tonemark/src/tests/downloader.test.ts

import { afterEach, beforeEach, describe, expect, it, vi } from 'vitest';
import { rm } from 'fs/promises';
import type { TranscriptResponse } from 'youtube-transcript';

const { mockExecFile, mockFetchTranscript } = vi.hoisted(() => ({
	mockExecFile: vi.fn(),
	mockFetchTranscript: vi.fn()
}));

const TEST_DATA_DIR = `/tmp/tonemark-downloader-test-${Date.now()}`;
vi.stubEnv('DATA_DIR', TEST_DATA_DIR);

vi.mock('child_process', () => ({
	execFile: mockExecFile
}));

vi.mock('youtube-transcript', () => ({
	fetchTranscript: mockFetchTranscript
}));

import { downloadYouTube, transcriptEntriesToSegments, saveUploadedFile, getUploadPath, cleanupUploadDir } from '$lib/server/downloader.js';

beforeEach(() => {
	vi.clearAllMocks();
	mockExecFile.mockImplementation((...args: unknown[]) => {
		const cb = args.at(-1) as (...callbackArgs: unknown[]) => void;
		cb(null, JSON.stringify({ title: 'Fetched Title' }), '');
	});
});

afterEach(async () => {
	await rm(TEST_DATA_DIR, { recursive: true, force: true }).catch(() => {});
});

describe('transcriptEntriesToSegments', () => {
	it('converts millisecond transcript offsets into second-based segments', () => {
		const entries: TranscriptResponse[] = [
			{ text: 'Hello everyone.', offset: 15240, duration: 4240, lang: 'en' },
			{ text: 'Um, welcome to this talk.', offset: 16600, duration: 5080, lang: 'en' }
		];

		expect(transcriptEntriesToSegments(entries)).toEqual([
			{ index: 0, start: 15.24, end: 19.48, text: 'Hello everyone.', words: [] },
			{ index: 1, start: 16.6, end: 21.68, text: 'Um, welcome to this talk.', words: [] }
		]);
	});

	it('preserves second-based transcript offsets and drops empty text', () => {
		const entries: TranscriptResponse[] = [
			{ text: '  ', offset: 0, duration: 1.5, lang: 'en' },
			{ text: 'Clean caption cue', offset: 91.08, duration: 3.72, lang: 'en' }
		];

		expect(transcriptEntriesToSegments(entries)).toEqual([
			{ index: 0, start: 91.08, end: 94.8, text: 'Clean caption cue', words: [] }
		]);
	});
});

describe('saveUploadedFile', () => {
	it('saves file to persistent DATA_DIR/uploads/{jobId}/ location', async () => {
		const buffer = Buffer.from('fake audio content');
		const filePath = await saveUploadedFile(buffer, 'test-recording.webm', 'upload-test-1');

		// Must be under DATA_DIR/uploads/, NOT under TMP_DIR/downloads/
		expect(filePath).toContain('/uploads/upload-test-1/test-recording.webm');
		expect(filePath).not.toContain('/downloads/');

		// File must exist on disk
		const { accessSync } = await import('fs');
		expect(() => accessSync(filePath)).not.toThrow();

		// Content must match
		const { readFileSync } = await import('fs');
		expect(readFileSync(filePath)).toEqual(buffer);

		// cleanup
		await cleanupUploadDir('upload-test-1');
	});

	it('getUploadPath returns correct path without creating file', async () => {
		const p = getUploadPath('some-job', 'audio.mp3');
		expect(p).toContain('/uploads/some-job/audio.mp3');
		expect(p).not.toContain('/downloads/');
	});

	it('cleanupUploadDir removes the job upload directory', async () => {
		const buffer = Buffer.from('data');
		await saveUploadedFile(buffer, 'temp-file.txt', 'cleanup-test');
		const p = getUploadPath('cleanup-test', 'temp-file.txt');

		const { accessSync } = await import('fs');
		expect(() => accessSync(p)).not.toThrow();

		await cleanupUploadDir('cleanup-test');
		expect(() => accessSync(p)).toThrow();
	});

	it('cleanupUploadDir does not throw for non-existent dir', async () => {
		await expect(cleanupUploadDir('non-existent-job')).resolves.toBeUndefined();
	});

	it('each job gets an isolated upload directory', async () => {
		const p1 = await saveUploadedFile(Buffer.from('a'), 'a.wav', 'job-iso-1');
		const p2 = await saveUploadedFile(Buffer.from('b'), 'b.wav', 'job-iso-2');

		expect(p1).toContain('/uploads/job-iso-1/a.wav');
		expect(p2).toContain('/uploads/job-iso-2/b.wav');

		await cleanupUploadDir('job-iso-1');
		await cleanupUploadDir('job-iso-2');
	});
});

describe('downloadYouTube', () => {
	it('uses fetched transcript entries directly for caption jobs', async () => {
		mockFetchTranscript.mockResolvedValue([
			{ text: 'Hello everyone.', offset: 15240, duration: 4240, lang: 'en' },
			{ text: 'Um, welcome to this talk.', offset: 16600, duration: 5080, lang: 'en' }
		] satisfies TranscriptResponse[]);

		const result = await downloadYouTube('https://youtube.com/watch?v=qdh_x-uRs9g', 'job-1');

		expect(mockFetchTranscript).toHaveBeenCalledWith('https://youtube.com/watch?v=qdh_x-uRs9g', {
			lang: 'en'
		});
		expect(result).toMatchObject({
			type: 'captions',
			segments: [
				{ index: 0, start: 15.24, end: 19.48, text: 'Hello everyone.', words: [] },
				{ index: 1, start: 16.6, end: 21.68, text: 'Um, welcome to this talk.', words: [] }
			]
		});
	});
});