Move the rewriteLocalLinks behavior to an AST pipeline plugin (github…

…#17550)

* 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

lib/page.js

@@ -5,7 +5,6 @@ const cheerio = require('cheerio')
 const patterns = require('./patterns')
 const getMapTopicContent = require('./get-map-topic-content')
 const rewriteAssetPathsToS3 = require('./rewrite-asset-paths-to-s3')
-const rewriteLocalLinks = require('./rewrite-local-links')
 const getApplicableVersions = require('./get-applicable-versions')
 const encodeBracketedParentheses = require('./encode-bracketed-parentheses')
 const generateRedirectsForPermalinks = require('./redirects/permalinks')
@@ -156,12 +155,6 @@ class Page {
     }
 
     this.intro = await renderContent(this.rawIntro, context)
-
-    // rewrite local links in the intro to include current language code and GHE version if needed
-    const introHtml = cheerio.load(this.intro)
-    rewriteLocalLinks(introHtml, context.currentVersion, context.currentLanguage)
-    this.intro = introHtml('body').html()
-
     this.introPlainText = await renderContent(this.rawIntro, context, { textOnly: true })
     this.title = await renderContent(this.rawTitle, context, { textOnly: true, encodeEntities: true })
     this.shortTitle = await renderContent(this.shortTitle, context, { textOnly: true, encodeEntities: true })
@@ -249,9 +242,6 @@ class Page {
       if (englishHeadings) useEnglishHeadings($, englishHeadings)
     }
 
-    // rewrite local links to include current language code and GHE version if needed
-    rewriteLocalLinks($, context.currentVersion, context.currentLanguage)
-
     // wrap ordered list images in a container div
     $('ol > li img').each((i, el) => {
       $(el).wrap('<div class="procedural-image-wrapper" />')

lib/render-content/create-processor.js

@@ -10,8 +10,9 @@ const html = require('rehype-stringify')
 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')
 
-module.exports = function createProcessor () {
+module.exports = function createProcessor (context) {
   return unified()
     .use(markdown)
     .use(remarkCodeExtra, { transform: codeHeader })
@@ -21,5 +22,6 @@ module.exports = function createProcessor () {
     .use(autolinkHeadings, { behavior: 'wrap' })
     .use(highlight, { languages: { graphql }, subset: false })
     .use(raw)
+    .use(rewriteLocalLinks, { languageCode: context.currentLanguage, version: context.currentVersion })
     .use(html)
 }

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

@@ -0,0 +1,102 @@
+const path = require('path')
+const visit = require('unist-util-visit')
+const externalRedirects = Object.keys(require('../../redirects/external-sites'))
+const { getPathWithoutLanguage, getVersionStringFromPath } = require('../../path-utils')
+const { getNewVersionedPath } = require('../../old-versions-utils')
+const patterns = require('../../patterns')
+const { deprecated, latest } = require('../../enterprise-server-releases')
+const nonEnterpriseDefaultVersion = require('../../non-enterprise-default-version')
+const allVersions = require('../../all-versions')
+const supportedVersions = Object.keys(allVersions)
+const supportedPlans = Object.values(allVersions).map(v => v.plan)
+const removeFPTFromPath = require('../../remove-fpt-from-path')
+
+// Matches any <a> tags with an href that starts with `/`
+const matcher = node => (
+  node.type === 'element' &&
+  node.tagName === 'a' &&
+  node.properties &&
+  node.properties.href &&
+  node.properties.href.startsWith('/')
+)
+
+// Content authors write links like `/some/article/path`, but they need to be
+// rewritten on the fly to match the current language and page version
+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 => {
+      const newHref = getNewHref(node, languageCode, version)
+      if (newHref) {
+        node.properties.href = newHref
+      }
+    })
+  }
+}
+
+function getNewHref (node, languageCode, version) {
+  const { href } = node.properties
+  // Exceptions to link rewriting
+  if (href.startsWith('/assets')) return
+  if (href.startsWith('/public')) return
+  if (externalRedirects.includes(href)) return
+
+  let newHref = href
+  // If the link has a hardcoded plan or version in it, do not update other than adding a language code
+  // Examples:
+  // /[email protected]/rest/reference/oauth-authorizations
+  // /enterprise-server/rest/reference/oauth-authorizations (this redirects to the latest version)
+  // /enterprise-server@latest/rest/reference/oauth-authorizations (this redirects to the latest version)
+  const firstLinkSegment = href.split('/')[1]
+  if ([...supportedPlans, ...supportedVersions, 'enterprise-server@latest'].includes(firstLinkSegment)) {
+    newHref = path.join('/', languageCode, href)
+  }
+
+  // If the link includes a deprecated version, do not update other than adding a language code
+  // Example: /enterprise/11.10.340/admin/articles/upgrading-to-the-latest-release
+  const oldEnterpriseVersionNumber = href.match(patterns.getEnterpriseVersionNumber)
+  if (oldEnterpriseVersionNumber && deprecated.includes(oldEnterpriseVersionNumber[1])) {
+    newHref = path.join('/', languageCode, href)
+  }
+
+  if (newHref === href) {
+    // start clean with no language (TOC pages already include the lang codes via lib/liquid-tags/link.js)
+    const hrefWithoutLang = getPathWithoutLanguage(href)
+
+    // normalize any legacy links so they conform to new link structure
+    newHref = path.posix.join('/', languageCode, getNewVersionedPath(hrefWithoutLang))
+
+    // get the current version from the link
+    const versionFromHref = getVersionStringFromPath(newHref)
+
+    // ------ BEGIN ONE-OFF OVERRIDES ------//
+    // dotcom-only links always point to dotcom
+    if (node.properties.className && node.properties.className.includes('dotcom-only')) {
+      version = nonEnterpriseDefaultVersion
+    }
+
+    // desktop links always point to dotcom
+    if (patterns.desktop.test(hrefWithoutLang)) {
+      version = nonEnterpriseDefaultVersion
+    }
+
+    // admin links on dotcom always point to Enterprise
+    if (patterns.adminProduct.test(hrefWithoutLang) && version === nonEnterpriseDefaultVersion) {
+      version = `enterprise-server@${latest}`
+    }
+
+    // insights links on dotcom always point to Enterprise
+    if (patterns.insightsProduct.test(hrefWithoutLang) && version === nonEnterpriseDefaultVersion) {
+      version = `enterprise-server@${latest}`
+    }
+    // ------ END ONE-OFF OVERRIDES ------//
+
+    // update the version in the link
+    newHref = removeFPTFromPath(newHref.replace(versionFromHref, version))
+  }
+
+  newHref = newHref.replace(patterns.trailingSlash, '$1')
+  return newHref
+}

lib/render-content/renderContent.js

@@ -52,7 +52,7 @@ module.exports = async function renderContent (
     // statements so that extra space doesn't mess with list numbering
     template = template.replace(/(\r?\n){3}/g, '\n\n')
 
-    const processor = createProcessor()
+    const processor = createProcessor(context)
     const vFile = await processor.process(template)
     let html = vFile.toString()