Normalize YouTube Shorts links to watch URLs

media-scrape/src/extract.ts

@@ -36,6 +36,15 @@ export interface ExtractedMedia {
     title: string | null;
 }
 
+/**
+ * Shorts links break the embed player: it needs `watch?v=`. Rewrite them for storage so the
+ * viewer, "View full", and refreshes all use the canonical form. Other URLs pass through.
+ */
+export function normalizeYouTubeUrl(url: string): string {
+    const match = /^https?:\/\/(?:www\.|m\.)?youtube\.com\/shorts\/([A-Za-z0-9_-]{6,})/.exec(url);
+    return match ? `https://www.youtube.com/watch?v=${match[1]}` : url;
+}
+
 function extension(filename: string | null | undefined): string {
     const match = /\.([a-z0-9]+)$/i.exec(filename ?? "");
     return match ? match[1].toLowerCase() : "";
@@ -75,9 +84,11 @@ function hostname(url: string | null): string | null {
 
 function extractEmbed(embed: CoreEmbed): ExtractedMedia | null {
     const provider = embed.provider ?? hostname(embed.url);
+    const embedUrl = embed.url ? normalizeYouTubeUrl(embed.url) : null;
+    const videoUrl = embed.videoUrl ? normalizeYouTubeUrl(embed.videoUrl) : null;
     const common = {
         source: "embed" as const,
-        playUrl: embed.videoUrl ?? null,
+        playUrl: videoUrl,
         provider,
         downloaded: false,
         size: null,
@@ -91,7 +102,7 @@ function extractEmbed(embed: CoreEmbed): ExtractedMedia | null {
         return {
             ...common,
             kind: "video",
-            url: embed.url ?? embed.videoUrl,
+            url: embedUrl ?? videoUrl ?? "",
             contentType: null,
             previewUrl: embed.thumbnailUrl ?? embed.imageUrl
         };
@@ -102,7 +113,7 @@ function extractEmbed(embed: CoreEmbed): ExtractedMedia | null {
         return {
             ...common,
             kind: "gif",
-            url: embed.url ?? embed.imageUrl ?? "",
+            url: embedUrl ?? embed.imageUrl ?? "",
             contentType: null,
             previewUrl: embed.thumbnailUrl ?? embed.imageUrl
         };
@@ -112,7 +123,7 @@ function extractEmbed(embed: CoreEmbed): ExtractedMedia | null {
         return {
             ...common,
             kind: "image",
-            url: embed.url ?? embed.imageUrl,
+            url: embedUrl ?? embed.imageUrl,
             contentType: null,
             previewUrl: embed.thumbnailUrl ?? embed.imageUrl
         };