From 45dc8e21c50c8240ea851bf678d95f18a8d7c1fa Mon Sep 17 00:00:00 2001
From: davidmurray <david.murray16@hotmail.com>
Date: Thu, 9 Nov 2023 11:02:58 -0500
Subject: [PATCH] uploads.socketRoutes.ts: Use node-stream-zip library to unzip
 the file

This does it using a stream rather than loading all files into RAM at once. Fixes #691
---
 packages/transition-backend/package.json      |  1 +
 .../src/api/uploads.socketRoutes.ts           | 23 ++++++++-----------
 yarn.lock                                     |  5 ++++
 3 files changed, 15 insertions(+), 14 deletions(-)

diff --git a/packages/transition-backend/package.json b/packages/transition-backend/package.json
index f32eb1ac7..25e8d4287 100644
--- a/packages/transition-backend/package.json
+++ b/packages/transition-backend/package.json
@@ -49,6 +49,7 @@
         "lodash": "^4.17.21",
         "moment": "^2.29.4",
         "morgan": "^1.10.0",
+        "node-stream-zip": "^1.15.0",
         "p-queue": "^6.6.2",
         "papaparse": "^5.3.1",
         "pbf": "^3.2.1",
diff --git a/packages/transition-backend/src/api/uploads.socketRoutes.ts b/packages/transition-backend/src/api/uploads.socketRoutes.ts
index 8d55f5524..40b4cfb63 100644
--- a/packages/transition-backend/src/api/uploads.socketRoutes.ts
+++ b/packages/transition-backend/src/api/uploads.socketRoutes.ts
@@ -4,8 +4,7 @@
  * This file is licensed under the MIT License.
  * License text available at https://opensource.org/licenses/MIT
  */
-import fs from 'fs';
-import JSZip from 'jszip';
+import StreamZip from 'node-stream-zip';
 import SocketIO from 'socket.io';
 
 import { directoryManager } from 'chaire-lib-backend/lib/utils/filesystem/directoryManager';
@@ -29,19 +28,15 @@ const gtfsImportFunction = async (socket: SocketIO.Socket, absoluteUserDir: stri
 
     // TODO: Consider moving to an `extract` method if this is needed anywhere else
     try {
-        const zipData = fs.readFileSync(filePath);
-        const zip = new JSZip();
-        const zipFileContent = await zip.loadAsync(zipData);
-        const filePromises = Object.keys(zipFileContent.files).map(async (filename) => {
-            const fileInfo = zip.file(filename);
-            if (fileInfo === null) {
-                return;
-            }
-            const content = await fileInfo.async('nodebuffer');
-            const dest = gtfsFilesDirectoryPath + filename;
-            fs.writeFileSync(dest, content);
+        const zip = new StreamZip.async({ file: filePath });
+        zip.on('extract', (entry, file) => {
+            console.log(`GTFS import: extracted ${entry.name} to ${file}`);
         });
-        Promise.all(filePromises);
+        // This extracts in a stream-based manner so the data is never full loaded in RAM.
+        const count = await zip.extract(null, gtfsFilesDirectoryPath);
+        console.log(`GTFS import: extracted ${count} entries`);
+
+        await zip.close();
         console.log('GTFS zip file upload Complete.');
         socket.emit('gtfsImporter.gtfsFileUnzipped');
 
diff --git a/yarn.lock b/yarn.lock
index 5a2ab204e..9c62370d5 100644
--- a/yarn.lock
+++ b/yarn.lock
@@ -10631,6 +10631,11 @@ node-releases@^2.0.1:
   resolved "https://registry.yarnpkg.com/node-releases/-/node-releases-2.0.1.tgz#3d1d395f204f1f2f29a54358b9fb678765ad2fc5"
   integrity sha512-CqyzN6z7Q6aMeF/ktcMVTzhAHCEpf8SOarwpzpf8pNBY2k5/oM34UHldUwp8VKI7uxct2HxSRdJjBaZeESzcxA==
 
+node-stream-zip@^1.15.0:
+  version "1.15.0"
+  resolved "https://registry.yarnpkg.com/node-stream-zip/-/node-stream-zip-1.15.0.tgz#158adb88ed8004c6c49a396b50a6a5de3bca33ea"
+  integrity sha512-LN4fydt9TqhZhThkZIVQnF9cwjU3qmUH9h78Mx/K7d3VvfRqqwthLwJEUOEL0QPZ0XQmNN7be5Ggit5+4dq3Bw==
+
 nodemailer-mock@^1.5.8:
   version "1.5.8"
   resolved "https://registry.yarnpkg.com/nodemailer-mock/-/nodemailer-mock-1.5.8.tgz#22e3d54db735f426fc590a0f975ded3a8cbd6b12"