@link-assistant/web-capture 2.0.0 → 2.0.2

package/tests/unit/cli.test.js

@@ -1,399 +0,0 @@
-// Unit tests for CLI argument parsing and functionality
-import { spawn } from 'child_process';
-import http from 'node:http';
-import { mkdtempSync, readFileSync, rmSync, writeFileSync } from 'fs';
-import { fileURLToPath } from 'url';
-import { dirname, join, resolve } from 'path';
-import packageJson from '../../package.json' with { type: 'json' };
-
-const __filename = fileURLToPath(import.meta.url);
-const __dirname = dirname(__filename);
-const cliPath = resolve(__dirname, '../../bin/web-capture.js');
-const searchFixturePath = resolve(__dirname, '../helpers/search-fixture.cjs');
-
-// Helper function to run CLI and capture output
-function runCli(args, options = {}) {
-  return new Promise((resolve, reject) => {
-    const proc = spawn(process.execPath, [cliPath, ...args], {
-      cwd: options.cwd || dirname(cliPath),
-      env: { ...process.env, ...options.env },
-    });
-
-    let stdout = '';
-    let stderr = '';
-
-    proc.stdout.on('data', (data) => {
-      stdout += data.toString();
-    });
-    proc.stderr.on('data', (data) => {
-      stderr += data.toString();
-    });
-
-    proc.on('close', (code) => {
-      resolve({ code, stdout, stderr });
-    });
-
-    proc.on('error', reject);
-  });
-}
-
-function startFixtureServer({
-  body = '<!doctype html><html><body><h1>Issue 68</h1><p>Captured from positional URL.</p></body></html>',
-  contentType = 'text/html; charset=utf-8',
-} = {}) {
-  return new Promise((resolve, reject) => {
-    const server = http.createServer((req, res) => {
-      res.writeHead(200, {
-        'Content-Type': contentType,
-        'Content-Length': Buffer.byteLength(body),
-      });
-      res.end(body);
-    });
-
-    server.on('error', reject);
-    server.listen(0, '127.0.0.1', () => {
-      const { port } = server.address();
-      resolve({
-        server,
-        url: `http://127.0.0.1:${port}/article`,
-      });
-    });
-  });
-}
-
-function stopFixtureServer(server) {
-  return new Promise((resolve, reject) => {
-    server.close((err) => {
-      if (err) {
-        reject(err);
-        return;
-      }
-      resolve();
-    });
-  });
-}
-
-describe('CLI', () => {
-  describe('--help', () => {
-    test('shows help message with --help flag', async () => {
-      const result = await runCli(['--help']);
-      expect(result.code).toBe(0);
-      expect(result.stdout).toContain('web-capture');
-      expect(result.stdout).toContain('Usage:');
-      expect(result.stdout).toContain('--serve');
-      expect(result.stdout).toContain('--format');
-      expect(result.stdout).toContain('--output');
-      expect(result.stdout).toContain('--engine');
-    });
-
-    test('shows help message with -h flag', async () => {
-      const result = await runCli(['-h']);
-      expect(result.code).toBe(0);
-      expect(result.stdout).toContain('web-capture');
-      expect(result.stdout).toContain('Usage:');
-    });
-
-    test('shows error and exits with code 1 when no arguments provided', async () => {
-      const result = await runCli([]);
-      expect(result.code).toBe(1);
-      // With lino-arguments and yargs, error messages go to stderr
-      expect(result.stderr).toContain('Error');
-    });
-  });
-
-  describe('--version', () => {
-    test('shows version with --version flag', async () => {
-      const result = await runCli(['--version']);
-      expect(result.code).toBe(0);
-      // yargs outputs just the version number
-      expect(result.stdout).toMatch(/\d+\.\d+\.\d+/);
-    });
-
-    test('shows version with -v flag', async () => {
-      const result = await runCli(['-v']);
-      expect(result.code).toBe(0);
-      // yargs outputs just the version number
-      expect(result.stdout).toMatch(/\d+\.\d+\.\d+/);
-    });
-
-    test('shows web-capture version when run from another npm project', async () => {
-      const callerProjectDir = mkdtempSync(join(process.cwd(), 'caller-'));
-      writeFileSync(
-        join(callerProjectDir, 'package.json'),
-        JSON.stringify({
-          name: 'caller-project',
-          version: '1.0.0',
-        })
-      );
-
-      try {
-        const result = await runCli(['--version'], { cwd: callerProjectDir });
-
-        expect(result.code).toBe(0);
-        expect(result.stdout.trim()).toBe(packageJson.version);
-      } finally {
