Migrate history-v1 recover_zip scripts from archiver to zip-stream (#32813)

* migrate recover_zip_from_backup from archiver to zip-stream

Replace the `archiver` package with `zip-stream` (the lower-level library
that `archiver` wraps) in the `recover_zip_from_backup.mjs` script and
`backupArchiver.mjs` library. The `archiver` package has known issues with
hanging when creating large zip files and is no longer actively maintained.

Changes:
- Add `zip-stream@^7.0.2` as a direct dependency
- Update `backupArchiver.mjs` to use promisified `ZipStream.entry()`
  instead of `Archiver.append()`
- Rewrite `recover_zip_from_backup.mjs` to use `ZipStream` with
  `stream/promises.pipeline` for cleaner async flow
- Keep `archiver` dependency for `project_archive.js` (separate code path)

Agent-Logs-Url: https://github.com/overleaf/internal/sessions/0df27a8b-97f1-43cc-ac26-f5247a84313f

Co-authored-by: briangough <7457354+briangough@users.noreply.github.com>

* extract finalize timeout to named constant

Agent-Logs-Url: https://github.com/overleaf/internal/sessions/0df27a8b-97f1-43cc-ac26-f5247a84313f

Co-authored-by: briangough <7457354+briangough@users.noreply.github.com>

* convert recover_zip.js to zip-stream, remove finalize timeout, add verbose logging

Agent-Logs-Url: https://github.com/overleaf/internal/sessions/9380d08a-d813-4e9f-a2ac-4891122c163b

Co-authored-by: briangough <7457354+briangough@users.noreply.github.com>

* add acceptance tests for recover_zip_from_backup in raw and latest modes

Agent-Logs-Url: https://github.com/overleaf/internal/sessions/9380d08a-d813-4e9f-a2ac-4891122c163b

Co-authored-by: briangough <7457354+briangough@users.noreply.github.com>

* fix comment formatting in recover_zip_from_backup.mjs

Agent-Logs-Url: https://github.com/overleaf/internal/sessions/9380d08a-d813-4e9f-a2ac-4891122c163b

Co-authored-by: briangough <7457354+briangough@users.noreply.github.com>

* restore EventEmitter.defaultMaxListeners in recover_zip.js, add acceptance test

Agent-Logs-Url: https://github.com/overleaf/internal/sessions/e7443126-22d5-4d0e-a176-a7a5dba49ffd

Co-authored-by: briangough <7457354+briangough@users.noreply.github.com>

* fix formatting

* refactor: simplify stream handling by using named imports for pipeline

* fix blob hash verification in backup acceptance tests

* fix recover_zip script and tests

* fix: exit with non-zero status on error in recover_zip.js

Agent-Logs-Url: https://github.com/overleaf/internal/sessions/ef3f109b-488f-47c9-84a5-b5269387166a

Co-authored-by: briangough <7457354+briangough@users.noreply.github.com>

* migrate from npm to yarn

---------

Co-authored-by: copilot-swe-agent[bot] <198982749+Copilot@users.noreply.github.com>
Co-authored-by: briangough <7457354+briangough@users.noreply.github.com>
Co-authored-by: Brian Gough <briangough@users.noreply.github.com>
GitOrigin-RevId: 6255f9610f3c846790e2ed8b1979ac08b7effece

services/history-v1/package.json

@@ -44,7 +44,8 @@
     "temp": "^0.8.3",
     "throng": "^4.0.0",
     "tsscmp": "^1.0.6",
-    "utf-8-validate": "^5.0.4"
+    "utf-8-validate": "^5.0.4",
+    "zip-stream": "^7.0.2"
   },
   "devDependencies": {
     "@overleaf/migrations": "workspace:*",

services/history-v1/storage/lib/backupArchiver.mjs

@@ -185,13 +185,30 @@ class BackupBlobStore {
  */
 
 /**
- * @typedef {(import('archiver').Archiver)} Archiver
+ * @typedef {(import('zip-stream').default)} ZipStream
  */
 
 /**
  * @typedef {(import('overleaf-editor-core').FileMap)} FileMap
  */
 
+/**
+ * Promisified wrapper for ZipStream's entry method.
+ *
+ * @param {ZipStream} archive
+ * @param {Buffer|NodeJS.ReadableStream|string} source
+ * @param {{ name: string }} data
+ * @return {Promise<void>}
+ */
+function addEntry(archive, source, data) {
+  return new Promise((resolve, reject) => {
+    archive.entry(source, data, err => {
+      if (err) reject(err)
+      else resolve()
+    })
+  })
+}
+
 /**
  *
  * @param historyId
@@ -254,14 +271,15 @@ async function fetchBlob(historyId, hash, persistor) {
 
