continente-mcp 3.1.0 → 4.0.0

package/check-change.js

@@ -1,157 +0,0 @@
-#!/usr/bin/env node
-/**
- * Continente Change Detector
- *
- * Checks whether the website structure has changed in a way that might break
- * the MCP server's scrapers. Run daily via cron.
- *
- * Usage:
- *   node check-change.js
- *
- * On first run it saves a baseline fingerprint. Subsequent runs compare against
- * it and log a warning if the structure changed significantly.
- *
- * To receive a webhook notification on change, set:
- *   CONTINENTE_ALERT_WEBHOOK=https://hooks.slack.com/...
- */
-
-import { createHash } from 'crypto';
-import { existsSync, readFileSync, writeFileSync, mkdirSync } from 'fs';
-import { chromium } from 'playwright';
-import * as cheerio from 'cheerio';
-import os from 'os';
-
-const CONTINENTE_SEARCH  = 'https://www.continente.pt/pesquisa/?q=leite';
-const STATE_DIR          = `${os.homedir()}/.continente`;
-const FINGERPRINT_FILE   = `${STATE_DIR}/fingerprint.json`;
-const LAST_ALERT_FILE    = `${STATE_DIR}/last_change_alert.txt`;
-const COOKIE_FILE        = `${STATE_DIR}/cookies.json`;
-const WEBHOOK_URL        = process.env.CONTINENTE_ALERT_WEBHOOK || null;
-
-function getPlatformUserAgent() {
-  if (process.platform === 'win32') {
-    return 'Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/121.0.0.0 Safari/537.36';
-  } else if (process.platform === 'linux') {
-    return 'Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/121.0.0.0 Safari/537.36';
-  }
-  return 'Mozilla/5.0 (Macintosh; Intel Mac OS X 10_15_7) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/121.0.0.0 Safari/537.36';
-}
-
-function extractFingerprint(html) {
-  const $ = cheerio.load(html);
-  $('script, style, noscript, iframe, svg').remove();
-
-  const counts = {
-    produtoLinks:  $('a[href*="/produto/"]').length,
-    priceElements: $('[class*="price"]').filter((i, el) => $(el).text().includes('€')).length,
-    forms:         $('form').length,
-    buttons:       $('button').length,
-    productCards:  $('[class*="product"], [class*="card"]').filter(
-      (i, el) => $(el).find('a[href*="/produto/"]').length > 0
-    ).length,
-  };
-
-  return {
-    ...counts,
-    hash:      createHash('sha256').update(JSON.stringify(counts)).digest('hex').substring(0, 12),
-    timestamp: new Date().toISOString(),
-  };
-}
-
-async function fetchHtml() {
-  const browser = await chromium.launch({ headless: true });
-  const context = await browser.newContext({ userAgent: getPlatformUserAgent() });
-
-  if (existsSync(COOKIE_FILE)) {
-    try {
-      const cookies = JSON.parse(readFileSync(COOKIE_FILE, 'utf8'));
-      if (cookies.length > 0) await context.addCookies(cookies);
-    } catch {}
-  }
-
-  const page = await context.newPage();
-  await page.goto(CONTINENTE_SEARCH, { waitUntil: 'domcontentloaded', timeout: 20000 });
-  await page.waitForTimeout(3000);
-
-  const html = await page.content();
-  await browser.close();
-  return html;
-}
-
-async function sendAlert(message) {
-  if (WEBHOOK_URL) {
-    // POST to webhook (works with Slack, Discord, n8n, etc.)
