feat: Add full-text page search to workspace (Closes #6)

apps/pdf-master/src/app/App.tsx

@@ -117,6 +117,19 @@ export function App() {
       }),
     [orderedDocuments, store.pageOrderByDocument, store.pages],
   );
+  const pageTextSearchIndex = useMemo(() => {
+    const index = new Map<string, string>();
+
+    for (const { pages } of groupedPages) {
+      for (const page of pages) {
+        if (page.textContent) {
+          index.set(page.id, page.textContent.toLowerCase());
+        }
+      }
+    }
+
+    return index;
+  }, [groupedPages]);
 
   const filteredGroups = useMemo(() => {
     const query = searchQuery.trim().toLowerCase();
@@ -132,7 +145,8 @@ export function App() {
           : pages.filter(
               (page) =>
                 page.label.toLowerCase().includes(query) ||
-                String(page.sourcePageIndex + 1).includes(query),
+                String(page.sourcePageIndex + 1).includes(query) ||
+                pageTextSearchIndex.get(page.id)?.includes(query),
             );
 
         if (!matchingPages.length) {
@@ -142,7 +156,7 @@ export function App() {
         return { document, pages: matchingPages };
       })
       .filter(Boolean) as typeof groupedPages;
-  }, [groupedPages, searchQuery]);
+  }, [groupedPages, pageTextSearchIndex, searchQuery]);
 
   const workspaceRevision = useMemo(
     () =>
@@ -671,7 +685,7 @@ export function App() {
                 <div className="p-4">
                   <EmptyState
                     title="No pages match this search"
-                    description="Try a page number, page label, or clear the search field to return to the full workspace."
+                    description="Try a page number, page label, text content, or clear the search field to return to the full workspace."
                   />
                 </div>
               )

apps/pdf-master/src/components/Toolbar/Toolbar.tsx

@@ -122,7 +122,7 @@ export function Toolbar({
               value={searchQuery}
               onChange={(event) => onSearchChange(event.target.value)}
               className="w-full border-0 bg-transparent p-0 text-sm text-[color:var(--pm-text-strong)] outline-none placeholder:text-[color:var(--pm-text-faint)]"
-              placeholder="Search pages"
+              placeholder="Search pages by text, label, or number"
             />
           </label>
 

apps/pdf-master/src/domain/commands.ts

@@ -56,6 +56,7 @@ export function addDocumentToWorkspace(
       height: page.height,
       rotation: 0,
       label: page.label,
+      textContent: page.textContent,
     };
     pages[page.id] = pageEntity;
     pageIds.push(page.id);
@@ -118,6 +119,7 @@ export function addDocumentToWorkspaceAtPosition(
       height: page.height,
       rotation: 0,
       label: page.label,
+      textContent: page.textContent,
     };
     pages[page.id] = pageEntity;
     pageIds.push(page.id);

apps/pdf-master/src/domain/types.ts

@@ -92,6 +92,7 @@ export interface PageEntity {
   height: number;
   rotation: number;
   label: string;
+  textContent?: string;
 }
 
 export interface SelectionState {
@@ -180,6 +181,7 @@ export interface IngestPagePayload {
   width: number;
   height: number;
   label: string;
+  textContent?: string;
 }
 
 export interface IngestDocumentPayload {

apps/pdf-master/src/services/pdfInspection.integration.test.ts

@@ -1,6 +1,6 @@
 import { describe, expect, it } from 'vitest';
 import { inspectPdfFile } from '@/services/pdfInspection';
-import { createFormPdfFile } from '@/test/pdfFixtures';
+import { createFormPdfFile, createPdfFile } from '@/test/pdfFixtures';
 
