast is generated

* Use the same logic as before, just to see if it'll pass

* Don't require languageCode/version

* Only work on local links

* Needs an href

* Only update href if there's a new one to use

* Check for node.properties

* Some links are just mean

* Move use-english-headings to be a plugin

* Bail if no englishHeadings were passed

* Install rehype-wrap

* Wrap ol > li img in div.procedural-image-wrapper

* Test for platform without cheerio

* Use a plugin for rewriteAssetPathsToS3

* Remove cheerio from page.render

* Fix require paths

* SImplify

* Fix some 🐛s

* Use our own rehype-wrap

* Move rewriteAssetPathsToS3 after HTML AST

* Remove some console logs

* Fix check for includesPlatformSpecificContent

* Rename ast => tree

lib/page.js

@@ -4,12 +4,10 @@ const path = require('path')
 const cheerio = require('cheerio')
 const patterns = require('./patterns')
 const getMapTopicContent = require('./get-map-topic-content')
-const rewriteAssetPathsToS3 = require('./rewrite-asset-paths-to-s3')
 const getApplicableVersions = require('./get-applicable-versions')
 const encodeBracketedParentheses = require('./encode-bracketed-parentheses')
 const generateRedirectsForPermalinks = require('./redirects/permalinks')
 const getEnglishHeadings = require('./get-english-headings')
-const useEnglishHeadings = require('./use-english-headings')
 const getTocItems = require('./get-toc-items')
 const pathUtils = require('./path-utils')
 const Permalink = require('./permalink')
@@ -154,6 +152,12 @@ class Page {
       this.markdown = await this.getMarkdown()
     }
 
+    // use English IDs/anchors for translated headings, so links don't break (see #8572)
+    if (this.languageCode !== 'en') {
+      const englishHeadings = getEnglishHeadings(this, context.pages)
+      context.englishHeadings = englishHeadings
+    }
+
     this.intro = await renderContent(this.rawIntro, context)
     this.introPlainText = await renderContent(this.rawIntro, context, { textOnly: true })
     this.title = await renderContent(this.rawTitle, context, { textOnly: true, encodeEntities: true })
@@ -185,6 +189,7 @@ class Page {
       }))
     }
 
+    context.relativePath = this.relativePath
     const html = await renderContent(markdown, context)
 
     // product frontmatter may contain liquid
@@ -227,29 +232,14 @@ class Page {
       })
     }
 
-    const $ = cheerio.load(html)
-
     // set a flag so layout knows whether to render a mac/windows/linux switcher element
-    this.includesPlatformSpecificContent = $('[class^="platform-"], .mac, .windows, .linux').length > 0
-
-    // rewrite asset paths to s3 if it's a dotcom article on any GHE version
-    // or if it's an enterprise article on any GHE version EXCEPT latest version
-    rewriteAssetPathsToS3($, context.currentVersion, this.relativePath)
-
-    // use English IDs/anchors for translated headings, so links don't break (see #8572)
-    if (this.languageCode !== 'en') {
-      const englishHeadings = getEnglishHeadings(this, context.pages)
-      if (englishHeadings) useEnglishHeadings($, englishHeadings)
-    }
-
-    // wrap ordered list images in a container div
-    $('ol > li img').each((i, el) => {
-      $(el).wrap('<div class="procedural-image-wrapper" />')
-    })
-
-    const cleanedHTML = $('body').html()
+    this.includesPlatformSpecificContent = (
+      html.includes('extended-markdown mac') ||
+      html.includes('extended-markdown windows') ||
+      html.includes('extended-markdown linux')
+    )
 
-    return cleanedHTML
+    return html
   }
 
   // Allow other modules (like custom liquid tags) to make one-off requests

lib/render-content/create-processor.js

@@ -11,6 +11,9 @@ const graphql = require('highlightjs-graphql').definer
 const remarkCodeExtra = require('remark-code-extra')
 const codeHeader = require('./plugins/code-header')
 const rewriteLocalLinks = require('./plugins/rewrite-local-links')
+const useEnglishHeadings = require('./plugins/use-english-headings')
+const rewriteAssetPathsToS3 = require('./plugins/rewrite-asset-paths-to-s3')
+const wrapInElement = require('./plugins/wrap-in-element')
 
 module.exports = function createProcessor (context) {
   return unified()
@@ -19,9 +22,12 @@ module.exports = function createProcessor (context) {
     .use(emoji)
     .use(remark2rehype, { allowDangerousHTML: true })
     .use(slug)
+    .use(useEnglishHeadings, context)
     .use(autolinkHeadings, { behavior: 'wrap' })
     .use(highlight, { languages: { graphql }, subset: false })
     .use(raw)
+    .use(rewriteAssetPathsToS3, context)
+    .use(wrapInElement, { selector: 'ol > li img', wrapper: 'div.procedural-image-wrapper' })
     .use(rewriteLocalLinks, { languageCode: context.currentLanguage, version: context.currentVersion })
     .use(html)
 }