 /**
  * @typedef {object} AddChunkOptions
- * @property {string} [prefix] Should include trailing slash (if length > 0)
+ * @property {string} [prefix]
  * @property {boolean} [useBackupGlobalBlobs]
+ * @property {boolean} [verbose]
  */
 
 /**
  *
  * @param {History} history
- * @param {Archiver} archive
+ * @param {ZipStream} archive
  * @param {CachedPerProjectEncryptedS3Persistor} projectCache
  * @param {string} historyId
  * @param {AddChunkOptions} [options]
@@ -272,7 +290,7 @@ async function addChunkToArchive(
   archive,
   projectCache,
   historyId,
-  { prefix = '', useBackupGlobalBlobs = false } = {}
+  { prefix = '', useBackupGlobalBlobs = false, verbose = false } = {}
 ) {
   const chunkBlobs = new Set()
   history.findBlobHashes(chunkBlobs)
@@ -334,9 +352,16 @@ async function addChunkToArchive(
         }
         content = await blobStore.getStream(hash)
       }
-      archive.append(content, {
+      if (content == null) {
+        logger.error({ filePath }, 'File content is empty')
+        continue
+      }
+      await addEntry(archive, content, {
         name: `${prefix}${filePath}`,
       })
+      if (verbose) {
+        logger.info({ filePath: `${prefix}${filePath}` }, 'added to archive')
+      }
     }
   })
 }
@@ -358,17 +383,20 @@ async function findStartVersionOfLatestChunk(historyId) {
 /**
  * Restore a project from the latest snapshot
  *
- * There is an assumption that the database backup has been restored.
+ * There is an assumption that the database backup
+ * has been restored.
  *
- * @param {Archiver} archive
+ * @param {ZipStream} archive
  * @param {string} historyId
  * @param {boolean} [useBackupGlobalBlobs]
+ * @param {boolean} [verbose]
  * @return {Promise<void>}
  */
 export async function archiveLatestChunk(
   archive,
   historyId,
-  useBackupGlobalBlobs = false
+  useBackupGlobalBlobs = false,
+  verbose = false
 ) {
   logger.info({ historyId, useBackupGlobalBlobs }, 'Archiving latest chunk')
 
@@ -386,20 +414,28 @@ export async function archiveLatestChunk(
 
   await addChunkToArchive(backedUpChunk, archive, projectCache, historyId, {
     useBackupGlobalBlobs,
+    verbose,
   })
 
   return archive
 }
 
 /**
- * Fetches all raw blobs from the project and adds them to the archive.
+ * Fetches all raw blobs from the project and adds
+ * them to the archive.
  *
  * @param {string} historyId
- * @param {Archiver} archive
+ * @param {ZipStream} archive
  * @param {CachedPerProjectEncryptedS3Persistor} projectCache
+ * @param {boolean} [verbose]
  * @return {Promise<void>}
  */
-async function addRawBlobsToArchive(historyId, archive, projectCache) {
+async function addRawBlobsToArchive(
+  historyId,
+  archive,
+  projectCache,
+  verbose = false
+) {
   const blobKeys = await projectCache.listDirectoryKeys(
     projectBlobsBucket,
     projectKey.format(historyId)
@@ -411,9 +447,13 @@ async function addRawBlobsToArchive(historyId, archive, projectCache) {
         key,
         { autoGunzip: true }
       )
-      archive.append(stream, {
-        name: path.join(historyId, 'blobs', key),
+      const entryName = path.join(historyId, 'blobs', key)
+      await addEntry(archive, stream, {
+        name: entryName,
       })
+      if (verbose) {
+        logger.info({ entryName }, 'added to archive')
+      }
     } catch (err) {
       logger.warn({ err, path: key }, 'Failed to append blob to archive')
     }
@@ -425,17 +465,20 @@ async function addRawBlobsToArchive(historyId, archive, projectCache) {
  *
  * This can work without the database being backed up.
  *
- * It will split the project into chunks per directory and download the blobs alongside the chunk.
+ * It will split the project into chunks per directory
+ * and download the blobs alongside the chunk.
  *
- * @param {Archiver} archive
+ * @param {ZipStream} archive
  * @param {string} historyId
  * @param {boolean} [useBackupGlobalBlobs]
+ * @param {boolean} [verbose]
  * @return {Promise<void>}
  */
 export async function archiveRawProject(
   archive,
   historyId,
-  useBackupGlobalBlobs = false
+  useBackupGlobalBlobs = false,
+  verbose = false
 ) {
   const projectCache = await getProjectPersistor(historyId)
 
