mirror of
https://github.com/jackwener/OpenCLI.git
synced 2026-09-14 18:25:42 +08:00
314bc6f231
1. Remove root-level shim files (errors.ts, logger.ts, registry.ts, types.ts, utils.ts, launcher.ts) — update all ~840 adapter imports to reference src/ directly 2. Consolidate interceptor: reuse shared DISGUISE_FN in tap interceptor instead of reimplementing 3. Unify error classification: single ClassifiedError type with icon/exitCode/hint lookup table, eliminating duplicated pattern matching between resolveExitCode and renderError 4. Simplify cascade probe: replace repetitive switch cases with PROBE_OPTIONS lookup map 12. Clean up synthesize.ts: remove deprecated snake_case field aliases (recommended_args, recommended_columns, recommendedColumnsLegacy) and unnecessary constant aliases Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
118 lines
4.2 KiB
TypeScript
118 lines
4.2 KiB
TypeScript
import { describe, expect, it, vi } from 'vitest';
|
|
import type { IPage } from '../../src/types.js';
|
|
import {
|
|
extractJsonLd,
|
|
forceEnglishUrl,
|
|
formatDuration,
|
|
getCurrentImdbId,
|
|
isChallengePage,
|
|
normalizeImdbTitleType,
|
|
normalizeImdbId,
|
|
waitForImdbPath,
|
|
waitForImdbReviewsReady,
|
|
waitForImdbSearchReady,
|
|
} from './utils.js';
|
|
|
|
describe('normalizeImdbId', () => {
|
|
it('passes through bare ids', () => {
|
|
expect(normalizeImdbId('tt1375666', 'tt')).toBe('tt1375666');
|
|
expect(normalizeImdbId('nm0634240', 'nm')).toBe('nm0634240');
|
|
});
|
|
|
|
it('extracts ids from supported urls', () => {
|
|
expect(normalizeImdbId('https://www.imdb.com/title/tt1375666/', 'tt')).toBe('tt1375666');
|
|
expect(normalizeImdbId('https://m.imdb.com/title/tt1375666/', 'tt')).toBe('tt1375666');
|
|
expect(normalizeImdbId('https://www.imdb.com/de/title/tt1375666/?ref_=nv_sr_srsg_0', 'tt')).toBe('tt1375666');
|
|
expect(normalizeImdbId('https://www.imdb.com/name/nm0634240/', 'nm')).toBe('nm0634240');
|
|
});
|
|
|
|
it('throws on invalid or mismatched ids', () => {
|
|
expect(() => normalizeImdbId('invalid', 'tt')).toThrow('Invalid IMDb ID');
|
|
expect(() => normalizeImdbId('tt1', 'tt')).toThrow('Invalid IMDb ID');
|
|
expect(() => normalizeImdbId('nm0634240', 'tt')).toThrow('Invalid IMDb ID');
|
|
});
|
|
});
|
|
|
|
describe('formatDuration', () => {
|
|
it('converts ISO-8601 durations to a short human format', () => {
|
|
expect(formatDuration('PT2H28M')).toBe('2h 28m');
|
|
expect(formatDuration('PT1H')).toBe('1h');
|
|
expect(formatDuration('PT45M')).toBe('45m');
|
|
expect(formatDuration('PT2H')).toBe('2h');
|
|
});
|
|
|
|
it('returns an empty string for invalid input', () => {
|
|
expect(formatDuration('')).toBe('');
|
|
expect(formatDuration('invalid')).toBe('');
|
|
});
|
|
});
|
|
|
|
describe('forceEnglishUrl', () => {
|
|
it('adds the English language parameter', () => {
|
|
expect(forceEnglishUrl('https://www.imdb.com/title/tt1375666/')).toContain('language=en-US');
|
|
});
|
|
|
|
it('preserves existing query parameters', () => {
|
|
const result = forceEnglishUrl('https://www.imdb.com/title/tt1375666/?ref_=nv');
|
|
expect(result).toContain('language=en-US');
|
|
expect(result).toContain('ref_=nv');
|
|
});
|
|
});
|
|
|
|
describe('normalizeImdbTitleType', () => {
|
|
it('maps internal imdb ids to readable labels', () => {
|
|
expect(normalizeImdbTitleType({ id: 'movie', text: '' })).toBe('Movie');
|
|
expect(normalizeImdbTitleType({ id: 'tvSeries', text: '' })).toBe('TV Series');
|
|
expect(normalizeImdbTitleType('short')).toBe('Short');
|
|
});
|
|
|
|
it('preserves explicit text labels', () => {
|
|
expect(normalizeImdbTitleType({ id: 'movie', text: 'Feature Film' })).toBe('Feature Film');
|
|
});
|
|
});
|
|
|
|
describe('extractJsonLd', () => {
|
|
it('returns the evaluated JSON-LD payload', async () => {
|
|
const page = {
|
|
evaluate: vi.fn().mockResolvedValue({ '@type': 'Movie', name: 'Inception' }),
|
|
} as unknown as IPage;
|
|
|
|
await expect(extractJsonLd(page, 'Movie')).resolves.toEqual({ '@type': 'Movie', name: 'Inception' });
|
|
expect(page.evaluate).toHaveBeenCalledTimes(1);
|
|
expect(page.evaluate).toHaveBeenCalledWith(expect.stringContaining('"Movie"'));
|
|
});
|
|
});
|
|
|
|
describe('isChallengePage', () => {
|
|
it('returns true when the page evaluation matches a challenge', async () => {
|
|
const page = {
|
|
evaluate: vi.fn().mockResolvedValue(true),
|
|
} as unknown as IPage;
|
|
|
|
await expect(isChallengePage(page)).resolves.toBe(true);
|
|
expect(page.evaluate).toHaveBeenCalledTimes(1);
|
|
});
|
|
});
|
|
|
|
describe('imdb browser helpers', () => {
|
|
it('reads the current imdb id from page metadata', async () => {
|
|
const page = {
|
|
evaluate: vi.fn().mockResolvedValue('nm0634240'),
|
|
} as unknown as IPage;
|
|
|
|
await expect(getCurrentImdbId(page, 'nm')).resolves.toBe('nm0634240');
|
|
expect(page.evaluate).toHaveBeenCalledTimes(1);
|
|
});
|
|
|
|
it('wait helpers resolve mocked readiness booleans', async () => {
|
|
const page = {
|
|
evaluate: vi.fn().mockResolvedValue(true),
|
|
} as unknown as IPage;
|
|
|
|
await expect(waitForImdbPath(page, '^/find/?$')).resolves.toBe(true);
|
|
await expect(waitForImdbSearchReady(page)).resolves.toBe(true);
|
|
await expect(waitForImdbReviewsReady(page)).resolves.toBe(true);
|
|
expect(page.evaluate).toHaveBeenCalledTimes(3);
|
|
});
|
|
});
|