From 1ab2b88ddece8184ac5cdbb817fa8a117a807a40 Mon Sep 17 00:00:00 2001
From: bensteinberg <bsteinberg@law.harvard.edu>
Date: Fri, 22 Nov 2024 13:58:19 -0500
Subject: [PATCH] Limit size of video attachment

---
 README.md        |  1 +
 Scoop.js         | 70 +++++++++++++++++++++++++++---------------------
 Scoop.test.js    | 17 ++++++++++++
 bin/cli.js       |  7 +++++
 options.js       |  1 +
 options.types.js |  4 ++-
 6 files changed, 68 insertions(+), 32 deletions(-)

diff --git a/README.md b/README.md
index 5affa1e0..b4b6c944 100644
--- a/README.md
+++ b/README.md
@@ -196,6 +196,7 @@ Options:
   --capture-window-x <number>                            Width of the browser window Scoop will open to capture, in pixels. (default: 1600)
   --capture-window-y <number>                            Height of the browser window Scoop will open to capture, in pixels. (default: 900)
   --max-capture-size <number>                            Size limit for the capture's exchanges list, in bytes. (default: 209715200)
+  --max-video-capture-size <number>                      Size limit for the video attachment, in bytes. Scoop will not capture video attachments larger than this. (default: 209715200)
   --auto-scroll <bool>                                   Should Scoop try to scroll through the page? (choices: "true", "false", default: "true")
   --auto-play-media <bool>                               Should Scoop try to autoplay `<audio>` and `<video>` tags? (choices: "true", "false", default: "true")
   --grab-secondary-resources <bool>                      Should Scoop try to download img srcsets and secondary stylesheets? (choices: "true", "false", default: "true")
diff --git a/Scoop.js b/Scoop.js
index 074a6377..7d8fe5f4 100644
--- a/Scoop.js
+++ b/Scoop.js
@@ -920,6 +920,7 @@ export class Scoop {
         '--output', `"${videoFilename}"`,
         '--no-check-certificate',
         '--proxy', `'http://${this.options.proxyHost}:${this.options.proxyPort}'`,
+        '--max-filesize', `"${this.options.maxVideoCaptureSize}"`,
         this.url
       ]
 
@@ -948,14 +949,19 @@ export class Scoop {
           const body = await readFile(`${this.captureTmpFolderPath}${file}`)
           const isEntryPoint = false // TODO: Reconsider whether this should be an entry point.
 
-          this.addGeneratedExchange(url, httpHeaders, body, isEntryPoint)
-          videoSaved = true
+          if (body.length) {
+            this.addGeneratedExchange(url, httpHeaders, body, isEntryPoint)
+            videoSaved = true
 
-          // Push to map of available videos and subtitles
-          const index = file.replace('.mp4', '')
+            // Push to map of available videos and subtitles
+            const index = file.replace('.mp4', '')
 
-          if (!(index in availableVideosAndSubtitles)) {
-            availableVideosAndSubtitles[index] = []
+            if (!(index in availableVideosAndSubtitles)) {
+              availableVideosAndSubtitles[index] = []
+            }
+          } else {
+            videoSaved = false
+            this.log.warn(`No content in ${file}.`)
           }
         } catch (err) {
           this.log.warn(`Error while creating exchange for ${file}.`)
@@ -964,7 +970,7 @@ export class Scoop {
       }
 
       // Subtitles
-      if (file.startsWith('video-extracted-') && file.endsWith('.vtt')) {
+      if (videoSaved && file.startsWith('video-extracted-') && file.endsWith('.vtt')) {
         try {
           const url = `file:///${file}`
           const httpHeaders = new Headers({ 'content-type': 'text/vtt' })
@@ -998,41 +1004,43 @@ export class Scoop {
     //
     // Try to add metadata to exchanges
     //
-    try {
-      metadataParsed = []
+    if (videoSaved) {
+      try {
+        metadataParsed = []
 
-      // yt-dlp returns JSONL when there is more than 1 video
-      for (const line of metadataRaw.split('\n')) {
-        if (line) {
-          metadataParsed.push(JSON.parse(line)) // May throw
+        // yt-dlp returns JSONL when there is more than 1 video
+        for (const line of metadataRaw.split('\n')) {
+          if (line) {
+            metadataParsed.push(JSON.parse(line)) // May throw
+          }
         }
-      }
 
-      if (!metadataParsed.length) {
-        throw new Error('yt-dlp reported success (returned 0) but produced no metadata.')
-      }
+        if (!metadataParsed.length) {
+          throw new Error('yt-dlp reported success (returned 0) but produced no metadata.')
+        }
 
-      // Merge parsed metadata into a single JSON string and clean it before saving it
-      const metadataAsJSON = JSON
-        .stringify(metadataParsed, null, 2)
-        .replaceAll(this.captureTmpFolderPath, '')
+        // Merge parsed metadata into a single JSON string and clean it before saving it
+        const metadataAsJSON = JSON
+          .stringify(metadataParsed, null, 2)
+          .replaceAll(this.captureTmpFolderPath, '')
 