@@ -454,11 +497,15 @@ export async function archiveRawProject(
 
     const { buffer } = await loadChunkByKey(projectCache, key)
 
-    archive.append(buffer, {
-      name: `${historyId}/chunks/${chunkId}/chunk.json`,
+    const entryName = `${historyId}/chunks/${chunkId}/chunk.json`
+    await addEntry(archive, buffer, {
+      name: entryName,
     })
+    if (verbose) {
+      logger.info({ entryName }, 'added to archive')
+    }
   }
-  await addRawBlobsToArchive(historyId, archive, projectCache)
+  await addRawBlobsToArchive(historyId, archive, projectCache, verbose)
 }
 
 export class BackupPersistorError extends OError {}

services/history-v1/storage/lib/streams.js

@@ -6,19 +6,19 @@
  */
 'use strict'
 
-const Stream = require('node:stream')
+const { pipeline } = require('node:stream/promises')
 const zlib = require('node:zlib')
 const { WritableBuffer } = require('@overleaf/stream-utils')
 
 /**
  * Create a promise for the result of reading a stream to a buffer.
  *
- * @param {Stream.Readable} readStream
+ * @param {import('node:stream').Readable} readStream
  * @return {Promise<Buffer>}
  */
 async function readStreamToBuffer(readStream) {
   const bufferStream = new WritableBuffer()
-  await Stream.promises.pipeline(readStream, bufferStream)
+  await pipeline(readStream, bufferStream)
   return bufferStream.contents()
 }
 
@@ -33,7 +33,7 @@ exports.readStreamToBuffer = readStreamToBuffer
 async function gunzipStreamToBuffer(readStream) {
   const gunzip = zlib.createGunzip()
   const bufferStream = new WritableBuffer()
-  await Stream.promises.pipeline(readStream, gunzip, bufferStream)
+  await pipeline(readStream, gunzip, bufferStream)
   return bufferStream.contents()
 }
 

services/history-v1/storage/scripts/recover_zip.js

@@ -17,9 +17,10 @@ const fs = require('node:fs')
 const os = require('node:os')
 const path = require('node:path')
 const util = require('node:util')
+const { pipeline } = require('node:stream/promises')
 
-// Something is registering 11 listeners, over the limit of 10, which generates
-// a lot of warning noise.
+// Something is registering 11 listeners, over the limit
+// of 10, which generates a lot of warning noise.
 require('node:events').EventEmitter.defaultMaxListeners = 11
 
 const config = require('config')
@@ -27,11 +28,23 @@ const config = require('config')
 // eslint-disable-next-line import/no-extraneous-dependencies
 const { Storage } = require('@google-cloud/storage')
 const isValidUtf8 = require('utf-8-validate')
+// zip-stream@7 uses ESM default export
+const ZipStream = require('zip-stream').default
+
+function createStorage() {
+  const opts = {}
+  if (config.has('persistor.gcs.endpoint.apiEndpoint')) {
+    opts.apiEndpoint = config.get('persistor.gcs.endpoint.apiEndpoint')
+  }
+  if (config.has('persistor.gcs.endpoint.projectId')) {
+    opts.projectId = config.get('persistor.gcs.endpoint.projectId')
+  }
+  return new Storage(opts)
+}
 
 const core = require('overleaf-editor-core')
 const projectKey = require('@overleaf/object-persistor/src/ProjectKey.js')
 const streams = require('../lib/streams')
-const ProjectArchive = require('../lib/project_archive')
 
 const {
   values: { verbose: VERBOSE },
@@ -53,7 +66,7 @@ if (HISTORY_IDS.length === 0) {
 
 async function listDeletedChunks(historyId) {
   const bucketName = config.get('chunkStore.bucket')
-  const storage = new Storage()
+  const storage = createStorage()
   const [files] = await storage.bucket(bucketName).getFiles({
     prefix: projectKey.format(historyId),
     versions: true,
@@ -137,7 +150,7 @@ class RecoveryBlobStore {
     if (VERBOSE) console.log('fetching blob', hash)
 
     const bucketName = config.get('blobStore.projectBucket')
-    const storage = new Storage()
+    const storage = createStorage()
     const [files] = await storage.bucket(bucketName).getFiles({
       prefix: this.makeProjectBlobKey(hash),
       versions: true,
@@ -158,7 +171,7 @@ class RecoveryBlobStore {
 
