ilia
55018a6e3e
Expose ExtractorSession for Playwright/camoufox-js so Node automations can import context dumps instead of shelling out to the Python package. Same fixture as the Python tests; wired into Makefile and Gitea CI.
118 lines
4.3 KiB
TypeScript
118 lines
4.3 KiB
TypeScript
import { afterAll, beforeAll, describe, expect, it } from 'vitest';
|
|
import { chromium, type Browser, type Page } from 'playwright-core';
|
|
import { fileURLToPath } from 'node:url';
|
|
import { dirname, join } from 'node:path';
|
|
import { pathToFileURL } from 'node:url';
|
|
import { ExtractorSession } from '../src/session.js';
|
|
|
|
// Same golden fixture the Python `automation/tests/` suite uses — one file,
|
|
// two automation layers, so a markdown-formatting bug can't silently diverge
|
|
// between them.
|
|
const here = dirname(fileURLToPath(import.meta.url));
|
|
const FIXTURE_URL = pathToFileURL(
|
|
join(here, '..', '..', 'automation', 'tests', 'fixtures', 'sample.html'),
|
|
).toString();
|
|
|
|
let browser: Browser;
|
|
|
|
beforeAll(async () => {
|
|
browser = await chromium.launch({ headless: true });
|
|
});
|
|
|
|
afterAll(async () => {
|
|
await browser.close();
|
|
});
|
|
|
|
async function withPage<T>(fn: (page: Page) => Promise<T>): Promise<T> {
|
|
const p = await browser.newPage();
|
|
try {
|
|
return await fn(p);
|
|
} finally {
|
|
await p.close();
|
|
}
|
|
}
|
|
|
|
describe('ExtractorSession', () => {
|
|
it('captures console messages, page errors, and network activity', async () => {
|
|
await withPage(async (page) => {
|
|
const session = new ExtractorSession(page);
|
|
await page.goto(FIXTURE_URL, { waitUntil: 'domcontentloaded' });
|
|
await page.waitForTimeout(150);
|
|
|
|
const store = session.getStore();
|
|
const levels = new Set(store.console.map((e) => e.level));
|
|
expect(levels.has('warn')).toBe(true);
|
|
expect(
|
|
levels.has('error') ||
|
|
store.console.some((e) => (e.msg || '').toLowerCase().includes('error')),
|
|
).toBe(true);
|
|
|
|
expect(store.errors.some((e) => (e.msg || '').includes('fixture boom'))).toBe(true);
|
|
expect(store.network.length).toBeGreaterThan(0);
|
|
});
|
|
});
|
|
|
|
it('extracts markdown from the full body and from a selector', async () => {
|
|
await withPage(async (page) => {
|
|
const session = new ExtractorSession(page);
|
|
await page.goto(FIXTURE_URL, { waitUntil: 'domcontentloaded' });
|
|
|
|
const full = await session.extractMarkdown();
|
|
expect(full.title).toBe('Context Extractor Fixture');
|
|
expect(full.markdown).toContain('Fixture Page');
|
|
expect(full.markdown).toContain('**world**');
|
|
expect(full.markdown).toContain('[docs](/docs)');
|
|
expect(full.markdown).toContain('- Alpha');
|
|
|
|
const scoped = await session.extractMarkdown('#main-content');
|
|
expect(scoped.selector).toBe('#main-content');
|
|
expect(scoped.markdown).toContain('Section');
|
|
expect(scoped.markdown).not.toContain('Skip this noise');
|
|
});
|
|
});
|
|
|
|
it('excludes hidden nodes (inline style, stylesheet class, [hidden], aria-hidden) from markdown', async () => {
|
|
await withPage(async (page) => {
|
|
const session = new ExtractorSession(page);
|
|
await page.goto(FIXTURE_URL, { waitUntil: 'domcontentloaded' });
|
|
|
|
const { markdown } = await session.extractMarkdown('#main-content');
|
|
expect(markdown).not.toContain('should-not-appear');
|
|
expect(markdown).not.toContain('secretConfig');
|
|
expect(markdown).not.toContain('lixTracking');
|
|
expect(markdown).toContain('Alpha');
|
|
expect(markdown).toContain('Beta');
|
|
});
|
|
});
|
|
|
|
it('builds an AI prompt with the expected sections', async () => {
|
|
await withPage(async (page) => {
|
|
const session = new ExtractorSession(page);
|
|
await page.goto(FIXTURE_URL, { waitUntil: 'domcontentloaded' });
|
|
await page.waitForTimeout(150);
|
|
|
|
const prompt = await session.buildAiPrompt('#main-content');
|
|
expect(prompt.startsWith('# Page Context')).toBe(true);
|
|
expect(prompt).toContain('## Page Content');
|
|
expect(prompt).toContain('Hello **world**');
|
|
expect(/JavaScript Errors|Console Errors/.test(prompt)).toBe(true);
|
|
expect(prompt).toContain('_Extracted by Context Extractor_');
|
|
});
|
|
});
|
|
|
|
it('clears the capture store, selectively and fully', async () => {
|
|
await withPage(async (page) => {
|
|
const session = new ExtractorSession(page);
|
|
await page.goto(FIXTURE_URL, { waitUntil: 'domcontentloaded' });
|
|
await page.waitForTimeout(100);
|
|
|
|
expect(session.getStore().console.length).toBeGreaterThan(0);
|
|
session.clear('console');
|
|
expect(session.getStore().console).toEqual([]);
|
|
|
|
session.clear();
|
|
expect(session.getStore()).toEqual({ console: [], errors: [], network: [] });
|
|
});
|
|
});
|
|
});
|