#!/usr/bin/env node
/**
 * Scrapes bank/day promotions via Chrome Relay (CDP).
 * Opens each store's medios-de-pago page, waits for JS render, extracts promos.
 *
 * Usage: node scrape_bank_promos_relay.js
 */
'use strict';

const WebSocket = require('/usr/lib/node_modules/openclaw/node_modules/ws');
const fs = require('fs/promises');
const path = require('path');

const TOKEN = '0537d84a67f2e43b525964bb43d93f6dfae1ec1b50946455';
const RELAY_WS = 'ws://127.0.0.1:18792/cdp';
const WORKSPACE = '/home/ubuntu/.openclaw/workspace';

const STORES = [
  { name: 'dia',       url: 'https://diaonline.supermercadosdia.com.ar/medios-de-pago' },
  { name: 'carrefour', url: 'https://www.carrefour.com.ar/medios-de-pago' },
  { name: 'disco',     url: 'https://www.disco.com.ar/medios-de-pago' },
  { name: 'jumbo',     url: 'https://www.jumbo.com.ar/medios-de-pago' },
  { name: 'changomas', url: 'https://www.masonline.com.ar/medios-de-pago' },
];

// ── CDP connection (top-level, no flat sessions) ─────────────────────────────
function connectCDP() {
  return new Promise((resolve, reject) => {
    const ws = new WebSocket(RELAY_WS, {
      headers: { 'x-openclaw-relay-token': TOKEN },
    });
    let msgId = 1;
    const pending = new Map();
    const events = {};

    const cdp = {
      send(method, params = {}, sessionId = null) {
        return new Promise((res, rej) => {
          const id = msgId++;
          pending.set(id, { res, rej });
          const msg = { id, method, params };
          if (sessionId) msg.sessionId = sessionId;
          ws.send(JSON.stringify(msg));
          setTimeout(() => {
            if (pending.has(id)) { pending.delete(id); rej(new Error(`CDP timeout: ${method}`)); }
          }, 30000);
        });
      },
      on(event, fn) {
        if (!events[event]) events[event] = [];
        events[event].push(fn);
      },
      off(event) { delete events[event]; },
      close() { ws.close(); },
    };

    ws.on('message', raw => {
      const msg = JSON.parse(raw);
      if (msg.id !== undefined && pending.has(msg.id)) {
        const { res, rej } = pending.get(msg.id);
        pending.delete(msg.id);
        if (msg.error) rej(new Error(JSON.stringify(msg.error)));
        else res(msg.result);
      } else if (msg.method) {
        (events[msg.method] || []).forEach(fn => fn(msg.params, msg.sessionId));
      }
    });

    ws.on('open', () => resolve(cdp));
    ws.on('error', reject);
  });
}

// ── Navigate to a URL using an existing tab ──────────────────────────────────
async function scrapeViaTab(cdp, targetId, url) {
  // Attach (creates a flat session)
  const { sessionId } = await cdp.send('Target.attachToTarget', { targetId, flatten: true });

  const s = (method, params = {}) => cdp.send(method, params, sessionId);

  // Enable events on this session
  await s('Page.enable');
  await s('Runtime.enable');
  await s('Network.enable');

  // Navigate then wait for network idle (no activity for 3s) or 30s hard cap
  await new Promise(async (resolve) => {
    let resolved = false;
    let idleTimer = null;
    const done = () => { if (!resolved) { resolved = true; resolve(); } };
    const bump = () => {
      if (idleTimer) clearTimeout(idleTimer);
      idleTimer = setTimeout(done, 3000);
    };
    cdp.on('Page.loadEventFired', (p, sid) => { if (sid === sessionId) bump(); });
    cdp.on('Network.requestWillBeSent', (p, sid) => { if (sid === sessionId) bump(); });
    cdp.on('Network.loadingFinished', (p, sid) => { if (sid === sessionId) bump(); });
    cdp.on('Network.loadingFailed', (p, sid) => { if (sid === sessionId) bump(); });
    setTimeout(done, 30000);
    await s('Page.navigate', { url });
    bump();
  });

  // Additional hydration buffer for VTEX React components
  await new Promise(r => setTimeout(r, 5000));

  // Extract text content
  const evalResult = await s('Runtime.evaluate', {
    expression: `(function() {
      const txt = document.body ? document.body.innerText : '';
      return {
        url: location.href,
        title: document.title,
        text: txt.slice(0, 12000),
        textLen: txt.length,
      };
    })()`,
    returnByValue: true,
  });

  await s('Target.detachFromTarget', { sessionId }).catch(() => {});
  return evalResult?.result?.value || null;
}

// ── Parse bank promos from page text ────────────────────────────────────────
function parsePromos(raw) {
  if (!raw || !raw.text) return { error: 'no text extracted', url: raw?.url };

  const text = raw.text;
  const DAYS_CANONICAL = {
    'lunes a domingo': 'Todos los días', 'todos los días': 'Todos los días',
    'lunes': 'Lunes', 'martes': 'Martes',
    'miércoles': 'Miércoles', 'miercoles': 'Miércoles',
    'jueves': 'Jueves', 'viernes': 'Viernes',
    'sábado': 'Sábado', 'sabado': 'Sábado', 'domingo': 'Domingo',
  };
  const DAY_RE = /\b(lunes a domingo|todos los d[ií]as|lunes|martes|mi[eé]rcoles|jueves|viernes|s[aá]bado|domingo)\b/i;
  const DISC_RE = /(\d+)\s*%\s*(?:de\s+)?descuento/i;
  const CUOTAS_RE = /(\d+)\s+cuotas?\s+sin\s+inter[eé]s/i;
  const BANK_RE = /\b(visa(?:\s+electr[oó]n)?|mastercard|amex|american express|naranja(?:\s*x)?|galicia|santander|macro|supervielle|icbc|mercado\s*pago|modo|naci[oó]n|provincia|ciudad|cabal|hsbc|bbva|patagonia|comafi|brubank|u[a]la|oca|lemon|prex|personal\s*pay|clar[ií]n\s*365|cuenta\s*dni|getnet|coto)\b/gi;