-        rmSync(callerProjectDir, { recursive: true, force: true });
-      }
-    });
-  });
-
-  describe('URL validation', () => {
-    test('accepts a URL as a positional argument', async () => {
-      const { server, url } = await startFixtureServer();
-
-      try {
-        const result = await runCli([
-          url,
-          '--format',
-          'markdown',
-          '--output',
-          '-',
-        ]);
-
-        expect(result.code).toBe(0);
-        expect(result.stderr).not.toContain('Unknown argument');
-        expect(result.stdout).toContain('Issue 68');
-        expect(result.stdout).toContain('Captured from positional URL.');
-      } finally {
-        await stopFixtureServer(server);
-      }
-    }, 15000);
-
-    test('accepts a URL after the -- argument separator', async () => {
-      const { server, url } = await startFixtureServer();
-
-      try {
-        const result = await runCli([
-          '--format',
-          'markdown',
-          '--output',
-          '-',
-          '--',
-          url,
-        ]);
-
-        expect(result.code).toBe(0);
-        expect(result.stderr).not.toContain('Missing URL');
-        expect(result.stdout).toContain('Issue 68');
-      } finally {
-        await stopFixtureServer(server);
-      }
-    }, 15000);
-
-    test('rejects unknown options', async () => {
-      const result = await runCli([
-        '--unknown-option-for-issue-68',
-        'https://example.com',
-      ]);
-
-      expect(result.code).toBe(1);
-      expect(result.stderr).toContain('Unknown argument');
-    });
-
-    test('rejects invalid URL', async () => {
-      const result = await runCli(['http://[invalid-url']);
-      // Should fail before trying to capture.
-      expect(result.code).toBe(1);
-      expect(result.stderr).toContain('Invalid URL');
-    }, 15000);
-  });
-
-  describe('format options', () => {
-    test('accepts --format html', async () => {
-      // Just test that the argument is accepted (actual fetch would fail without network)
-      const result = await runCli(['--help']);
-      expect(result.stdout).toContain('--format');
-      expect(result.stdout).toContain('html');
-    });
-
-    test('accepts --format markdown', async () => {
-      const result = await runCli(['--help']);
-      expect(result.stdout).toContain('markdown');
-    });
-
-    test('accepts --format image', async () => {
-      const result = await runCli(['--help']);
-      expect(result.stdout).toContain('image');
-    });
-  });
-
-  describe('engine options', () => {
-    test('help shows engine options', async () => {
-      const result = await runCli(['--help']);
-      expect(result.stdout).toContain('--engine');
-      expect(result.stdout).toContain('puppeteer');
-      expect(result.stdout).toContain('playwright');
-    });
-
-    test('accepts --engine option', async () => {
-      const result = await runCli(['--help']);
-      expect(result.stdout).toContain('--engine');
-    });
-  });
-
-  describe('server mode options', () => {
-    test('help shows --serve option', async () => {
-      const result = await runCli(['--help']);
-      expect(result.stdout).toContain('--serve');
-      expect(result.stdout).toContain('-s');
-    });
-
-    test('help shows --port option', async () => {
-      const result = await runCli(['--help']);
-      expect(result.stdout).toContain('--port');
-      expect(result.stdout).toContain('-p');
-    });
-
-    test('help shows API endpoints', async () => {
-      const result = await runCli(['--help']);
-      expect(result.stdout).toContain('/html');
-      expect(result.stdout).toContain('/markdown');
-      expect(result.stdout).toContain('/image');
-      expect(result.stdout).not.toContain('/gdocs');
-    });
-  });
-
-  describe('output options', () => {
-    test('help shows --output option', async () => {
-      const result = await runCli(['--help']);
-      expect(result.stdout).toContain('--output');
-      expect(result.stdout).toContain('-o');
-    });
-  });
-
-  describe('FormalAI-compatible CLI contract', () => {
-    test('writes HTML, Markdown, and text captures to stdout with -o -', async () => {
-      const { server, url } = await startFixtureServer({
-        body: '<!doctype html><html><body><h1>FormalAI Fixture</h1><p>Stable CLI shape.</p></body></html>',
-      });
-
-      try {
-        const html = await runCli([url, '--format', 'html', '--output', '-']);