   async fetchGlobalBlob(hash, destination) {
     const bucketName = config.get('blobStore.globalBucket')
-    const storage = new Storage()
+    const storage = createStorage()
     const file = storage.bucket(bucketName).file(this.makeGlobalBlobKey(hash))
     await file.download({ destination })
   }
@@ -203,9 +216,18 @@ class RecoveryBlobStore {
 async function uploadZip(historyId, zipPathname) {
   const bucketName = config.get('zipStore.bucket')
   const deadline = 24 * 3600 * 1000 // lifecycle limit on the zips bucket
-  const storage = new Storage()
+  const storage = createStorage()
   const destination = `${historyId}-recovered.zip`
-  await storage.bucket(bucketName).upload(zipPathname, { destination })
+  await storage.bucket(bucketName).upload(zipPathname, {
+    destination,
+    resumable: false,
+  })
+
+  if (config.has('persistor.gcs.endpoint.apiEndpoint')) {
+    // In emulator mode, signed URLs aren't available
+    const apiEndpoint = config.get('persistor.gcs.endpoint.apiEndpoint')
+    return `${apiEndpoint}/storage/v1/b/${bucketName}/o/${encodeURIComponent(destination)}?alt=media`
+  }
 
   const signedUrls = await storage
     .bucket(bucketName)
@@ -219,6 +241,23 @@ async function uploadZip(historyId, zipPathname) {
   return signedUrls[0]
 }
 
+/**
+ * Promisified wrapper for ZipStream's entry method.
+ *
+ * @param {ZipStream} archive
+ * @param {Buffer|NodeJS.ReadableStream|string} source
+ * @param {{ name: string }} data
+ * @return {Promise<void>}
+ */
+function addEntry(archive, source, data) {
+  return new Promise((resolve, reject) => {
+    archive.entry(source, data, err => {
+      if (err) reject(err)
+      else resolve()
+    })
+  })
+}
+
 async function restoreProject(historyId) {
   const tmp = await fs.promises.mkdtemp(
     path.join(os.tmpdir(), historyId.toString())
@@ -237,9 +276,40 @@ async function restoreProject(historyId) {
   if (VERBOSE) console.log('zipping', historyId)
 
   const zipPathname = path.join(tmp, `${historyId}.zip`)
-  const zipTimeoutMs = 60 * 1000
-  const archive = new ProjectArchive(snapshot, zipTimeoutMs)
-  await archive.writeZip(blobStore, zipPathname)
+  const outputFile = fs.createWriteStream(zipPathname)
+  const archive = new ZipStream()
+
+  const pipelinePromise = pipeline(archive, outputFile)
+
+  for (const pathname of snapshot.getFilePathnames()) {
+    const file = snapshot.getFile(pathname)
+    if (!file) continue
+
+    await file.load('eager', blobStore)
+    let content = file.getContent({
+      filterTrackedDeletes: true,
+    })
+
+    if (content === null) {
+      const hash = file.getHash()
+      content = await blobStore.getStream(hash)
+    }
+
+    if (content == null) continue
+
+    if (typeof content === 'string') {
+      content = Buffer.from(content)
+    }
+    await addEntry(archive, content, { name: pathname })
+    if (VERBOSE) console.log(`${pathname} added`)
+  }
+
+  archive.finalize()
+  await pipelinePromise
+
+  if (VERBOSE) {
+    console.log(`Wrote ${archive.getBytesWritten()} bytes`)
+  }
 
   if (VERBOSE) console.log('uploading', historyId)
 
@@ -252,4 +322,7 @@ async function main() {
     console.log(signedUrl)
   }
 }
-main().catch(console.error)
+main().catch(err => {
+  console.error(err)
+  process.exit(1)
+})

services/history-v1/storage/scripts/recover_zip_from_backup.mjs

@@ -4,6 +4,7 @@ import commandLineArgs from 'command-line-args'
 import assert from '../lib/assert.js'
 import fs from 'node:fs'
 import { setTimeout } from 'node:timers/promises'
+import { pipeline } from 'node:stream/promises'
 import {
   archiveLatestChunk,
   archiveRawProject,
@@ -11,17 +12,13 @@ import {
 } from '../lib/backupArchiver.mjs'
 import knex from '../lib/knex.js'
 import { client } from '../lib/mongodb.js'
-import archiver from 'archiver'
-import Events from 'node:events'
+import ZipStream from 'zip-stream'
 import { Chunk } from 'overleaf-editor-core'
 import _ from 'lodash'
 
-// Silence warning.
-Events.setMaxListeners(20)
-
 const SUPPORTED_MODES = ['raw', 'latest']
 
