feat: add local OneNote and ONEPKG reader

src/onenote/onepkg/joplin-fixture.test.ts

@@ -0,0 +1,114 @@
+import { createHash } from 'node:crypto';
+import { readFileSync } from 'node:fs';
+import { resolve } from 'node:path';
+
+import { describe, expect, it } from 'vitest';
+
+import { parseOneNoteSection } from '../parser/parse-section.js';
+import { enumerateCabinet, extractCabinet } from './cabinet.js';
+import { openOneNotePackage } from './package.js';
+
+/*
+ * Joplin CLI test fixture, AGPL-3.0-or-later. See tests/fixtures/README.md.
+ * Pinned upstream revision: 1e73aad7eb08fde5d9e4cb533df40052a5cd32d7
+ */
+const encodedFixture = readFileSync(
+  resolve(process.cwd(), 'tests/fixtures/joplin-test.onepkg.base64'),
+  'utf8'
+);
+const JOPLIN_ONEPKG = Uint8Array.from(
+  Buffer.from(encodedFixture.replaceAll(/\s/gu, ''), 'base64')
+);
+
+const ONE_MAGIC = Uint8Array.from([
+  0xe4, 0x52, 0x5c, 0x7b, 0x8c, 0xd8, 0xa7, 0x4d, 0xae, 0xb1, 0x53, 0x78, 0xd0,
+  0x29, 0x96, 0xd3,
+]);
+
+const EXPECTED_PATHS = [
+  'Another section.one',
+  'Open Notebook.onetoc2',
+  'Section group/A.one',
+  'Section group/B.one',
+  'Section group/Open Notebook.onetoc2',
+  'Tést!.one',
+  '⅀⸨ Unicode ⸩.one',
+];
+
+describe('pinned Joplin .onepkg fixture', () => {
+  it('retains the reviewed upstream bytes', () => {
+    expect(JOPLIN_ONEPKG).toHaveLength(11_066);
+    expect(createHash('sha256').update(JOPLIN_ONEPKG).digest('hex')).toBe(
+      '83cb622d40f0ab127038b629553b6926be3e2a6ce642f169df0e2614c13e8469'
+    );
+  });
+
+  it('enumerates all seven UTF-8/nested entries using LZX 0x1203', () => {
+    const listing = enumerateCabinet(JOPLIN_ONEPKG);
+    expect(listing.entries.map((entry) => entry.normalizedPath)).toEqual(
+      EXPECTED_PATHS
+    );
+    expect(listing.folders).toHaveLength(1);
+    expect(listing.folders[0]).toMatchObject({
+      dataBlockCount: 3,
+      compression: { kind: 'lzx', raw: 0x1203, windowBits: 18 },
+    });
+    expect(
+      listing.diagnostics.filter(
+        (diagnostic) => diagnostic.code === 'cab-unmarked-utf8-name'
+      )
+    ).toHaveLength(2);
+  });
+
+  it('extracts every entry and preserves native .one headers', async () => {
+    const extraction = await extractCabinet(JOPLIN_ONEPKG);
+    expect(
+      extraction.extractedEntries.map((entry) => entry.normalizedPath)
+    ).toEqual(EXPECTED_PATHS);
+    const sections = extraction.extractedEntries.filter((entry) =>
+      entry.normalizedPath.toLowerCase().endsWith('.one')
+    );
+    expect(sections).toHaveLength(5);
+    for (const section of sections) {
+      expect(section.bytes.subarray(0, ONE_MAGIC.length)).toEqual(ONE_MAGIC);
+    }
+
+    const opened = await openOneNotePackage(JOPLIN_ONEPKG, {
+      sourceName: 'test.onepkg',
+    });
+    expect(opened.package.entries).toHaveLength(7);
+    expect(opened.package.sections).toHaveLength(5);
+    expect(
+      opened.package.sections.every(
+        (section) => section.parseStatus === 'not-requested'
+      )
+    ).toBe(true);
+  });
+
+  it('opens and parses all five packaged sections end to end', async () => {
+    const opened = await openOneNotePackage(JOPLIN_ONEPKG, {
+      sourceName: 'test.onepkg',
+      parseSection: ({ bytes, normalizedPath }) => ({
+        status: 'parsed',
+        value: parseOneNoteSection(bytes, { sourceName: normalizedPath }),
+      }),
+    });
+
+    expect(opened.package.sections).toHaveLength(5);
+    expect(
+      opened.package.sections.map((section) => section.parseStatus)
+    ).toEqual(['parsed', 'parsed', 'parsed', 'parsed', 'parsed']);
+    expect(
+      opened.package.entries.some((entry) => entry.parseStatus === 'failed')
+    ).toBe(false);
+
+    const testSection = opened.package.sections.find(
+      (section) => section.path === 'Tést!.one'
+    );
+    expect(testSection?.section?.sectionName).toBe('Tést!');
+    expect(testSection?.section?.pages[0]).toMatchObject({
+      title: 'Testing…',
+      text: 'Link to page: Page 2',
+    });
+  });
+});