#!/usr/bin/env node
/* Second look, in a real browser (Chrome, phone size, JavaScript on), at EVERY sampled website that opened in the first
   check. It answers tap-to-call, hours, address and "a way to reach them" again, both ways (a wrong "yes" counts as much as a wrong "no"):
     tel    a tel: link with a real phone number (10+ digits)
     hours  opening hours in the text a visitor sees, or in the page data (JSON-LD openingHours)
   and flags pages that open but show an error ("Not found", "404", almost no text).
   The rules for hours come from the 30-site review (20 Oct): "M - F: 8AM - 5PM" and "Open today / 08:30 am – 05:00 pm"
   were missed before. Reads out/frame.json + the scan, writes out/browser-all.json.
   Run: node browser-all.mjs <scan prospects.json> <out dir> [--pw <dir with node_modules/playwright-core>, default: the current folder] */
import fs from 'node:fs';
import path from 'node:path';
import { pathToFileURL } from 'node:url';
const [scan, out] = process.argv.slice(2);
const pi = process.argv.indexOf('--pw');
const pw = pi > 0 ? process.argv[pi + 1] : process.cwd();
const { chromium } = await import(pathToFileURL(path.join(pw, 'node_modules/playwright-core/index.mjs')).href);
const S = JSON.parse(fs.readFileSync(scan, 'utf8'));
const F = JSON.parse(fs.readFileSync(path.join(out, 'frame.json'), 'utf8'));
const picked = new Set(Object.values(F.cells).flatMap((c) => c.picked));
const sites = S.rows.filter((r) => picked.has(r.place_id) && r.site_kind === 'own' && r.check?.fetch_status === 'ok').map((r) => ({ id: r.place_id, url: r.check.raw.final_url }));

const DAY = '(?:mon|tue|wed|thu|fri|sat|sun)[a-z]*\\.?|m\\s*[-–]\\s*f|m\\s*[-–]\\s*th';
const TIME = '\\d{1,2}(?::\\d{2})?\\s*(?:am|pm|a\\.m\\.|p\\.m\\.|a\\b|p\\b)';   // "8:00a-12:00p" seen in the second review
const RANGE = `${TIME}\\s*(?:-|–|—|to)\\s*${TIME}`;
const HOURS = [
  new RegExp(`\\b(?:${DAY})(?:\\s*(?:-|–|to|&|thru|through)\\s*(?:${DAY}))?[:,\\s]+(?:${RANGE}|open 24 hours|closed)`, 'i'),
  new RegExp(`\\b(?:hours|open today|open daily|we.re open)\\b[\\s\\S]{0,80}?${RANGE}`, 'i'),
  /\bopen 24 hours\b|\b24\/7\b/i,
];
/* a US street address: number + street name + street word (any case), e.g. "5022 South 33rd street", "515 W. WASHINGTON AVE" */
const ADDRESS = /\b\d{1,6}\s+(?:[NSEW]\.?\s+|north |south |east |west )?(?:[A-Za-z0-9.'-]+\s+){0,4}(?:st|street|ave|avenue|rd|road|blvd|boulevard|dr|drive|ln|lane|hwy|highway|way|pkwy|parkway|ct|court|pl|place|sq|square|cir|circle|trl|trail|pike|loop)\b\.?/i;
const b = await chromium.launch({ executablePath: '/Applications/Google Chrome.app/Contents/MacOS/Google Chrome' });
const ctx = await b.newContext({ viewport: { width: 390, height: 844 }, isMobile: true, hasTouch: true,
  userAgent: 'Mozilla/5.0 (iPhone; CPU iPhone OS 18_0 like Mac OS X) AppleWebKit/605.1.15 (KHTML, like Gecko) Version/18.0 Mobile/15E148 Safari/604.1' });
const res = {};
let i = 0;
async function worker() {
  while (i < sites.length) {
    const s = sites[i++];
    const p = await ctx.newPage();
    try {
      const r = await p.goto(s.url, { waitUntil: 'domcontentloaded', timeout: 30000 });
      await p.waitForLoadState('networkidle', { timeout: 12000 }).catch(() => {});
      await p.waitForTimeout(1500);
      const look = await p.evaluate(() => ({
        tels: [...document.querySelectorAll('a[href^="tel:" i]')].map((a) => a.getAttribute('href').replace(/\D/g, '')),
        mail: [...document.querySelectorAll('a[href^="mailto:" i]')].some((a) => /@/.test(a.getAttribute('href'))),
        contactLink: [...document.querySelectorAll('a[href]')].some((a) => /contact/i.test(a.getAttribute('href') + ' ' + a.textContent)),
        /* a form a customer can write in: a message box, or 3+ fields (a lone e-mail box is a newsletter sign-up) */
        form: [...document.querySelectorAll('form')].some((f) => f.querySelector('textarea') || f.querySelectorAll('input:not([type=hidden]):not([type=submit]):not([type=button])').length >= 3),
        text: document.body?.innerText || '', title: document.title,
        ld: [...document.querySelectorAll('script[type="application/ld+json"]')].map((x) => x.textContent).join(' '),
        es: /^es\b/i.test(document.documentElement.lang) || !!document.querySelector('link[hreflang^="es" i]') || [...document.querySelectorAll('a[href]')].some((a) => /espa(ñ|n)ol|\bspanish\b/i.test(a.textContent) || /\/es(\/|$)|\/\/es\./i.test(a.getAttribute('href'))),
      }));
      const t = look.text.replace(/\s+/g, ' ').trim();
      /* not the business's own website: a parked or for-sale domain (seen in the second review) */
      const parked = /parked domain|domain (is )?for sale|buy this domain|this domain (may be|is) for sale/i.test(t.slice(0, 600));
      const notFound = parked || t.length < 40 || /^(not found|404|page not found)\b/i.test(t) || /\b404\b.*not found|page (can.t|cannot) be found/i.test(t.slice(0, 300)) || /^404|not found/i.test(look.title.trim());
      res[s.id] = { ok: true, status: r ? r.status() : 0, notFound,
        tel: look.tels.some((d) => d.length >= 10),
        hours: /openingHours/i.test(look.ld) || HOURS.some((re) => re.test(look.text)),
        address: /"streetAddress"/i.test(look.ld) || ADDRESS.test(look.text),
        reach: look.mail || look.contactLink || look.form, spanish: look.es, parked };
    } catch (e) { res[s.id] = { ok: false, error: String(e.message || e).split('\n')[0].slice(0, 120) }; }
    await p.close();
    process.stderr.write(`\r${i}/${sites.length}`);
  }
}
await Promise.all([worker(), worker(), worker(), worker()]);
await b.close();
fs.writeFileSync(path.join(out, 'browser-all.json'), JSON.stringify(res, null, 1));
const v = Object.values(res);
console.error(`\nok ${v.filter((x) => x.ok).length} · notFound ${v.filter((x) => x.notFound).length} · tel ${v.filter((x) => x.tel).length} · hours ${v.filter((x) => x.hours).length} · address ${v.filter((x) => x.address).length} · reach ${v.filter((x) => x.reach).length} · errors ${v.filter((x) => !x.ok).length}`);