-    try {
-      const res = await fetch(WEBHOOK_URL, {
-        method: 'POST',
-        headers: { 'Content-Type': 'application/json' },
-        body: JSON.stringify({ text: message }),
-      });
-      if (!res.ok) throw new Error(`HTTP ${res.status}`);
-      console.error('[ChangeDetector] Webhook notified.');
-    } catch (e) {
-      console.error(`[ChangeDetector] Webhook failed: ${e.message}`);
-    }
-  } else {
-    // No webhook configured — log prominently to stdout so cron picks it up
-    console.log('\n' + '='.repeat(60));
-    console.log('CONTINENTE STRUCTURE CHANGE DETECTED');
-    console.log('='.repeat(60));
-    console.log(message);
-    console.log('='.repeat(60) + '\n');
-  }
-}
-
-async function main() {
-  mkdirSync(STATE_DIR, { recursive: true });
-
-  console.error(`[ChangeDetector] Checking ${CONTINENTE_SEARCH}`);
-  const html    = await fetchHtml();
-  const current = extractFingerprint(html);
-
-  if (!existsSync(FINGERPRINT_FILE)) {
-    writeFileSync(FINGERPRINT_FILE, JSON.stringify({ baseline: current, savedAt: current.timestamp }, null, 2));
-    console.error(`[ChangeDetector] Baseline saved (${current.hash}). Run again tomorrow to start detecting changes.`);
-    return;
-  }
-
-  const { baseline } = JSON.parse(readFileSync(FINGERPRINT_FILE, 'utf8'));
-
-  if (baseline.hash === current.hash) {
-    console.error(`[ChangeDetector] Unchanged (${current.hash}). All good.`);
-    return;
-  }
-
-  const diffs = {};
-  for (const key of Object.keys(baseline)) {
-    if (key === 'hash' || key === 'timestamp') continue;
-    if (JSON.stringify(baseline[key]) !== JSON.stringify(current[key])) {
-      diffs[key] = { before: baseline[key], after: current[key] };
-    }
-  }
-
-  const diffSummary = Object.entries(diffs).map(([k, v]) => `${k}: ${v.before} → ${v.after}`).join(', ');
-  console.error(`[ChangeDetector] CHANGE DETECTED: ${diffSummary}`);
-
-  // Rate-limit: only alert once per 24 h
-  if (existsSync(LAST_ALERT_FILE)) {
-    const lastAlert = readFileSync(LAST_ALERT_FILE, 'utf8').trim();
-    const hoursSince = (Date.now() - new Date(lastAlert).getTime()) / 3_600_000;
-    if (hoursSince < 24) {
-      console.error(`[ChangeDetector] Alert suppressed (last sent ${hoursSince.toFixed(1)}h ago).`);
-      return;
-    }
-  }
-
-  const alertMsg =
-    `⚠️ Continente website structure changed — the MCP scraper may need updating.\n\n` +
-    `Baseline: ${baseline.hash}\n` +
-    `Current:  ${current.hash}\n` +
-    `Changes:  ${diffSummary}\n\n` +
-    `Check https://github.com/davidmarqu3s/continente-mcp for updates.`;
-
-  await sendAlert(alertMsg);
-  writeFileSync(LAST_ALERT_FILE, new Date().toISOString());
-}
-
-main().catch(e => { console.error('[ChangeDetector] Fatal:', e.message); process.exit(1); });

package/continente-backup.js

@@ -1,166 +0,0 @@
-#!/usr/bin/env node
-/**
- * Continente Order Backup — incremental
- *
- * Checks for orders not yet in the backup, fetches their details,
- * and appends them. Run daily via cron.