-        expect(html.code).toBe(0);
-        expect(html.stdout).toContain('<h1>FormalAI Fixture</h1>');
-
-        const markdown = await runCli([
-          url,
-          '--format',
-          'markdown',
-          '--output',
-          '-',
-        ]);
-        expect(markdown.code).toBe(0);
-        expect(markdown.stdout).toContain('FormalAI Fixture');
-        expect(markdown.stdout).toContain('Stable CLI shape.');
-
-        const text = await runCli([url, '--format', 'txt', '--output', '-']);
-        expect(text.code).toBe(0);
-        expect(text.stdout).toContain('FormalAI Fixture');
-      } finally {
-        await stopFixtureServer(server);
-      }
-    }, 20000);
-
-    test('writes archive captures as ZIP files for binary CLI consumers', async () => {
-      const outputDir = mkdtempSync(join(process.cwd(), 'formalai-cli-'));
-      const outputPath = join(outputDir, 'capture.zip');
-      const { server, url } = await startFixtureServer({
-        body: '<!doctype html><html><body><h1>FormalAI Archive</h1></body></html>',
-      });
-
-      try {
-        const result = await runCli([
-          url,
-          '--archive',
-          'zip',
-          '--output',
-          outputPath,
-        ]);
-
-        expect(result.code).toBe(0);
-        const zipBytes = readFileSync(outputPath);
-        expect(zipBytes.length).toBeGreaterThan(50);
-        expect(zipBytes[0]).toBe(0x50);
-        expect(zipBytes[1]).toBe(0x4b);
-      } finally {
-        await stopFixtureServer(server);
-        rmSync(outputDir, { recursive: true, force: true });
-      }
-    }, 20000);
-
-    test('emits normalized search JSON from the search subcommand', async () => {
-      const nodeOptions = [
-        process.env.NODE_OPTIONS,
-        `--require=${searchFixturePath}`,
-      ]
-        .filter(Boolean)
-        .join(' ');
-      const result = await runCli(
-        ['search', 'formal-ai', '--provider', 'wikipedia', '--limit', '1'],
-        {
-          env: { NODE_OPTIONS: nodeOptions },
-        }
-      );
-
-      expect(result.code).toBe(0);
-      const body = JSON.parse(result.stdout);
-      expect(body).toMatchObject({
-        query: 'formal-ai',
-        provider: 'wikipedia',
-        captureMode: 'fetch',
-      });
-      expect(body.results).toEqual([
-        {
-          rank: 1,
-          title: 'Formal methods',
-          url: 'https://en.wikipedia.org/wiki/Formal_methods',
-          snippet: 'the study of formal methods',
-        },
-      ]);
-      expect(body.diagnostics).toMatchObject({
-        status: 200,
-        blockedByCors: false,
-        blockedByCaptcha: false,
-      });
-      expect(body.diagnostics.sourceUrl).toContain('en.wikipedia.org');
-    }, 20000);
-
-    test('emits shared-dialog demo_memory from the shared-dialog subcommand', async () => {
-      const fixturePath = resolve(
-        __dirname,
-        '../../../docs/case-studies/issue-141/raw-data/chatgpt-share-6a3825b9.html'
-      );
-      const body = readFileSync(fixturePath, 'utf-8');
-      const { server, url } = await startFixtureServer({ body });
-      const shareUrl = url.replace(
-        '/article',
-        '/share/6a3825b9-8de4-83ee-9c24-52fd1eb38d24'
-      );
-
-      try {
-        const result = await runCli(
-          [
-            'shared-dialog',
-            shareUrl,
-            '--format',
-            'demo-memory',
-            '--output',
-            '-',
-          ],
-          {
-            cwd: resolve(__dirname, '../..'),
-          }
-        );
-
-        expect(result.code).toBe(0);
-        expect(result.stdout.startsWith('demo_memory')).toBe(true);
-        expect(result.stdout).toContain('demo_memory');
-        expect(result.stdout.match(/\n {2}event "/g)).toHaveLength(4);
-        expect(result.stdout).toContain('role "user"');
-        expect(result.stdout).toContain('role "assistant"');
-        expect(result.stdout).toContain(
-          'conversationTitle "Infinite loop script"'
-        );
-      } finally {
-        await stopFixtureServer(server);
-      }
-    }, 20000);
-  });
-});

package/tests/unit/convertGoogleDriveUrl.test.js

@@ -1,89 +0,0 @@
-import {
-  convertGoogleDriveUrl,
-  getGoogleDriveFileId,
-  googleDriveImageTextForUrl,
-  isGoogleDriveFileUrl,
-} from '../../src/lib.js';
-
-const FIRST_FILE_ID = '1Cxkx6-428EQAX0-eiaq66H829ohnPp7q';
-const SECOND_FILE_ID = '1fgJaftjv53xCN7vgiJOaQlbfWkUqPgyd';
