Files
researchChat/lib/tools.js
T

155 lines
4.9 KiB
JavaScript

const path = require('path');
const fs = require('fs');
const { chromium } = require('playwright');
const DASHBOARD_DIR = path.join(__dirname, '..', 'dashboards');
let browser = null;
const USER_AGENT =
'Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/137.0.0.0 Safari/537.36';
/** DDG-Redirect-URLs (duckduckgo.com/l/?uddg=...) auf die echte URL auflösen. */
function realUrl(href) {
try {
const u = new URL(href);
if (u.hostname === 'duckduckgo.com' && u.pathname.startsWith('/l/') && u.searchParams.get('uddg')) {
return u.searchParams.get('uddg');
}
} catch {}
return href;
}
async function getBrowser() {
if (!browser || !browser.isConnected()) {
browser = await chromium.launch({
channel: 'chromium',
headless: true,
args: ['--disable-blink-features=AutomationControlled', '--no-sandbox'],
});
}
return browser;
}
async function withPage(url, fn) {
const b = await getBrowser();
const page = await b.newPage({ userAgent: USER_AGENT, viewport: { width: 1280, height: 900 } });
try {
return await fn(page);
} finally {
await page.close();
}
}
async function goto(page, url, timeout = 30000) {
let resp;
try {
resp = await page.goto(url, { waitUntil: 'domcontentloaded', timeout });
} catch (e) {
return { ok: false, error: `Navigation fehlgeschlagen: ${e.message.slice(0, 200)}` };
}
await page.waitForTimeout(1200);
return {
ok: resp ? resp.status() < 400 : true,
status: resp ? resp.status() : 0,
finalUrl: page.url(),
};
}
/** DuckDuckGo HTML-Suche: Ergebnisse parsen. */
async function search(query, log) {
const url = `https://duckduckgo.com/html/?q=${encodeURIComponent(query)}`;
const results = await withPage(url, async (page) => {
const nav = await goto(page, url);
if (!nav.ok) return { nav, results: [], note: 'Suchseite nicht erreichbar' };
await page.waitForTimeout(2500);
const links = await page
.locator('a.result__a')
.evaluateAll((els) =>
els.slice(0, 10).map((a) => ({
title: (a.textContent || '').trim(),
url: a.href,
}))
);
for (const l of links) l.url = realUrl(l.url);
const snippets = await page
.locator('.result__snippet')
.evaluateAll((els) => els.slice(0, 10).map((e) => (e.textContent || '').trim()));
const blocked =
(await page.title().catch(() => '')) === 'CAPTCHA' ||
(await page.locator('text=CAPTCHA').count().catch(() => 0)) > 0;
return {
nav,
blocked,
results: links.map((l, i) => ({ ...l, snippet: snippets[i] || '' })),
};
});
log?.(`Suche "${query}" -> ${results.results.length} Ergebnisse`);
return results;
}
/** Seite besuchen und lesbaren Text extrahieren. */
async function browse(url, log) {
return withPage(url, async (page) => {
const nav = await goto(page, url);
if (!nav.ok) return { nav, error: 'Seite konnte nicht geladen werden' };
const data = await page.evaluate(() => {
const title = document.title || '';
const text = (document.body ? document.body.innerText : '').replace(/\s+\n/g, '\n');
return { title, text: text.slice(0, 15000) };
});
log?.(`Besuche ${url} -> ${data.title}`);
return { nav, title: data.title, text: data.text };
});
}
/** Link prüfen: lädt die Seite erfolgreich und hat sie echten Inhalt? */
async function verify(url, log) {
return withPage(url, async (page) => {
const nav = await goto(page, url, 25000);
if (!nav.ok) {
log?.(`Verifizierung FEHLGESCHLAGEN (${nav.status || 'Timeout'}): ${url}`);
return { url, ok: false, reason: `Seite lädt nicht (${nav.status || 'Timeout'})` };
}
const title = await page.title().catch(() => '');
const bodyLen = await page.evaluate(() => document.body?.innerText?.length || 0).catch(() => 0);
const badTitle = /404|not found|seite nicht gefunden|page not found/i.test(title);
const ok = !badTitle && bodyLen > 200;
log?.(`${ok ? 'VERIFIZIERT' : 'FEHLGESCHLAGEN'}: ${url}`);
return {
url,
ok,
reason: ok ? '' : badTitle ? 'Fehlerseite' : 'Kein nennenswerter Inhalt',
title,
};
});
}
/** Dashboard als HTML-Datei speichern. Gibt den URL-Pfad zurück. */
function saveDashboard(sessionId, filename, html, log) {
fs.mkdirSync(DASHBOARD_DIR, { recursive: true });
const safeName = (filename || 'dashboard')
.toLowerCase()
.replace(/[^a-z0-9äöü-]+/g, '-')
.replace(/^-+|-+$/g, '')
.slice(0, 80) || 'dashboard';
const file = `${sessionId}-${safeName}.html`;
fs.writeFileSync(path.join(DASHBOARD_DIR, file), html, 'utf8');
log?.(`Dashboard gespeichert: ${file}`);
return { file, url: `/dashboards/${file}` };
}
async function closeBrowser() {
if (browser) {
await browser.close().catch(() => {});
browser = null;
}
}
module.exports = { search, browse, verify, saveDashboard, closeBrowser };