- *
- * Usage:
- *   node continente-backup.js
- *
- * Output:
- *   ~/.continente/orders-backup.json
- */
-import { chromium } from 'playwright';
-import { existsSync, readFileSync, writeFileSync, mkdirSync } from 'fs';
-import { normalizeCookies } from './src/utils.js';
-import os from 'os';
-
-const CONTINENTE_BASE = 'https://www.continente.pt';
-const STATE_DIR       = `${os.homedir()}/.continente`;
-const BACKUP_FILE     = `${STATE_DIR}/orders-backup.json`;
-const COOKIE_FILE     = `${STATE_DIR}/cookies.json`;
-
-const log = (...args) => console.error('[backup]', ...args);
-
-// ─── Cookie helpers ───────────────────────────────────────────────────────────
-
-function loadCookies() {
-  if (!existsSync(COOKIE_FILE)) {
-    log(`No cookies found at ${COOKIE_FILE}`);
-    log('Run: python3 continente-cookie-reader.py');
-    return null;
-  }
-  try {
-    return JSON.parse(readFileSync(COOKIE_FILE, 'utf8'));
-  } catch (e) {
-    log('Failed to parse cookies:', e.message);
-    return null;
-  }
-}
-
-// ─── Scraping helpers ─────────────────────────────────────────────────────────
-
-async function getOrderLinks(page) {
-  await page.goto(`${CONTINENTE_BASE}/conta/encomendas/`, { waitUntil: 'networkidle', timeout: 30000 });
-  await page.waitForTimeout(3000);
-
-  if (page.url().includes('/login')) throw new Error('Not authenticated — cookies may have expired');
-
-  return page.evaluate(() => {
-    const seen = new Set();
-    return Array.from(document.querySelectorAll('a[href*="detalhe-encomenda"]'))
-      .map(a => a.href)
-      .filter(h => { if (seen.has(h)) return false; seen.add(h); return true; });
-  });
-}
-
-async function getOrderDetail(page, url) {
-  await page.goto(url, { waitUntil: 'networkidle', timeout: 20000 });
-  await page.waitForTimeout(2000);
-
-  return page.evaluate(() => {
-    const bodyText = document.body.innerText;
-    const orderNumMatch = bodyText.match(/Encomenda (\d{9}_\d{3})/);
-    const totalMatch    = bodyText.match(/Total Encomendado\s+([\d,.]+€)/);
-    const slotMatch     = bodyText.match(/(\d{1,2}:\d{2} - \d{1,2}:\d{2})/);
-    const dateMatch     = bodyText.match(/(?:Jan|Feb|Mar|Apr|May|Jun|Jul|Aug|Sep|Oct|Nov|Dec)\s+\d{1,2},\s+\d{4}/i);
-
-    const products = [];
-    document.querySelectorAll('[class*="product-line"]').forEach(el => {
-      const qtyEl = el.querySelector('[class*="qty"], [class*="quantity"], [class*="amount"]');
-      if (!qtyEl) return;
-      const nameEl  = el.querySelector('[class*="product-name"], [class*="name"], a[href*="produto"]');
-      const rawName = (nameEl || el).textContent.trim();
-      const name    = rawName.split('\n')[0].trim();
-      const qtyMatch = qtyEl.textContent.trim().match(/^(\d+)/);
-      const qty     = qtyMatch ? parseInt(qtyMatch[1], 10) : 1;
-      const priceEl = el.querySelector('[class*="price"]');
-      const price   = priceEl ? priceEl.textContent.trim().split('\n')[0].trim() : null;
-      if (name && name.length > 3) products.push({ name, qty, price });
-    });
-
-    return {
-      orderNumber:  orderNumMatch ? orderNumMatch[1] : null,
-      date:         dateMatch     ? dateMatch[0]     : null,
-      total:        totalMatch    ? totalMatch[1]    : null,
-      deliverySlot: slotMatch     ? slotMatch[1]     : null,
-      productCount: products.length,
-      products,
-    };
-  });
-}
-
-// ─── Main ─────────────────────────────────────────────────────────────────────
-
-async function main() {
-  const cookies = loadCookies();
-  if (!cookies) process.exit(1);
-
-  // Load existing backup
-  let backup = { backedUpAt: null, orderCount: 0, orders: [] };
-  if (existsSync(BACKUP_FILE)) {
-    try { backup = JSON.parse(readFileSync(BACKUP_FILE, 'utf8')); } catch {}
-  }
-  const knownUrls = new Set(backup.orders.map(o => o.url).filter(Boolean));
-
-  const browser = await chromium.launch({ headless: true });
-  const context = await browser.newContext({
-    userAgent: getPlatformUserAgent(),
-    locale: 'pt-PT',
-  });
-  await context.addCookies(normalizeCookies(cookies));
-  const page = await context.newPage();
-
-  let orderLinks;
-  try {
-    orderLinks = await getOrderLinks(page);
-  } catch (e) {
-    log('Failed to load orders page:', e.message);
-    await browser.close();
-    process.exit(1);
-  }