-// Pads the mode name to a fixed length for better alignment in output.
+// Pads the mode name to a fixed length for alignment.
 const padModeName = _.partialRight(
   _.padEnd,
   Math.max(...SUPPORTED_MODES.map(mode => mode.length))
@@ -65,25 +62,6 @@ function usage() {
   })
 }
 
-/**
- * @typedef {import('archiver').ZipArchive} ZipArchive
- */
-
-/**
- * @typedef {import('archiver').ProgressData} ProgressData
- */
-
-/**
- * @typedef {import('archiver').EntryData} EntryData
- */
-
-/**
- * @typedef {Object} ArchiverError
- * @property {string} message
- * @property {string} code
- * @property {Object} data
- */
-
 let historyId, help, mode, output, useBackupGlobalBlobs, verbose
 
 try {
@@ -136,80 +114,37 @@ await loadGlobalBlobs()
 
 outputFile = fs.createWriteStream(output)
 
-const archive = archiver.create('zip', {})
-
-archive.on('close', function () {
-  console.log(archive.pointer() + ' total bytes')
-  console.log(`Wrote ${output}`)
-  shutdown().catch(e => console.error('Error shutting down', e))
-})
-
-archive.on(
-  'error',
-  /**
-   *
-   * @param {ArchiverError} e
-   */
-  function (e) {
-    console.error(`Error writing archive: ${e.message}`)
-  }
-)
+const archive = new ZipStream()
 
-archive.on('end', function () {
-  console.log(`Wrote ${archive.pointer()} total bytes to ${output}`)
-  shutdown().catch(e => console.error('Error shutting down', e))
+archive.on('error', function (e) {
+  console.error(`Error writing archive: ${e.message}`)
 })
 
-archive.on(
-  'progress',
-  /**
-   *
-   * @param {ProgressData} progress
-   */
-  function (progress) {
-    if (verbose) {
-      console.log(
-        `${progress.entries.processed} processed out of ${progress.entries.total}`
-      )
-    }
-  }
-)
-
-archive.on(
-  'entry',
-  /**
-   *
-   * @param {EntryData} entry
-   */
-  function (entry) {
-    if (verbose) {
-      console.log(`${entry.name} added`)
-    }
-  }
-)
-
-archive.on(
-  'warning',
-  /**
-   *
-   * @param {ArchiverError} warning
-   */
-  function (warning) {
-    console.warn(`Warning encountered when writing archive: ${warning.message}`)
-  }
-)
-
 try {
+  // Pipe archive to the output file before adding entries.
+  // pipeline handles backpressure and will resolve when
+  // the archive stream ends.
+  const pipelinePromise = pipeline(archive, outputFile)
+
   switch (mode) {
     case 'latest':
-      await archiveLatestChunk(archive, historyId, useBackupGlobalBlobs)
+      await archiveLatestChunk(
+        archive,
+        historyId,
+        useBackupGlobalBlobs,
+        verbose
+      )
       break
     case 'raw':
     default:
-      await archiveRawProject(archive, historyId, useBackupGlobalBlobs)
+      await archiveRawProject(archive, historyId, useBackupGlobalBlobs, verbose)
       break
   }
-  archive.pipe(outputFile)
+
+  archive.finalize()
+  await pipelinePromise
+
+  console.log(`Wrote ${archive.getBytesWritten()} total bytes to ${output}`)
 } catch (error) {
   if (error instanceof BackupPersistorError) {
     console.error(error.message)
@@ -222,12 +157,7 @@ try {
   } else {
     console.error('Error encountered when writing archive')
   }
-} finally {
-  await Promise.race([
-    await archive.finalize(),
-    setTimeout(10000).then(() => {
-      console.error('Archive did not finalize in time')
-      return shutdown(1)
-    }),
-  ])
+  await shutdown(1)
 }
+
+await shutdown(0)