lib/render-content/plugins/rewrite-asset-paths-to-s3.js

@@ -0,0 +1,45 @@
+const visit = require('unist-util-visit')
+const latestEnterpriseRelease = require('../../enterprise-server-releases').latest
+const nonEnterpriseDefaultVersion = require('../../non-enterprise-default-version')
+const { getS3BucketPathFromVersion } = require('../../s3-bucket-path-utils')
+const allVersions = require('../../all-versions')
+const s3BasePath = 'https://github-images.s3.amazonaws.com'
+
+const matcher = node => (
+  node.type === 'element' &&
+  node.tagName === 'img' &&
+  node.properties.src &&
+  node.properties.src.startsWith('/assets/images')
+)
+
+// This module rewrites asset paths on Enterprise versions to S3 paths.
+// Source example: /assets/images/foo.png
+// Rewritten: https://github-images.s3.amazonaws.com/enterprise/2.20/assets/images/foo.png
+// The one exception is Admin pages on the latest GHES release.
+module.exports = function rewriteAssetPathsToS3 ({ currentVersion, relativePath }) {
+  // Bail if we don't have a relativePath in this context
+  if (!relativePath) return
+
+  // skip if this is the homepage
+  if (relativePath === 'index.md') return
+
+  // if the current version is non-enterprise, do not rewrite
+  if (currentVersion === nonEnterpriseDefaultVersion) return
+
+  // the relativePath starts with the product, like /admin/foo or /github/foo
+  const product = relativePath.split('/')[0]
+
+  // if this is an Admin page on the latest GHES release, do not rewrite
+  if (product === 'admin' && allVersions[currentVersion].currentRelease === latestEnterpriseRelease) return
+
+  // if the version is enterprise-server@2.22, use `enterprise/2.22` as the bucket path
+  // otherwise, use the plan name, e.g., `github-ae`
+  const bucketPath = getS3BucketPathFromVersion(currentVersion)
+
+  return tree => {
+    visit(tree, matcher, node => {
+      // Rewrite the node's src
+      node.properties.src = `${s3BasePath}/${bucketPath}${node.properties.src}`
+    })
+  }
+}

lib/render-content/plugins/rewrite-local-links.js

@@ -26,8 +26,8 @@ module.exports = function rewriteLocalLinks ({ languageCode, version }) {
   // There's no languageCode or version passed, so nothing to do
   if (!languageCode || !version) return
 
-  return ast => {
-    visit(ast, matcher, node => {
+  return tree => {
+    visit(tree, matcher, node => {
       const newHref = getNewHref(node, languageCode, version)
       if (newHref) {
         node.properties.href = newHref

lib/render-content/plugins/use-english-headings.js

@@ -0,0 +1,26 @@
+const GithubSlugger = require('github-slugger')
+const Entities = require('html-entities').XmlEntities
+const visit = require('unist-util-visit')
+const slugger = new GithubSlugger()
+const entities = new Entities()
+
+const matcher = node => (
+  node.type === 'element' &&
+  ['h2', 'h3', 'h4'].includes(node.tagName)
+)
+
+// replace translated IDs and links in headings with English
+module.exports = function useEnglishHeadings ({ englishHeadings }) {
+  if (!englishHeadings) return
+  return tree => {
+    visit(tree, matcher, node => {
+      const text = node.children.find(n => n.type === 'text').value
+      // find English heading in the collection
+      const englishHeading = englishHeadings[entities.encode(text)]
+      // get English slug
+      const englishSlug = slugger.slug(englishHeading)
+      // use English slug for heading ID and link
+      node.properties.id = englishSlug
+    })
+  }
+}

lib/render-content/plugins/wrap-in-element.js

@@ -0,0 +1,33 @@
+const visit = require('unist-util-visit')
+const { selectAll } = require('hast-util-select')
+const parseSelector = require('hast-util-parse-selector')
+
+/*
+ * Attacher
+ */
+module.exports = options => {
+  options = options || {}
+  const selector = options.selector || options.select || 'body'
+  const wrapper = options.wrapper || options.wrap
+
+  /*
+   * Transformer
+   */
+  return tree => {
+    if (typeof wrapper !== 'string') {
+      throw new TypeError('Expected a `string` as wrapper')
+    }
+
+    if (typeof selector !== 'string') {
+      throw new TypeError('Expected a `string` as selector')
+    }
+
+    for (const match of selectAll(options.selector, tree)) {
+      visit(tree, match, (node, i, parent) => {
+        const wrapper = parseSelector('div')
+        wrapper.children = [node]
+        parent.children[i] = wrapper
+      })
+    }
+  }
+}

package.json

@@ -49,6 +49,8 @@
     "got": "^9.6.0",
     "gray-matter": "^4.0.1",
     "hast-util-from-parse5": "^6.0.1",
+    "hast-util-parse-selector": "^2.2.5",
+    "hast-util-select": "^4.0.2",
     "hastscript": "^6.0.0",
     "helmet": "^3.21.2",
     "highlightjs-graphql": "^1.0.2",