From 6e421f87547142138022589f2c78d516b8126e3d Mon Sep 17 00:00:00 2001
From: Mathias Jakobsen <mathias.jakobsen@overleaf.com>
Date: Fri, 10 Feb 2023 15:13:00 +0000
Subject: [PATCH] Merge pull request #11765 from overleaf/mj-count-images

[web] Add script for counting images in Mongo

GitOrigin-RevId: 1f2795d73d9afa0f0ae46b1e29a03a93b1d85ee5
---
 services/web/scripts/count_image_files.js | 76 +++++++++++++++++++++++
 1 file changed, 76 insertions(+)
 create mode 100644 services/web/scripts/count_image_files.js

diff --git a/services/web/scripts/count_image_files.js b/services/web/scripts/count_image_files.js
new file mode 100644
index 0000000000..7b41a836f6
--- /dev/null
+++ b/services/web/scripts/count_image_files.js
@@ -0,0 +1,76 @@
+const { ReadPreference } = require('mongodb')
+const { db, waitForDb } = require('../app/src/infrastructure/mongodb')
+const { extname } = require('node:path')
+
+const FILE_TYPES = [
+  '.jpg',
+  '.jpeg',
+  '.png',
+  '.bmp',
+  '.webp',
+  '.svg',
+  '.pdf',
+  '.eps',
+  '.gif',
+  '.ico',
+  '.tiff',
+]
+
+const longestFileType = Math.max(...FILE_TYPES.map(fileType => fileType.length))
+
+async function main() {
+  await waitForDb()
+  const projects = db.projects.find(
+    {},
+    { projection: { rootFolder: 1 }, readPreference: ReadPreference.SECONDARY }
+  )
+  let projectsProcessed = 0
+  const result = new Map(FILE_TYPES.map(fileType => [fileType, 0]))
+  for await (const project of projects) {
+    projectsProcessed += 1
+    if (projectsProcessed % 100000 === 0) {
+      console.log(projectsProcessed, 'projects processed')
+    }
+    countFiles(project.rootFolder[0], result)
+  }
+
+  const sortedResults = [...result.entries()].sort(
+    ([, countA], [, countB]) => countB - countA
+  )
+
+  sortedResults.forEach(([fileType, count]) => {
+    console.log(
+      `${fileType.padStart(longestFileType, ' ')}: ${count
+        .toString()
+        .padStart(7, ' ')}`
+    )
+  })
+}
+
+function countFiles(folder, result) {
+  if (folder.folders) {
+    for (const subfolder of folder.folders) {
+      countFiles(subfolder, result)
+    }
+  }
+
+  if (folder.fileRefs) {
+    for (const file of folder.fileRefs) {
+      const fileType = extname(file.name).toLowerCase()
+      const current = result.get(fileType)
+      if (current !== undefined) {
+        result.set(fileType, current + 1)
+      }
+    }
+  }
+  return result
+}
+
+main()
+  .then(() => {
+    process.exit(0)
+  })
+  .catch(err => {
+    console.error(err)
+    process.exit(1)
+  })