services/history-v1/test/acceptance/js/storage/backup.test.mjs

@@ -647,6 +647,68 @@ describe('backup script', function () {
       expect(newBackupStatus.backupStatus.lastBackedUpVersion).to.equal(50) // backup fails on final chunk
       expect(newBackupStatus.currentEndVersion).to.equal(54) // backup is incomplete due to missing blob
     })
+
+    it('can recover zip file from backup in raw mode', async function () {
+      // First, run backup so data is available
+      await runBackupScript(['--projectId', projectId])
+
+      const zipPath = `/tmp/test-recover-raw-${historyId}.zip`
+      try {
+        await runRecoverZipFromBackupScript([
+          '--historyId',
+          historyId,
+          '--output',
+          zipPath,
+          '--mode=raw',
+        ])
+
+        // Verify the zip file is valid
+        const { stdout } = await promisify(execFile)('unzip', ['-l', zipPath], {
+          encoding: 'utf-8',
+        })
+
+        // Raw mode includes chunk and blob keys
+        // Verify chunks are present
+        expect(stdout).to.include('chunk.json')
+
+        // Verify blob hashes are present (hashes are stored as {hash[0:2]}/{hash[2:]})
+        expect(stdout).to.include(testFiles.GRAPH_PNG_HASH.slice(2))
+        expect(stdout).to.include(testFiles.NON_BMP_TXT_HASH.slice(2))
+      } finally {
+        await fs.promises.unlink(zipPath).catch(() => {})
+      }
+    })
+
+    it('can recover zip file from backup in latest mode', async function () {
+      // First, run backup so data is available
+      await runBackupScript(['--projectId', projectId])
+
+      const zipPath = `/tmp/test-recover-latest-${historyId}.zip`
+      try {
+        await runRecoverZipFromBackupScript([
+          '--historyId',
+          historyId,
+          '--output',
+          zipPath,
+          '--mode=latest',
+        ])
+
+        // Verify the zip file is valid
+        const { stdout } = await promisify(execFile)('unzip', ['-l', zipPath], {
+          encoding: 'utf-8',
+        })
+
+        // Latest mode includes the project files
+        expect(stdout).to.include('main.tex')
+        expect(stdout).to.include('chapter1.tex')
+        expect(stdout).to.include('chapter2.tex')
+        expect(stdout).to.include('bibliography.bib')
+        expect(stdout).to.include('graph.png')
+        expect(stdout).to.include('unicodeFile.tex')
+      } finally {
+        await fs.promises.unlink(zipPath).catch(() => {})
+      }
+    })
   })
 })
 
@@ -683,3 +745,37 @@ async function runBackupScript(args) {
   }
   return result
 }
+
+/**
+ * Run the recover_zip_from_backup script with given arguments
+ * @param {string[]} args
+ */
+async function runRecoverZipFromBackupScript(args) {
+  const TIMEOUT = 30 * 1000
+  let result
+  try {
+    result = await promisify(execFile)(
+      'node',
+      ['storage/scripts/recover_zip_from_backup.mjs', ...args],
+      {
+        encoding: 'utf-8',
+        timeout: TIMEOUT,
+        env: {
+          ...process.env,
+          LOG_LEVEL: 'debug',
+        },
+      }
+    )
+    result.status = 0
+  } catch (err) {
+    const { stdout, stderr, code } = err
+    if (typeof code !== 'number') {
+      console.log(err)
+    }
+    result = { stdout, stderr, status: code }
+  }
+  if (result.status !== 0) {
+    throw new Error(`recover_zip_from_backup failed: ${result.stderr}`)
+  }
+  return result
+}

