From ac02aae67fefb74fbe953f610a3d77cadc356aa6 Mon Sep 17 00:00:00 2001
From: Joe Mou <dev@mou.fo>
Date: Thu, 12 Mar 2026 15:46:12 -0400
Subject: Fix GitHub scraping for new JSON structure, update TypeScript
 interfaces

GitHub restructured their embedded JSON: tree/blob data is now split
across codeViewRepoRoute, codeViewTreeRoute, codeViewBlobLayoutRoute,
codeViewBlobLayoutRoute.StyledBlob, codeViewBlobRoute, and the shared
codeViewLayoutRoute (repo, refInfo, path).

Replace RawPayload and OverviewFile with typed interfaces for each
route, add AppPayload as the top-level union, and simplify parsePayload
to always extract react-app.embeddedData.
---
 src/scraper.test.ts | 7 +++++--
 1 file changed, 5 insertions(+), 2 deletions(-)

(limited to 'src/scraper.test.ts')

diff --git a/src/scraper.test.ts b/src/scraper.test.ts
index 125e6a6..ad3ed46 100644
--- a/src/scraper.test.ts
+++ b/src/scraper.test.ts
@@ -95,8 +95,11 @@ describe("GitHub scraper", () => {
       assert.strictEqual(data.language, "Markdown");
       assert.strictEqual(data.image, false);
 
-      assert.strictEqual(data.textLines, null);
-      assert.strictEqual(data.htmlLines, null);
+      assert.strictEqual(data.textLines?.[0], "# deploy-pages 🚀");
+      assert.strictEqual(
+        data.htmlLines?.[0],
+        '<span class="pl-mh"># deploy-pages 🚀</span>deploy-pages 🚀<span class="pl-en">deploy-pages 🚀</span>',
+      );
       const firstLine =
         /^<article class="markdown-body entry-content container-lg" itemprop="text"><div class="markdown-heading" dir="auto"><h1 tabindex="-1" class="heading-element" dir="auto">deploy-pages 🚀<\/h1>/;
       assert.match(data.htmlContent ?? "", firstLine);
-- 
cgit v1.3.1