-      const url = 'file:///video-extracted-metadata.json'
-      const httpHeaders = new Headers({ 'content-type': 'application/json' })
-      const body = Buffer.from(metadataAsJSON)
-      const isEntryPoint = false
+        const url = 'file:///video-extracted-metadata.json'
+        const httpHeaders = new Headers({ 'content-type': 'application/json' })
+        const body = Buffer.from(metadataAsJSON)
+        const isEntryPoint = false
 
-      this.addGeneratedExchange(url, httpHeaders, body, isEntryPoint)
-      metadataSaved = true
-    } catch (err) {
-      this.log.warn('Error while creating exchange for file:///video-extracted-medatadata.json.')
-      this.log.trace(err)
+        this.addGeneratedExchange(url, httpHeaders, body, isEntryPoint)
+        metadataSaved = true
+      } catch (err) {
+        this.log.warn('Error while creating exchange for file:///video-extracted-medatadata.json.')
+        this.log.trace(err)
+      }
     }
 
     //
     // Generate summary page
     //
-    if ((videoSaved || metadataSaved || subtitlesSaved) === false) {
+    if (videoSaved === false) {
       this.log.warn('yt-dlp reported success (returned 0), but produced no output.')
       return
     }
diff --git a/Scoop.test.js b/Scoop.test.js
index b075112c..37db6c5e 100644
--- a/Scoop.test.js
+++ b/Scoop.test.js
@@ -34,6 +34,8 @@ await test('Scoop - capture of a web page.', async (t) => {
   app.get('/redirect', (req, res) => res.redirect(parseInt(req.query.statusCode), req.query.path))
   app.get('/:path', (req, res) => res.sendFile(FIXTURES_PATH + req.params.path))
 
+  const testVideoFixture = await readFile(`${FIXTURES_PATH}video.mp4`)
+
   /*
    * TESTS
    */
@@ -72,6 +74,21 @@ await test('Scoop - capture of a web page.', async (t) => {
       'file:///video-extracted-summary.html'
     ]
     assert.deepEqual(expected.filter(url => urls.includes(url)), expected)
+    const attachment = exchanges.filter(
+      ex => ex.url === 'file:///video-extracted-1.mp4'
+    )[0]
+    assert.deepEqual(attachment.response.body, testVideoFixture)
+  })
+
+  await t.test('Scoop observes maxVideoCaptureSize', async (_t) => {
+    const { exchanges } = await Scoop.capture(`${URL}/test.html`, { ...options, captureVideoAsAttachment: true, maxVideoCaptureSize: 50000 })
+    const urls = exchanges.map(ex => ex.url)
+    const expected = [
+      'file:///video-extracted-1.mp4',
+      'file:///video-extracted-metadata.json',
+      'file:///video-extracted-summary.html'
+    ]
+    assert.deepEqual(expected.filter(url => urls.includes(url)), [])
   })
 
   await t.test('Scoop can be configured for different window dimensions', async (_t) => {
diff --git a/bin/cli.js b/bin/cli.js
index 5cbead7d..924fac14 100755
--- a/bin/cli.js
+++ b/bin/cli.js
@@ -181,6 +181,13 @@ program.addOption(
     .default(defaults.maxCaptureSize)
 )
 
+program.addOption(
+  new Option(
+    '--max-video-capture-size <number>',
+    'Size limit for the video attachment, in bytes. Scoop will not capture video attachments larger than this.')
+    .default(defaults.maxVideoCaptureSize)
+)
+
 //
 // Behaviors
 //
diff --git a/options.js b/options.js
index 9c7df05f..6c3f5102 100644
--- a/options.js
+++ b/options.js
@@ -26,6 +26,7 @@ export const defaults = {
   captureWindowY: 900,
 
   maxCaptureSize: 200 * 1024 * 1024,
+  maxVideoCaptureSize: 200 * 1024 * 1024,
 
   autoScroll: true,
   autoPlayMedia: true,
diff --git a/options.types.js b/options.types.js
index f14e012f..17126b47 100644
--- a/options.types.js
+++ b/options.types.js
@@ -21,7 +21,9 @@
  * @property {number} captureWindowX=1600 - Browser window resolution in pixels: X axis.
  * @property {number} captureWindowY=900 - Browser window resolution in pixels: Y axis.
  *
- * @property {number} maxCaptureSize=209715200 - Maximum size, in bytes, for the exchanges list. Scoop stop intercepting exchanges at this threshold.
+ * @property {number} maxCaptureSize=209715200 - Maximum size, in bytes, for the exchanges list. Scoop stops intercepting exchanges at this threshold.
+ *
+ * @property {number} maxVideoCaptureSize=209715200 - Maximum size, in bytes, for video attachments. Scoop will not capture video attachments larger than this.
  *
  * @property {boolean} autoScroll=true - Should Scoop try to scroll through the page?
  * @property {boolean} autoPlayMedia=true - Should Scoop try to autoplay `<audio>` and `<video>` tags?