services/history-v1/test/acceptance/js/storage/recover_zip.test.mjs

@@ -0,0 +1,169 @@
+import { expect } from 'chai'
+import config from 'config'
+import { execFile } from 'node:child_process'
+import fs from 'node:fs'
+import { promisify } from 'node:util'
+import { Change, Operation, File, TextOperation } from 'overleaf-editor-core'
+// We depend on this via object-persistor.
+// eslint-disable-next-line import/no-extraneous-dependencies
+import { Storage } from '@google-cloud/storage'
+import {
+  loadGlobalBlobs,
+  BlobStore,
+} from '../../../../storage/lib/blob_store/index.js'
+import ChunkStore from '../../../../storage/lib/chunk_store/index.js'
+import persistChanges from '../../../../storage/lib/persist_changes.js'
+import testFiles from '../storage/support/test_files.js'
+import cleanup from './support/cleanup.js'
+import { getZipEntries } from './support/unzip.js'
+
+describe('recover_zip script', function () {
+  let projectId
+  let limitsToPersistImmediately
+
+  before(async function () {
+    const farFuture = new Date()
+    farFuture.setTime(farFuture.getTime() + 7 * 24 * 3600 * 1000)
+    limitsToPersistImmediately = {
+      minChangeTimestamp: farFuture,
+      maxChangeTimestamp: farFuture,
+      maxChanges: 10,
+      maxChunkChanges: 10,
+    }
+
+    const gcsEndpoint = config.get('persistor.gcs.endpoint')
+    const storage = new Storage({
+      apiEndpoint: gcsEndpoint.apiEndpoint,
+      projectId: gcsEndpoint.projectId,
+    })
+    const bucketName = config.get('zipStore.bucket')
+    try {
+      const [exists] = await storage.bucket(bucketName).exists()
+      if (!exists) {
+        await storage.createBucket(bucketName)
+      }
+    } catch (err) {
+      if (err.code !== 409) throw err
+    }
+  })
+
+  beforeEach(cleanup.everything)
+
+  beforeEach(async function () {
+    await loadGlobalBlobs()
+    projectId = '123'
+
+    // Initialize the project in the chunk store
+    await ChunkStore.initializeProject(projectId)
+
+    const blobStore = new BlobStore(projectId)
+
+    // Upload binary file blob
+    await blobStore.putFile(testFiles.path('graph.png'))
+
+    // Create initial snapshot with text and binary files
+    const addMainTex = Operation.addFile(
+      'main.tex',
+      File.fromString('hello world')
+    )
+    const addGraphPng = Operation.addFile(
+      'graph.png',
+      File.fromHash(testFiles.GRAPH_PNG_HASH)
+    )
+    const change1 = new Change([addMainTex, addGraphPng], new Date(), [])
+    await persistChanges(projectId, [change1], limitsToPersistImmediately, 0)
+
+    // Add a text edit
+    const textOp = TextOperation.fromJSON({
+      textOperation: ['hello world'.length, ' more'],
+    })
+    const editOp = Operation.editFile('main.tex', textOp)
+    const change2 = new Change([editOp], new Date(), [])
+    await persistChanges(projectId, [change2], limitsToPersistImmediately, 1)
+  })
+
+  it('creates a valid zip from GCS data', async function () {
+    this.timeout(30 * 1000)
+
+    const zipPath = `/tmp/test-recover-zip-${projectId}.zip`
+    try {
+      const { stdout } = await runRecoverZipScript([projectId])
+
+      // The script logs the signed URL to stdout
+      const urlMatch = stdout.match(/(https?:\/\/[^\s]+)/)
+      expect(urlMatch).to.not.be.null
+      const signedUrl = urlMatch[1]
+
+      // Download the zip via fetch
+      const res = await fetch(signedUrl)
+      expect(res.ok).to.be.true
+      const buffer = await res.arrayBuffer()
+      await fs.promises.writeFile(zipPath, Buffer.from(buffer))
+
+      const zipEntries = await getZipEntries(zipPath)
+      const fileNames = zipEntries.map(e => e.fileName).sort()
+
+      expect(fileNames).to.deep.equal(['graph.png', 'main.tex'])
+
+      // Verify text content size (after edit)
+      const mainTexEntry = zipEntries.find(e => e.fileName === 'main.tex')
+      expect(mainTexEntry.uncompressedSize).to.equal('hello world more'.length)
+
+      // Verify binary content size
+      const graphEntry = zipEntries.find(e => e.fileName === 'graph.png')
+      expect(graphEntry.uncompressedSize).to.equal(
+        testFiles.GRAPH_PNG_BYTE_LENGTH
+      )
+    } finally {
+      await fs.promises.unlink(zipPath).catch(() => {})
+    }
+  })
+
+  it('supports the --verbose flag', async function () {
+    this.timeout(30 * 1000)
+
+    const { stdout } = await runRecoverZipScript(['--verbose', projectId])
+
+    // Verbose mode logs each file as it's added
+    expect(stdout).to.include('main.tex added')
+    expect(stdout).to.include('graph.png added')
+  })
+})
+
+/**
+ * Run the recover_zip.js script with given arguments
+ * @param {string[]} args
+ */
+async function runRecoverZipScript(args) {
+  const TIMEOUT = 30 * 1000
+  let result
+  try {
+    result = await promisify(execFile)(
+      'node',
+      ['storage/scripts/recover_zip.js', ...args],
+      {
+        encoding: 'utf-8',
+        timeout: TIMEOUT,
+        env: {
+          ...process.env,
+          LOG_LEVEL: 'debug',
+        },
+      }
+    )
+    result.status = 0
+  } catch (err) {
+    const { stdout, stderr, code } = err
+    if (typeof code !== 'number') {
+      console.log(err)
+    }
+    result = { stdout, stderr, status: code }
+  }
+  if (result.status !== 0 || result.stderr) {
+    throw new Error(
+      `recover_zip failed (exit ${result.status}):\n` +
+        `stdout: ${result.stdout}\n` +
+        `stderr: ${result.stderr}`
+    )
+  }
+  return result
+}

