feat: experimental pdf support

We are using the [Google Vertex AI API](https://cloud.google.com/vertex-ai/docs/reference/rest).
2anki · Oct 20, 2024 · 7374194 · 7374194
1 parent 35d591b
commit 7374194
Show file tree

Hide file tree

Showing 13 changed files with 356 additions and 72 deletions.
diff --git a/package-lock.json b/package-lock.json
diff --git a/package.json b/package.json
@@ -31,6 +31,7 @@
   "license": "MIT",
   "dependencies": {
     "@2anki/csv-to-apkg": "^1.4.4",
+    "@google-cloud/vertexai": "^1.9.0",
     "@notionhq/client": "^2.2.13",
     "@sendgrid/mail": "^8.1.3",
     "aws-sdk": "^2.1502.0",

diff --git a/src/controllers/SettingsController/SettingsController.test.ts b/src/controllers/SettingsController/SettingsController.test.ts
@@ -1,4 +1,6 @@
-import SettingsService, { IServiceSettings } from '../../services/SettingsService';
+import SettingsService, {
+  IServiceSettings,
+} from '../../services/SettingsService';
 import SettingsController from './SettingsController';
 import { SettingsInitializer } from '../../data_layer/public/Settings';
 
@@ -7,69 +9,70 @@ class FakeSettingsService implements IServiceSettings {
     return Promise.resolve([]);
   }
   delete(owner: string, id: string): Promise<void> {
-    return Promise.resolve()
+    return Promise.resolve();
   }
 
   getById(id: string): Promise<SettingsInitializer> {
     return Promise.resolve({
-      object_id: "1",
-      owner: "1",
-      payload: "payload"
-    })
+      object_id: '1',
+      owner: '1',
+      payload: 'payload',
+    });
   }
 }
 
-describe("SettingsController", () => {
-
-  test("returns default settings for client", () => {
-    const settingsController = new SettingsController(new FakeSettingsService());
-    const defaultOptions = settingsController.getDefaultSettingsCardOptions('client');
-
-    expect(defaultOptions).toStrictEqual(
-      {
-        "add-notion-link": "false",
-        "use-notion-id": "true",
-        "all": "true",
-        "paragraph": "false",
-        "cherry": "false",
-        "avocado": "false",
-        "tags": "false",
-        "cloze": "true",
-        "markdown-nested-bullet-points": "false",
-        "enable-input": "false",
-        "basic-reversed": "false",
-        "reversed": "false",
-        "no-underline": "false",
-        "max-one-toggle-per-card": "true",
-        "remove-mp3-links": "true",
-        "perserve-newlines": "true"
-      }
+describe('SettingsController', () => {
+  test('returns default settings for client', () => {
+    const settingsController = new SettingsController(
+      new FakeSettingsService()
     );
-  })
+    const defaultOptions =
+      settingsController.getDefaultSettingsCardOptions('client');
 
-  test("returns default settings for server", () => {
-    const settingsController = new SettingsController(new FakeSettingsService());
-    const defaultOptions = settingsController.getDefaultSettingsCardOptions('server');
+    expect(defaultOptions).toStrictEqual({
+      'add-notion-link': 'false',
+      'use-notion-id': 'true',
+      all: 'true',
+      paragraph: 'false',
+      cherry: 'false',
+      avocado: 'false',
+      tags: 'false',
+      cloze: 'true',
+      'markdown-nested-bullet-points': 'false',
+      'enable-input': 'false',
+      'basic-reversed': 'false',
+      reversed: 'false',
+      'no-underline': 'false',
+      'max-one-toggle-per-card': 'true',
+      'remove-mp3-links': 'true',
+      'perserve-newlines': 'true',
+      'vertex-ai-pdf-questions': 'false',
+    });
+  });
 
-    expect(defaultOptions).toStrictEqual(
-      {
-        'add-notion-link': 'false',
-        'use-notion-id': 'true',
-        "all": 'true',
-        "paragraph": 'false',
-        "cherry": 'false',
-        "avocado": 'false',
-        "tags": 'true',
-        "cloze": 'true',
-        'enable-input': 'false',
-        'basic-reversed': 'false',
-        "reversed": 'false',
-        'no-underline': 'false',
-        'max-one-toggle-per-card': 'true',
-        'perserve-newlines': 'false',
-        'page-emoji': 'first-emoji',
-      }
+  test('returns default settings for server', () => {
+    const settingsController = new SettingsController(
+      new FakeSettingsService()
     );
-  })
+    const defaultOptions =
+      settingsController.getDefaultSettingsCardOptions('server');
 
-})
+    expect(defaultOptions).toStrictEqual({
+      'add-notion-link': 'false',
+      'use-notion-id': 'true',
+      all: 'true',
+      paragraph: 'false',
+      cherry: 'false',
+      avocado: 'false',
+      tags: 'true',
+      cloze: 'true',
+      'enable-input': 'false',
+      'basic-reversed': 'false',
+      reversed: 'false',
+      'no-underline': 'false',
+      'max-one-toggle-per-card': 'true',
+      'perserve-newlines': 'false',
+      'page-emoji': 'first-emoji',
+    });
+  });
+});
diff --git a/src/controllers/SettingsController/supportedOptions.ts b/src/controllers/SettingsController/supportedOptions.ts
@@ -118,6 +118,12 @@ const supportedOptions = (): CardOption[] => {
       'Enable conversion of bullet and sub bullet points in Markdown. If you are a Obsidian user, enable this',
       false
     ),
+    new CardOption(
+      'vertex-ai-pdf-questions',
+      'Generate Questions from Single PDF File Uploads',
+      'Use Vertex AI API to generate questions from PDFs. This is a paid feature and if enabled will send your notes to Google Cloud.',
+      false
+    ),
   ];
 
   return v.filter(Boolean);

diff --git a/src/lib/anki/zip.tsx b/src/lib/anki/zip.tsx
@@ -2,7 +2,7 @@ import { strFromU8, unzipSync } from 'fflate';
 import { Body } from 'aws-sdk/clients/s3';
 import { renderToStaticMarkup } from 'react-dom/server';
 import { getUploadLimits } from '../misc/getUploadLimits';
-import { isHTMLFile, isMarkdownFile } from '../storage/checks';
+import { isHTMLFile, isMarkdownFile, isPDFFile } from '../storage/checks';
 
 interface File {
   name: string;
@@ -46,6 +46,14 @@ class ZipHandler {
     for (const name of this.fileNames) {
       const file = loadedZip[name];
       let contents = file;
+
+      /**
+       * For now disable batch processing of PDF files. We only want single uploads to avoid creating too many requests.
+       */
+      if (name.includes('__MACOSX/') || isPDFFile(name)) {
+        continue;
+      }
+
       if ((isHTMLFile(name) || isMarkdownFile(name)) && contents) {
         this.files.push({ name, contents: strFromU8(file) });
       } else if (contents) {