-
-  const newLinks = orderLinks.filter(l => !knownUrls.has(l));
-  log(`${orderLinks.length} orders on site, ${knownUrls.size} already backed up, ${newLinks.length} new`);
-
-  if (newLinks.length === 0) {
-    log('Nothing new to backup.');
-    await browser.close();
-    return;
-  }
-
-  const newOrders = [];
-  for (let i = 0; i < newLinks.length; i++) {
-    try {
-      const detail = await getOrderDetail(page, newLinks[i]);
-      detail.url = newLinks[i];
-      newOrders.push(detail);
-      log(`${i + 1}/${newLinks.length} — ${detail.orderNumber || '?'} (${detail.productCount} products)`);
-    } catch (e) {
-      log(`${i + 1}/${newLinks.length} FAILED: ${e.message}`);
-      newOrders.push({ url: newLinks[i], error: e.message });
-    }
-  }
-
-  backup.orders    = [...newOrders, ...backup.orders];
-  backup.orderCount = backup.orders.filter(o => !o.error).length;
-  backup.backedUpAt = new Date().toISOString();
-
-  mkdirSync(STATE_DIR, { recursive: true });
-  writeFileSync(BACKUP_FILE, JSON.stringify(backup, null, 2));
-  log(`Done — added ${newOrders.length} orders. Total: ${backup.orderCount} → ${BACKUP_FILE}`);
-
-  await browser.close();
-}
-
-function getPlatformUserAgent() {
-  if (process.platform === 'win32') {
-    return 'Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/121.0.0.0 Safari/537.36';
-  } else if (process.platform === 'linux') {
-    return 'Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/121.0.0.0 Safari/537.36';
-  }
-  return 'Mozilla/5.0 (Macintosh; Intel Mac OS X 10_15_7) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/121.0.0.0 Safari/537.36';
-}
-
-main().catch(e => { log('Fatal:', e.message); process.exit(1); });

package/continente-cookie-reader.py

@@ -1,164 +0,0 @@
-#!/usr/bin/env python3
-"""
-continente-cookie-reader — export Continente.pt cookies from your browser.
-
-Reads browser cookies for continente.pt and writes them to ~/.continente/cookies.json
-so the MCP server can use them. Auto-detects your browser or let you specify one.
-
-Usage:
-  python3 continente-cookie-reader.py                    # auto-detect
-  python3 continente-cookie-reader.py --browser chrome
-  python3 continente-cookie-reader.py --browser firefox
-  python3 continente-cookie-reader.py --output /path/to/cookies.json
-  python3 continente-cookie-reader.py --list-browsers
-
-Requirements:
-  pip install browser-cookie3
-
-Environment variables:
-  CONTINENTE_COOKIES_PATH      Override the default output path (~/.continente/cookies.json)
-  CONTINENTE_VAULT_COOKIE_PATH If set, also write cookies to this path (useful for syncing
-                               to a second machine via a shared folder, e.g. Obsidian vault)
-"""
-import argparse
-import json
-import os
-import stat
-import sys
-from pathlib import Path
-
-DOMAIN = "continente.pt"
-DEFAULT_OUTPUT = Path(
-    os.environ.get("CONTINENTE_COOKIES_PATH", str(Path.home() / ".continente" / "cookies.json"))
-)
-
-# If set, cookies are also written here (e.g. for syncing to a remote machine)
-VAULT_PATH = Path(os.environ["CONTINENTE_VAULT_COOKIE_PATH"]) if "CONTINENTE_VAULT_COOKIE_PATH" in os.environ else None
-
-# Tried in this order when auto-detecting
-BROWSERS = ["arc", "chrome", "edge", "brave", "firefox", "chromium", "vivaldi", "opera", "safari"]
-
-
-def get_browser_cookies(browser_name: str) -> list[dict]:
-    """Read continente.pt cookies from the given browser."""
-    import browser_cookie3
-
-    getter = getattr(browser_cookie3, browser_name, None)
-    if getter is None:
-        raise ValueError(f"Browser '{browser_name}' not supported by browser-cookie3.")
-
-    jar = getter(domain_name=DOMAIN)
-    cookies = []
-    for c in jar:
-        cookies.append({
-            "domain": c.domain,
-            "name": c.name,
-            "value": c.value,
-            "path": getattr(c, "path", "/"),
-            "secure": bool(getattr(c, "secure", False)),
-            "httpOnly": c.has_nonstandard_attr("HttpOnly") if hasattr(c, "has_nonstandard_attr") else False,
-            "sameSite": "Lax",
-            "expires": getattr(c, "expires", None),
-        })
-    return cookies
-
-
-def find_cookies(browser: str | None) -> tuple[list[dict], str]:
-    """Return (cookies, browser_name). Tries specified browser or auto-detects."""
