From 726f69484fe67f27747ffe087fb06ac7f1a68cc2 Mon Sep 17 00:00:00 2001
From: Jackson Harper <jacksonh@gmail.com>
Date: Tue, 15 Nov 2022 18:43:55 +0800
Subject: [PATCH] Better content extraction of title and description for GitHub
 URLs

---
 .../src/websites/github-handler.ts            | 21 ++++++++++++++++++-
 1 file changed, 20 insertions(+), 1 deletion(-)

diff --git a/packages/content-handler/src/websites/github-handler.ts b/packages/content-handler/src/websites/github-handler.ts
index c31fe2224..92da72bcd 100644
--- a/packages/content-handler/src/websites/github-handler.ts
+++ b/packages/content-handler/src/websites/github-handler.ts
@@ -12,10 +12,29 @@ export class GitHubHandler extends ContentHandler {
 
   async preParse(url: string, dom: Document): Promise<Document> {
     const body = dom.querySelector('body')
+    const head = dom.querySelector('head')
     const article = dom.querySelector('article')
+    const twitterTitle = dom.querySelector(`meta[name='twitter:title']`)
+    const linkAuthor = dom.querySelector(`span[itemprop='author']`)
 
     if (body && article) {
-      body?.replaceChildren(article)
+      body.replaceChildren(article)
+
+      // Attempt to set the author also. This is available on repo homepages
+      // but not on things like PRs. Ideally we want PRs and issues to have
+      // author set to the author of the PR/issue.
+      if (linkAuthor && linkAuthor.textContent) {
+        const author = dom.createElement('span')
+        author.setAttribute('rel', 'author')
+        author.innerHTML = linkAuthor.textContent
+        article.appendChild(author)
+      }
+    }
+
+    // Remove the GitHub - and repo org from the title
+    const twitterTitleContent = twitterTitle?.getAttribute('content')
+    if (twitterTitle && twitterTitleContent) {
+      twitterTitle.setAttribute('content', twitterTitleContent.replace(/GitHub - .*\//, ''))
     }
 
     return Promise.resolve(dom)