fix: normalize extracted video poster paths

Fixes #414
nuxt-modules · Jan 28, 2025 · ed18278 · ed18278
1 parent e022206
commit ed18278
Show file tree

Hide file tree

Showing 3 changed files with 51 additions and 15 deletions.
diff --git a/src/prerender.ts b/src/prerender.ts
@@ -8,6 +8,7 @@ import chalk from 'chalk'
 import { dirname } from 'pathe'
 import { defu } from 'defu'
 import type { ConsolaInstance } from 'consola'
+import { withSiteUrl } from 'nuxt-site-config/kit'
 import { extractSitemapMetaFromHtml } from './util/extractSitemapMetaFromHtml'
 import type { ModuleRuntimeConfig, SitemapUrl } from './runtime/types'
 import { splitForLocales } from './runtime/utils-pure'
@@ -80,12 +81,17 @@ export function setupPrerenderHandler(_options: { runtimeConfig: ModuleRuntimeCo
           route._sitemap._sitemap = _sitemap
         }
       }
+
       route._sitemap = defu(extractSitemapMetaFromHtml(html, {
         images: options.discoverImages,
         videos: options.discoverVideos,
         // TODO configurable?
         lastmod: true,
         alternatives: true,
+        resolveUrl(s) {
+          // if the match is relative
+          return s.startsWith('/') ? withSiteUrl(s) : s
+        },
       }), route._sitemap) as SitemapUrl
     })
     nitro.hooks.hook('prerender:done', async () => {

diff --git a/src/util/extractSitemapMetaFromHtml.ts b/src/util/extractSitemapMetaFromHtml.ts
@@ -1,6 +1,4 @@
-import { withSiteUrl } from 'nuxt-site-config/kit'
 import { parseURL } from 'ufo'
-import { tryUseNuxt } from '@nuxt/kit'
 import type { ResolvedSitemapUrl, SitemapUrl, VideoEntry } from '../runtime/types'
 
 const videoRegex = /<video[^>]*>([\s\S]*?)<\/video>/g
@@ -20,13 +18,14 @@ const videoLiveRegex = /<video[^>]*\sdata-live="([^"]+)"/
 const videoTagRegex = /<video[^>]*\sdata-tag="([^"]+)"/
 const sourceRegex = /<source[^>]*\ssrc="([^"]+)"/g
 
-export function extractSitemapMetaFromHtml(html: string, options?: { images?: boolean, videos?: boolean, lastmod?: boolean, alternatives?: boolean }) {
+export function extractSitemapMetaFromHtml(html: string, options?: { images?: boolean, videos?: boolean, lastmod?: boolean, alternatives?: boolean, resolveUrl?: (s: string) => string }) {
   options = options || { images: true, videos: true, lastmod: true, alternatives: true }
   const payload: Partial<SitemapUrl> = {}
+  const resolveUrl = options?.resolveUrl || ((s: string) => s)
+  const mainRegex = /<main[^>]*>([\s\S]*?)<\/main>/
+  const mainMatch = mainRegex.exec(html)
   if (options?.images) {
     const images = new Set<string>()
-    const mainRegex = /<main[^>]*>([\s\S]*?)<\/main>/
-    const mainMatch = mainRegex.exec(html)
     if (mainMatch?.[1] && mainMatch[1].includes('<img')) {
       // Extract image src attributes using regex on the HTML, but ignore elements with invalid values such as data:, blob:, or file:
       // eslint-disable-next-line regexp/no-useless-lazy,regexp/no-super-linear-backtracking
@@ -37,10 +36,7 @@ export function extractSitemapMetaFromHtml(html: string, options?: { images?: bo
         // This is necessary to avoid infinite loops with zero-width matches
         if (match.index === imgRegex.lastIndex)
           imgRegex.lastIndex++
-        let url = match[1]
-        // if the match is relative
-        if (url.startsWith('/'))
-          url = tryUseNuxt() ? withSiteUrl(url) : url
+        const url = resolveUrl(match[1])
         images.add(url)
       }
     }
@@ -50,9 +46,6 @@ export function extractSitemapMetaFromHtml(html: string, options?: { images?: bo
 
   if (options?.videos) {
     const videos = []
-    const mainRegex = /<main[^>]*>([\s\S]*?)<\/main>/
-    const mainMatch = mainRegex.exec(html)
-
     if (mainMatch?.[1] && mainMatch[1].includes('<video')) {
       let videoMatch
       while ((videoMatch = videoRegex.exec(mainMatch[1])) !== null) {
@@ -109,11 +102,12 @@ export function extractSitemapMetaFromHtml(html: string, options?: { images?: bo
 
         if (sources.length > 0) {
           videos.push(...sources.map((source) => {
-            if (source.startsWith('/'))
-              source = tryUseNuxt() ? withSiteUrl(source) : source
+            if (videoObj.thumbnail_loc) {
+              videoObj.thumbnail_loc = resolveUrl(String(videoObj.thumbnail_loc))
+            }
             return {
               ...videoObj,
-              content_loc: source,
+              content_loc: resolveUrl(source),
             }
           }))
         }

diff --git a/test/unit/extractSitemapMetaFromHtml.test.ts b/test/unit/extractSitemapMetaFromHtml.test.ts
@@ -309,4 +309,40 @@ describe('extractSitemapMetaFromHtml', () => {
       }
     `)
   })
+  it('extracts relative poster as absolute', async () => {
+    const testcase5 = extractSitemapMetaFromHtml(`
+<main>
+      <video
+        controls
+        src="https://archive.org/download/BigBuckBunny_124/Content/big_buck_bunny_720p_surround.mp4"
+        poster="/poster.jpg"
+        width="620"
+        data-title="Big Buck Bunny"
+        data-description="Big Buck Bunny in DivX 720p."
+      >
+              <source
+          src="https://archive.org/download/DuckAndCover_185/CivilDefenseFilm-DuckAndCoverColdWarNuclearPropaganda_512kb.mp4"
+          type="video/mp4"
+        />
+        </video>
+        </main>
+       `, {
+      videos: true,
+      resolveUrl(s) {
+        return s.startsWith('/') ? `https://example.com${s}` : s
+      },
+    })
+    expect(testcase5).toMatchInlineSnapshot(`
+      {
+        "videos": [
+          {
+            "content_loc": "https://archive.org/download/DuckAndCover_185/CivilDefenseFilm-DuckAndCoverColdWarNuclearPropaganda_512kb.mp4",
+            "description": "Big Buck Bunny in DivX 720p.",
+            "thumbnail_loc": "https://example.com/poster.jpg",
+            "title": "Big Buck Bunny",
+          },
+        ],
+      }
+    `)
+  })
 })