gigafibre-fsm/services/cogeco-checker/lib/cogeco-session.js

'use strict'
/**
 * cogeco-session.js — drives Cogeco's public address-availability checker with
 * a real headless Chromium (Playwright) so the reCAPTCHA token (x-rc-token)
 * and short-lived JWT the endpoint requires are generated legitimately by the
 * page's own JS. A pure HTTP call can't produce those, hence the browser.
 *
 * Flow (reverse-engineered 2026-06):
 *   1. load /en/internet/packages
 *   2. click "Check Availability" → address dialog
 *   3. type the address into the autocomplete combobox
 *   4. pick the first suggestion (triggers GET /boutique/api/address/search
 *      then the serviceability lookup)
 *   5. capture the JSON responses + the rendered result text
 *
 * We intercept every /boutique/api/* and /api/check-avail/* response and also
 * read the visible result, then return a normalized verdict. Cogeco can change
 * this flow at any time — treat parsing defensively and keep `raw` for debug.
 */

// playwright-extra + stealth masks the headless automation signals
// (navigator.webdriver, missing plugins, headless UA quirks) that reCAPTCHA
// Enterprise scores against. Falls back to vanilla playwright if the stealth
// stack isn't installed.
let chromium
try {
  chromium = require('playwright-extra').chromium
  const stealth = require('puppeteer-extra-plugin-stealth')()
  chromium.use(stealth)
} catch {
  chromium = require('playwright').chromium
}

const PAGE_URL = 'https://www.cogeco.ca/en/internet/packages'
const NAV_TIMEOUT = 45000
const STEP_TIMEOUT = 20000

let _browser = null
async function getBrowser () {
  if (_browser && _browser.isConnected()) return _browser
  _browser = await chromium.launch({
    headless: true,
    args: ['--no-sandbox', '--disable-dev-shm-usage', '--disable-blink-features=AutomationControlled'],
  })
  return _browser
}

// Normalize Cogeco's serviceability payload into a stable verdict. The exact
// shape varies, so we probe several likely fields and fall back to scanning
// the captured JSON + UI text for availability keywords + speed numbers.
function interpret (captured, uiText) {
  const verdict = { available: null, max_download_mbps: null, plans: [], confidence: 'low' }

  // 1. Look for an explicit serviceability object in the captured responses.
  for (const c of captured) {
    const b = c.body
    if (!b || typeof b !== 'object') continue
    const flat = JSON.stringify(b).toLowerCase()
    // Common serviceability flags
    if (verdict.available === null) {
      if (/"serviceable"\s*:\s*true|"available"\s*:\s*true|"iseligible"\s*:\s*true|"qualified"\s*:\s*true/.test(flat)) {
        verdict.available = true; verdict.confidence = 'high'
      } else if (/"serviceable"\s*:\s*false|"available"\s*:\s*false|"iseligible"\s*:\s*false|"qualified"\s*:\s*false/.test(flat)) {
        verdict.available = false; verdict.confidence = 'high'
      }
    }
    // Speed markers anywhere in the payload (e.g. download 1000)
    const speeds = [...flat.matchAll(/"(?:download|downloadspeed|speed|maxspeed)"\s*:\s*"?(\d{2,5})"?/g)].map(m => parseInt(m[1], 10))
    if (speeds.length) verdict.max_download_mbps = Math.max(verdict.max_download_mbps || 0, ...speeds)
  }

  // 2. Fall back to the rendered result text.
  if (verdict.available === null && uiText) {
    const t = uiText.toLowerCase()
    if (/available|disponible|good news|great news|we('| a)re in your area|select your plan|choose your/i.test(t)) {
      verdict.available = true; verdict.confidence = 'medium'
    } else if (/not available|non disponible|unfortunately|pas (encore )?disponible|sorry/i.test(t)) {
      verdict.available = false; verdict.confidence = 'medium'
    }
  }
  return verdict
}

async function checkAddress (address, { debug = false } = {}) {
  const browser = await getBrowser()
  // Fresh context per check — avoids carrying a stale reCAPTCHA/session score
  // between addresses and keeps each lookup independent.
  const ctx = await browser.newContext({
    locale: 'en-CA',
    userAgent: 'Mozilla/5.0 (Macintosh; Intel Mac OS X 10_15_7) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/149.0.0.0 Safari/537.36',
    viewport: { width: 1280, height: 900 },
  })
  const page = await ctx.newPage()
  const captured = []
  page.on('response', async (resp) => {
    const u = resp.url()
    if (/\/(boutique\/api|api\/check-avail)\//.test(u)) {
      let body = null
      try { body = await resp.json() } catch { try { body = (await resp.text()).slice(0, 2000) } catch { /* ignore */ } }
      captured.push({ url: u, status: resp.status(), body })
    }
  })

  const result = { address, queried_at: new Date().toISOString() }
  try {
    await page.goto(PAGE_URL, { waitUntil: 'domcontentloaded', timeout: NAV_TIMEOUT })

    // Dismiss a cookie/consent banner if present (best-effort, non-fatal).
    for (const label of [/accept all/i, /accept/i, /agree/i, /tout accepter/i, /j'accepte/i]) {
      const btn = page.getByRole('button', { name: label })
      if (await btn.count().catch(() => 0)) { await btn.first().click().catch(() => {}); break }
    }

    // Open the address dialog.
    await page.getByRole('button', { name: /check availability/i }).first()
      .click({ timeout: STEP_TIMEOUT })

    // Wait for the dialog, then target the combobox inside it (more robust
    // than matching the accessible name, which differs EN/FR).
    const dialog = page.getByRole('dialog')
    await dialog.waitFor({ state: 'visible', timeout: STEP_TIMEOUT }).catch(() => {})
    const input = (await dialog.count().catch(() => 0))
      ? dialog.getByRole('combobox').first()
      : page.getByRole('combobox', { name: /address|adresse/i })
    await input.waitFor({ state: 'visible', timeout: STEP_TIMEOUT })
    await input.fill('')
    await input.pressSequentially(address, { delay: 60 })

    // Wait for autocomplete suggestions, then pick the first one.
    let picked = false
    try {
      const firstOption = page.getByRole('option').first()
      await firstOption.waitFor({ state: 'visible', timeout: 8000 })
      await firstOption.click()
      picked = true
    } catch {
      // No dropdown option appeared — try pressing ArrowDown+Enter as a fallback.
      try { await input.press('ArrowDown'); await input.press('Enter'); picked = true } catch { /* ignore */ }
    }

    // Give the serviceability lookup time to fire + render.
    await page.waitForTimeout(5000)

    // Grab the visible result text (whatever the page now shows).
    const uiText = (await page.locator('body').innerText().catch(() => '') || '').slice(0, 4000)

    Object.assign(result, interpret(captured, uiText), { picked_suggestion: picked })
    if (debug) {
      result.captured = captured
      result.ui_excerpt = uiText.slice(0, 1200)
      result.screenshot = (await page.screenshot({ fullPage: false }).catch(() => null))?.toString('base64') || null
    }
  } catch (e) {
    result.error = e.message
    if (debug) {
      result.captured = captured
      try { result.ui_excerpt = (await page.locator('body').innerText()).slice(0, 1200) } catch { /* ignore */ }
    }
  } finally {
    await ctx.close().catch(() => {})
  }
  return result
}

async function shutdown () {
  if (_browser) { await _browser.close().catch(() => {}); _browser = null }
}

module.exports = { checkAddress, shutdown }