-    if browser:
-        try:
-            cookies = get_browser_cookies(browser)
-        except Exception as e:
-            print(f"Error reading {browser}: {e}", file=sys.stderr)
-            sys.exit(1)
-        if not cookies:
-            print(f"No Continente cookies found in {browser}. Are you logged in?", file=sys.stderr)
-            sys.exit(1)
-        return cookies, browser
-
-    print("Auto-detecting browser with Continente cookies...", file=sys.stderr)
-    for name in BROWSERS:
-        try:
-            cookies = get_browser_cookies(name)
-            if cookies:
-                return cookies, name
-        except Exception as e:
-            print(f"  {name}: {e}", file=sys.stderr)
-            continue
-
-    print(
-        "\nNo Continente cookies found in any supported browser.\n"
-        "Make sure you are logged into continente.pt, then try again.\n"
-        f"Supported browsers: {', '.join(BROWSERS)}",
-        file=sys.stderr,
-    )
-    sys.exit(1)
-
-
-def write_cookie_file(cookies: list[dict], path: Path) -> None:
-    """Write cookies JSON and set owner-only permissions."""
-    path.parent.mkdir(parents=True, exist_ok=True)
-    path.write_text(json.dumps(cookies, indent=2, ensure_ascii=False))
-    os.chmod(path, stat.S_IRUSR | stat.S_IWUSR)  # 600 — cookies = full account access
-
-
-def main() -> None:
-    parser = argparse.ArgumentParser(
-        description="Export Continente.pt browser cookies for use with continente-mcp.",
-        formatter_class=argparse.RawDescriptionHelpFormatter,
-    )
-    parser.add_argument(
-        "--browser", "-b",
-        choices=BROWSERS,
-        metavar="BROWSER",
-        help=f"Browser to read from (default: auto-detect). One of: {', '.join(BROWSERS)}",
-    )
-    parser.add_argument(
-        "--output", "-o",
-        type=Path,
-        default=DEFAULT_OUTPUT,
-        help=f"Output path (default: {DEFAULT_OUTPUT})",
-    )
-    parser.add_argument(
-        "--list-browsers",
-        action="store_true",
-        help="List supported browsers and exit",
-    )
-    args = parser.parse_args()
-
-    if args.list_browsers:
-        print("Supported browsers: " + ", ".join(BROWSERS))
-        sys.exit(0)
-
-    try:
-        import browser_cookie3  # noqa: F401
-    except ImportError:
-        print(
-            "browser-cookie3 is not installed.\n"
-            "Run: pip install browser-cookie3",
-            file=sys.stderr,
-        )
-        sys.exit(1)
-
-    cookies, browser_name = find_cookies(args.browser)
-
-    write_cookie_file(cookies, args.output)
-    print(f"Saved {len(cookies)} cookies → {args.output}  (browser: {browser_name})")
-
-    # Optional: also write to a secondary path (e.g. for syncing to a remote machine)
-    if VAULT_PATH:
-        existing = []
-        if VAULT_PATH.exists():
-            try:
-                existing = json.loads(VAULT_PATH.read_text())
-            except Exception:
-                pass
-        if cookies != existing:
-            write_cookie_file(cookies, VAULT_PATH)
-            print(f"Saved {len(cookies)} cookies → {VAULT_PATH}  (vault sync)")
-        else:
-            print(f"Vault unchanged — skipping write to {VAULT_PATH}")
-
-
-if __name__ == "__main__":
-    main()

package/continente-keepalive.py

@@ -1,135 +0,0 @@
-#!/usr/bin/env python3
-"""
-Continente Session Keepalive
-
-Pings continente.pt to reset the idle session timeout.
-If the session has expired, re-reads cookies from your browser automatically.