 describe('pdf inspection integration', () => {
   it('extracts page inventory and AcroForm metadata from a real PDF', async () => {
@@ -13,4 +13,21 @@ describe('pdf inspection integration', () => {
     expect(payload.hasForms).toBe(true);
     expect(payload.formFields.map((field) => field.name)).toEqual(['name', 'approved', 'status']);
   });
+
+  it('extracts text content from PDF pages', async () => {
+    const file = await createPdfFile('text-test.pdf', [
+      [400, 600], // Page 1 with text "Fixture 1"
+      [400, 600], // Page 2 with text "Fixture 2"
+    ]);
+
+    const payload = await inspectPdfFile(file, 'doc-text');
+
+    expect(payload.pageCount).toBe(2);
+    // Text extraction may vary depending on PDF library, but we should get some text
+    expect(payload.pages[0]?.textContent).toBeDefined();
+    expect(payload.pages[1]?.textContent).toBeDefined();
+    // The fixture draws "Fixture 1" and "Fixture 2" on pages
+    expect(payload.pages[0]?.textContent?.toLowerCase()).toContain('fixture');
+    expect(payload.pages[1]?.textContent?.toLowerCase()).toContain('fixture');
+  });
 });

apps/pdf-master/src/services/pdfInspection.ts

@@ -7,10 +7,15 @@ import {
   PDFTextField,
 } from 'pdf-lib';
 import type { PDFField } from 'pdf-lib';
+import * as pdfjs from 'pdfjs-dist/legacy/build/pdf.mjs';
 import type { FormFieldModel, IngestDocumentPayload } from '@/domain/types';
 
 export async function inspectPdfFile(file: File, documentId: string): Promise<IngestDocumentPayload> {
-  const pdf = await PDFDocument.load(await file.arrayBuffer(), { updateMetadata: false });
+  const bytes = new Uint8Array(await file.arrayBuffer());
+  const [pdf, pageTextContent] = await Promise.all([
+    PDFDocument.load(bytes.slice(), { updateMetadata: false }),
+    extractPageTextContent(bytes.slice()),
+  ]);
   const pages = pdf.getPages().map((page, index) => {
     const size = page.getSize();
     return {
@@ -19,6 +24,7 @@ export async function inspectPdfFile(file: File, documentId: string): Promise<In
       width: size.width,
       height: size.height,
       label: `Page ${index + 1}`,
+      textContent: pageTextContent[index],
     };
   });
 
@@ -47,6 +53,40 @@ export async function inspectPdfFile(file: File, documentId: string): Promise<In
   };
 }
 
+async function extractPageTextContent(bytes: Uint8Array): Promise<string[]> {
+  const loadingTask = pdfjs.getDocument({
+    data: bytes,
+    disableWorker: true,
+    stopAtErrors: false,
+  } as Parameters<typeof pdfjs.getDocument>[0] & { disableWorker: boolean });
+
+  const pdf = await loadingTask.promise;
+
+  try {
+    const textContent: string[] = [];
+
+    for (let index = 1; index <= pdf.numPages; index += 1) {
+      const page = await pdf.getPage(index);
+      try {
+        const content = await page.getTextContent();
+        textContent.push(
+          content.items
+            .map((item) => ('str' in item ? item.str : ''))
+            .join(' ')
+            .trim(),
+        );
+      } finally {
+        page.cleanup();
+      }
+    }
+
+    return textContent;
+  } finally {
+    await pdf.destroy();
+    loadingTask.destroy();
+  }
+}
+
 function readFormField(field: PDFField): FormFieldModel {
   if (field instanceof PDFTextField) {
     return buildField(field.getName(), 'text', field.getText() ?? '', undefined, field.isReadOnly(), field.isRequired());

apps/pdf-master/src/store/pdfStore.ts

@@ -481,6 +481,7 @@ export const usePdfStore = create<PdfStore>((set, get) => ({
           height: page.height,
           rotation: 0,
           label: page.label,
+          textContent: page.textContent,
         };
         newPageIds.push(page.id);
       }

static/utilities/pdf-master/app.html

@@ -5,7 +5,7 @@
     <link rel="icon" type="image/svg+xml" href="./favicon.svg" />
     <meta name="viewport" content="width=device-width, initial-scale=1.0" />
     <title>pdf-master</title>
-    <script type="module" crossorigin src="./assets/app-nlhTV1tO.js"></script>
+    <script type="module" crossorigin src="./assets/app-CugQOqfU.js"></script>
     <link rel="stylesheet" crossorigin href="./assets/app-Cf9cPR_H.css">
   </head>
   <body>