yarn.lock

@@ -15123,6 +15123,19 @@ __metadata:
   languageName: node
   linkType: hard
 
+"compress-commons@npm:^7.0.0":
+  version: 7.0.1
+  resolution: "compress-commons@npm:7.0.1"
+  dependencies:
+    crc-32: "npm:^1.2.0"
+    crc32-stream: "npm:^7.0.1"
+    is-stream: "npm:^4.0.0"
+    normalize-path: "npm:^3.0.0"
+    readable-stream: "npm:^4.0.0"
+  checksum: 10c0/9837b9971c7e536f14b113178e944741a0cb76051db2c84ff23da2b6321f2f6da8ed922d713bb643987d4cd39eeaa537c73d9d7843f9eb74f8fe47afe93d0733
+  languageName: node
+  linkType: hard
+
 "compressible@npm:~2.0.16":
   version: 2.0.18
   resolution: "compressible@npm:2.0.18"
@@ -15550,6 +15563,16 @@ __metadata:
   languageName: node
   linkType: hard
 
+"crc32-stream@npm:^7.0.1":
+  version: 7.0.1
+  resolution: "crc32-stream@npm:7.0.1"
+  dependencies:
+    crc-32: "npm:^1.2.0"
+    readable-stream: "npm:^4.0.0"
+  checksum: 10c0/0d8d217ca4f328bba859a6bef8593028d24841bef2fe4dc44ae892c7cace4301ce6d5676b568642ddb0a709ee3f0f7af1d1ada4c9c399bc5a4ff9b93d5d1071d
+  languageName: node
+  linkType: hard
+
 "create-storybook@npm:10.3.5":
   version: 10.3.5
   resolution: "create-storybook@npm:10.3.5"
@@ -21520,6 +21543,13 @@ __metadata:
   languageName: node
   linkType: hard
 
+"is-stream@npm:^4.0.0":
+  version: 4.0.1
+  resolution: "is-stream@npm:4.0.1"
+  checksum: 10c0/2706c7f19b851327ba374687bc4a3940805e14ca496dc672b9629e744d143b1ad9c6f1b162dece81c7bfbc0f83b32b61ccc19ad2e05aad2dd7af347408f60c7f
+  languageName: node
+  linkType: hard
+
 "is-string@npm:^1.1.1":
   version: 1.1.1
   resolution: "is-string@npm:1.1.1"
@@ -25938,6 +25968,7 @@ __metadata:
     typescript: "npm:^5.0.4"
     utf-8-validate: "npm:^5.0.4"
     yauzl: "npm:^2.9.1"
+    zip-stream: "npm:^7.0.2"
   languageName: unknown
   linkType: soft
 
@@ -29037,6 +29068,19 @@ __metadata:
   languageName: node
   linkType: hard
 
+"readable-stream@npm:^4.0.0":
+  version: 4.7.0
+  resolution: "readable-stream@npm:4.7.0"
+  dependencies:
+    abort-controller: "npm:^3.0.0"
+    buffer: "npm:^6.0.3"
+    events: "npm:^3.3.0"
+    process: "npm:^0.11.10"
+    string_decoder: "npm:^1.3.0"
+  checksum: 10c0/fd86d068da21cfdb10f7a4479f2e47d9c0a9b0c862fc0c840a7e5360201580a55ac399c764b12a4f6fa291f8cee74d9c4b7562e0d53b3c4b2769f2c98155d957
+  languageName: node
+  linkType: hard
+
 "readdir-glob@npm:^1.1.2":
   version: 1.1.3
   resolution: "readdir-glob@npm:1.1.3"
@@ -31434,7 +31478,7 @@ __metadata:
   languageName: node
   linkType: hard
 
-"string_decoder@npm:^1.1.1":
+"string_decoder@npm:^1.1.1, string_decoder@npm:^1.3.0":
   version: 1.3.0
   resolution: "string_decoder@npm:1.3.0"
   dependencies:
@@ -34968,6 +35012,17 @@ __metadata:
   languageName: node
   linkType: hard
 
+"zip-stream@npm:^7.0.2":
+  version: 7.0.5
+  resolution: "zip-stream@npm:7.0.5"
+  dependencies:
+    compress-commons: "npm:^7.0.0"
+    normalize-path: "npm:^3.0.0"
+    readable-stream: "npm:^4.0.0"
+  checksum: 10c0/e1669e17031c3c7243cb9014eacfaa66f4cd2e0d613a57dbee9caf7122ae869f8b2ea2e5891b5d9eee2897060c01db048b9b2a544ba83e227f6c162905282e48
+  languageName: node
+  linkType: hard
+
 "zod-validation-error@npm:4.0.1":
   version: 4.0.1
   resolution: "zod-validation-error@npm:4.0.1"