-
-Run every 20 minutes via cron or launchd:
-  */20 * * * * python3 /path/to/continente-keepalive.py >> ~/.continente/keepalive.log 2>&1
-"""
-import json
-import os
-import stat
-import subprocess
-import sys
-from pathlib import Path
-
-import requests
-
-LOCAL_COOKIE = Path(
-    os.environ.get("CONTINENTE_COOKIES_PATH", str(Path.home() / ".continente" / "cookies.json"))
-)
-PING_URL = "https://www.continente.pt/on/demandware.store/Sites-continente-Site/default/Account-Show"
-CHECK_URL = "https://www.continente.pt/conta/encomendas/"
-LOG_PREFIX = "[continente-keepalive]"
-
-USER_AGENT = (
-    "Mozilla/5.0 (Windows NT 10.0; Win64; x64) "
-    "AppleWebKit/537.36 (KHTML, like Gecko) Chrome/121.0.0.0 Safari/537.36"
-)
-
-
-def log(msg: str) -> None:
-    print(f"{LOG_PREFIX} {msg}", flush=True)
-
-
-def load_cookies(path: Path) -> list[dict] | None:
-    if not path.exists():
-        return None
-    try:
-        data = json.loads(path.read_text())
-        return data if isinstance(data, list) else data.get("cookies")
-    except Exception as e:
-        log(f"Failed to read {path}: {e}")
-        return None
-
-
-def save_cookies(cookies: list[dict]) -> None:
-    LOCAL_COOKIE.parent.mkdir(parents=True, exist_ok=True)
-    LOCAL_COOKIE.write_text(json.dumps(cookies, indent=2, ensure_ascii=False))
-    os.chmod(LOCAL_COOKIE, stat.S_IRUSR | stat.S_IWUSR)
-
-
-def is_logged_in(session: requests.Session) -> bool:
-    try:
-        r = session.get(CHECK_URL, timeout=10, allow_redirects=True)
-        return "login" not in r.url and r.status_code == 200
-    except Exception:
-        return False
-
-
-def refresh_from_browser() -> list[dict] | None:
-    """Re-run continente-cookie-reader.py to pull fresh cookies from the browser."""
-    script = Path(__file__).parent / "continente-cookie-reader.py"
-    if not script.exists():
-        log("continente-cookie-reader.py not found next to this script.")
-        return None
-    try:
-        result = subprocess.run(
-            [sys.executable, str(script)],
-            capture_output=True, text=True, timeout=30
-        )
-        output = (result.stdout + result.stderr).strip()
-        if output:
-            log(f"Cookie reader: {output}")
-        return load_cookies(LOCAL_COOKIE)
-    except Exception as e:
-        log(f"Cookie reader failed: {e}")
-        return None
-
-
-def ping(cookies: list[dict]) -> bool:
-    session = requests.Session()
-    session.headers.update({
-        "User-Agent": USER_AGENT,
-        "Accept-Language": "pt-PT,pt;q=0.9",
-    })
-    for c in cookies:
-        if c.get("name") and c["name"] != "undefined":
-            session.cookies.set(c["name"], c["value"], domain=c.get("domain", ".continente.pt"))
-
-    logged_in = is_logged_in(session)
-    if logged_in:
-        try:
-            session.get(PING_URL, timeout=10)
-        except Exception:
-            pass
-    return logged_in
-
-
-def main() -> None:
-    cookies = load_cookies(LOCAL_COOKIE)
-
-    if not cookies:
-        log("No cookies found — trying to refresh from browser...")
-        cookies = refresh_from_browser()
-        if not cookies:
-            log("ERROR: No cookies found. Run continente-cookie-reader.py first.")
-            sys.exit(1)
-
-    if ping(cookies):
-        log("Session alive.")
-        save_cookies(cookies)
-        return
-
-    log("Session expired — refreshing from browser...")
-    cookies = refresh_from_browser()
-    if not cookies:
-        log("ERROR: Could not refresh cookies. Run continente-cookie-reader.py manually.")
-        sys.exit(1)
-
-    if ping(cookies):
-        save_cookies(cookies)
-        log("Session restored.")
-        return
-
-    log(
-        "ERROR: Still not logged in after refresh. "
-        "Log into Continente in your browser and re-run continente-cookie-reader.py."
-    )
-    sys.exit(1)
-
-
-if __name__ == "__main__":
-    main()