  const lines = text.split('\n').map(l => l.trim()).filter(l => l.length > 3);
  const byDay = {};
  let currentDay = 'Todos los días';

  for (let i = 0; i < lines.length; i++) {
    const line = lines[i];
    const lineLc = line.toLowerCase();

    // Update current day context
    const dayM = DAY_RE.exec(line);
    if (dayM) {
      const canonical = DAYS_CANONICAL[dayM[1].toLowerCase()] || dayM[1];
      currentDay = canonical;
    }

    // Check for discount/cuotas in this line
    const discM = DISC_RE.exec(line);
    const cuotasM = CUOTAS_RE.exec(line);
    if (!discM && !cuotasM) continue;

    const discount = discM ? discM[0] : cuotasM[0];

    // Context window: ±3 lines
    const ctx = lines.slice(Math.max(0, i - 3), Math.min(lines.length, i + 4)).join(' ');

    // Day from context
    const ctxDayM = DAY_RE.exec(ctx);
    const day = ctxDayM
      ? (DAYS_CANONICAL[ctxDayM[1].toLowerCase()] || ctxDayM[1])
      : currentDay;

    // Banks from context
    BANK_RE.lastIndex = 0;
    const banks = [...new Set((ctx.match(BANK_RE) || []).map(b => b.trim()))];

    // Tope
    const topeM = ctx.match(/tope[^$\n]*\$\s*[\d.,]+/i);
    const vigenciaM = ctx.match(/hasta\s+el\s+\d+\/\d+/i);

    const promo = {
      descuento: discount,
      bancos: banks,
      tope: topeM ? topeM[0].trim() : '',
      vigencia: vigenciaM ? vigenciaM[0].trim() : '',
      detalle: ctx.replace(/\s+/g, ' ').trim().slice(0, 250),
    };

    if (!byDay[day]) byDay[day] = [];

    // Deduplicate
    const key = `${discount}|${banks.sort().join(',')}`;
    if (!byDay[day].some(p => `${p.descuento}|${p.bancos.sort().join(',')}` === key)) {
      byDay[day].push(promo);
    }
  }

  const total = Object.values(byDay).reduce((s, a) => s + a.length, 0);
  return {
    source: 'chrome-relay-cdp',
    url: raw.url,
    title: raw.title,
    textLen: raw.textLen,
    totalPromos: total,
    byDay,
  };
}

// ── Main ─────────────────────────────────────────────────────────────────────
async function main() {
  console.log('Connecting to Chrome Relay...');
  const cdp = await connectCDP();

  const { targetInfos } = await cdp.send('Target.getTargets');
  const pages = targetInfos.filter(t => t.type === 'page');
  console.log(`Tabs open: ${pages.length}`);

  // Use first available page tab as our scraping tab
  let scrapeTargetId = pages[0]?.targetId;
  if (!scrapeTargetId) {
    const { targetId } = await cdp.send('Target.createTarget', { url: 'about:blank' });
    scrapeTargetId = targetId;
    console.log('Created new tab for scraping');
  } else {
    console.log(`Using tab: ${pages[0].url.slice(0, 60)}`);
  }

  const results = {};

  for (const store of STORES) {
    process.stdout.write(`\n[${store.name}] Scraping ${store.url} ... `);
    try {
      const raw = await scrapeViaTab(cdp, scrapeTargetId, store.url);
      if (!raw) throw new Error('null result from CDP');
      console.log(`${raw.textLen} chars extracted`);
      results[store.name] = parsePromos(raw);
      const r = results[store.name];
      if (r.error) {
        console.log(`  -> ERROR: ${r.error}`);
      } else {
        console.log(`  -> ${r.totalPromos} promos, días: [${Object.keys(r.byDay).join(', ')}]`);
        for (const [day, promos] of Object.entries(r.byDay)) {
          for (const p of promos.slice(0, 4)) {
            console.log(`     ${day}: ${p.descuento} — ${p.bancos.join(', ') || '?'} ${p.tope}`);
          }
        }
      }
    } catch (e) {
      console.log(`FAILED: ${e.message}`);
      results[store.name] = { source: 'chrome-relay-cdp', error: e.message };
    }
    await new Promise(r => setTimeout(r, 800));
  }

  cdp.close();

  // Save
  const outFile = path.join(WORKSPACE, 'data', 'discounts', 'bank-promos-relay-latest.json');
  const out = { generatedAt: new Date().toISOString(), source: 'chrome-relay-cdp', stores: results };
  await fs.writeFile(outFile, JSON.stringify(out, null, 2));

  const ts = new Date().toISOString().replace(/[:.]/g, '-').slice(0, 19);
  await fs.writeFile(
    path.join(WORKSPACE, 'data', 'discounts', `bank-promos-relay-${ts}.json`),
    JSON.stringify(out, null, 2)
  );

  console.log(`\nSaved: ${outFile}`);
  return results;
}

main().catch(e => { console.error(e); process.exit(1); });