-
-function directDownloadUrl(fileId) {
-  return `https://drive.google.com/uc?export=download&id=${fileId}`;
-}
-
-describe('Google Drive URL helpers', () => {
-  it.each([
-    [
-      'issue URL with /view',
-      `https://drive.google.com/file/d/${FIRST_FILE_ID}/view`,
-      FIRST_FILE_ID,
-    ],
-    [
-      'issue URL without /view',
-      `https://drive.google.com/file/d/${FIRST_FILE_ID}`,
-      FIRST_FILE_ID,
-    ],
-    [
-      'second issue URL with /view',
-      `https://drive.google.com/file/d/${SECOND_FILE_ID}/view`,
-      SECOND_FILE_ID,
-    ],
-    [
-      'second issue URL without /view',
-      `https://drive.google.com/file/d/${SECOND_FILE_ID}`,
-      SECOND_FILE_ID,
-    ],
-    [
-      'open?id URL',
-      `https://drive.google.com/open?id=${FIRST_FILE_ID}`,
-      FIRST_FILE_ID,
-    ],
-    [
-      'direct uc URL',
-      `https://drive.google.com/uc?export=download&id=${FIRST_FILE_ID}`,
-      FIRST_FILE_ID,
-    ],
-    [
-      'redirected usercontent URL',
-      `https://drive.usercontent.google.com/download?id=${FIRST_FILE_ID}&export=download`,
-      FIRST_FILE_ID,
-    ],
-  ])('extracts and converts %s', (_name, input, fileId) => {
-    expect(getGoogleDriveFileId(input)).toBe(fileId);
-    expect(isGoogleDriveFileUrl(input)).toBe(true);
-    expect(convertGoogleDriveUrl(input)).toBe(directDownloadUrl(fileId));
-  });
-
-  it('ignores non-Google Drive URLs and malformed Drive IDs', () => {
-    const example = 'https://example.com/image.jpg';
-    const malformed = 'https://drive.google.com/file/d/not%20valid/view';
-
-    expect(getGoogleDriveFileId(example)).toBeNull();
-    expect(isGoogleDriveFileUrl(example)).toBe(false);
-    expect(convertGoogleDriveUrl(example)).toBe(example);
-    expect(getGoogleDriveFileId(malformed)).toBeNull();
-    expect(isGoogleDriveFileUrl(malformed)).toBe(false);
-    expect(convertGoogleDriveUrl(malformed)).toBe(malformed);
-  });
-
-  it('keeps empty or missing inputs unchanged when converting', () => {
-    expect(convertGoogleDriveUrl(null)).toBeNull();
-    expect(convertGoogleDriveUrl(undefined)).toBeUndefined();
-    expect(convertGoogleDriveUrl('')).toBe('');
-  });
-
-  it('renders text output with the original file URL and direct image URL', () => {
-    const input = `https://drive.google.com/file/d/${FIRST_FILE_ID}/view`;
-
-    expect(googleDriveImageTextForUrl(input)).toContain(
-      `Original Google Drive file: ${input}`
-    );
-    expect(googleDriveImageTextForUrl(input)).toContain(
-      `Image: ${directDownloadUrl(FIRST_FILE_ID)}`
-    );
-    expect(
-      googleDriveImageTextForUrl('https://example.com/image.jpg')
-    ).toBeNull();
-  });
-});

package/tests/unit/convertRelativeUrls.test.js

@@ -1,34 +0,0 @@
-import { convertRelativeUrls } from '../../src/lib.js';
-
-describe('convertRelativeUrls', () => {
-  const baseUrl = 'https://example.com/';
-
-  it('should NOT inject runtimeHook if there are no <script> tags', () => {
-    const html = `<!DOCTYPE html><html><head><title>Test</title></head><body><h1>Hello</h1><p>World</p></body></html>`;
-    const result = convertRelativeUrls(html, baseUrl);
-    // Should not inject the runtimeHook
-    expect(result).not.toMatch(/MutationObserver/);
-    expect(result).not.toMatch(/fixAllUrls/);
-    expect(result).not.toMatch(/runtime JS hook/);
-    // Should not add any <script> tag
-    expect(result.match(/<script/gi)).toBeNull();
-  });
-
-  it('should inject runtimeHook if there is a <script> tag', () => {
-    const html = `<!DOCTYPE html><html><head><title>Test</title><script src="foo.js"></script></head><body><h1>Hello</h1></body></html>`;
-    const result = convertRelativeUrls(html, baseUrl);
-    // Should inject the runtimeHook
-    expect(result).toMatch(/MutationObserver/);
-    expect(result).toMatch(/fixAllUrls/);
-    // Should add a <script> tag for the hook
-    expect((result.match(/<script/gi) || []).length).toBeGreaterThan(1);
-  });
-
-  it('should inject runtimeHook if there is an inline <script>', () => {
-    const html = `<!DOCTYPE html><html><head><title>Test</title><script>console.log('hi');</script></head><body></body></html>`;
-    const result = convertRelativeUrls(html, baseUrl);
-    expect(result).toMatch(/MutationObserver/);
-    expect(result).toMatch(/fixAllUrls/);
-    expect((result.match(/<script/gi) || []).length).toBeGreaterThan(1);
-  });
-});