telecomkz_scraper/tools/inspect_page.py

"""
Diagnostic probes against the live WebView.

Consolidates what inspect_amplitude / inspect_bridge / inspect_window / inspect_vue_app /
read_amplitude_events / debug_cdp used to do separately. All of them shared the same
three defects: they assumed an adb forward already existed, they grabbed targets[0]
without checking its type, and they read one WebSocket frame after Page.enable and
treated whatever arrived as the command reply. The shared core fixes all three.

    py tools/inspect_page.py                 # everything
    py tools/inspect_page.py amplitude bridge
"""

import argparse
import json
import sys
from pathlib import Path

sys.path.insert(0, str(Path(__file__).resolve().parent))

import telecom_cdp as T

PROBES = {
    "page": """
(() => ({
  url: location.href,
  route: location.pathname,
  title: document.title,
  innerWidth: window.innerWidth,
  innerHeight: window.innerHeight,
  devicePixelRatio: window.devicePixelRatio,
  scrollHeight: document.documentElement.scrollHeight,
  scrollY: window.scrollY,
  readyState: document.readyState
}))()
""",
    "amplitude": """
(() => {
  const connectors = window.analyticsConnectorInstances || {};
  return {
    hasAmplitudeGlobal: typeof window.amplitude !== 'undefined',
    connectorInstances: Object.keys(connectors),
    connectorShape: Object.keys(connectors).reduce((acc, k) => {
      const inst = connectors[k] || {};
      acc[k] = {
        keys: Object.keys(inst).slice(0, 20),
        hasEventBridge: !!inst.eventBridge,
        hasSetEventReceiver: !!(inst.eventBridge && typeof inst.eventBridge.setEventReceiver === 'function')
      };
      return acc;
    }, {}),
    hasYandexMetrika: typeof window.ym === 'function',
    hasGtag: typeof window.gtag === 'function'
  };
})()
""",
    "bridge": """
(() => {
  const out = {};
  ['AndroidBridge', 'Android', 'Aitu', 'NativeBridge', 'webkit'].forEach(name => {
    const obj = window[name];
    if (!obj) return;
    const methods = [];
    for (const k in obj) {
      try { if (typeof obj[k] === 'function') methods.push(k); } catch (e) { /* guarded getter */ }
    }
    out[name] = methods.slice(0, 60);
  });
  return out;
})()
""",
    "window": """
(() => {
  const builtin = new Set(Object.getOwnPropertyNames(Object.getPrototypeOf(window)));
  const custom = Object.keys(window).filter(k => !builtin.has(k) && !k.startsWith('webkit'));
  return {
    customGlobals: custom.slice(0, 120),
    total: custom.length
  };
})()
""",
    "vue": """
(() => {
  const root = document.querySelector('#app') || document.body;
  const vue = root && (root.__vue_app__ || root.__vue__);
  return {
    hasVueApp: !!(root && root.__vue_app__),
    hasVueInstance: !!(root && root.__vue__),
    version: (window.Vue && window.Vue.version) || (vue && vue.version) || null,
    rootId: root ? root.id : null,
    routerPath: (window.$nuxt && window.$nuxt.$route && window.$nuxt.$route.path) || location.pathname,
    componentClasses: [...new Set([...document.querySelectorAll('[class]')]
      .map(e => String(e.className).split(' ')[0]).filter(Boolean))].slice(0, 60)
  };
})()
""",
    "storage": """
(() => {
  const pick = store => {
    const out = {};
    for (let i = 0; i < store.length; i++) {
      const k = store.key(i);
      if (!/amplitude|analytics|event|session|device/i.test(k)) continue;
      const v = store.getItem(k) || '';
      out[k] = v.length > 400 ? v.slice(0, 400) + '...(truncated)' : v;
    }
    return out;
  };
  return { localStorage: pick(window.localStorage), sessionStorage: pick(window.sessionStorage) };
})()
""",
}


def main():
    parser = argparse.ArgumentParser(description="Probe the TelecomKz WebView.")
    parser.add_argument("probes", nargs="*", choices=list(PROBES) + [], default=None)
    args = parser.parse_args()
    names = args.probes or list(PROBES)

    try:
        device, target = T.connect()
    except T.DeviceError as exc:
        print("Error: " + str(exc))
        return 1

    print("Device: " + device)
    print("Target: " + str(target.get("title")) + "  " + str(target.get("url")))

    with T.CdpSession(target["webSocketDebuggerUrl"]) as session:
        for name in names:
            print("\n=== " + name + " ===")
            try:
                value = session.evaluate("JSON.stringify(" + PROBES[name] + ")")
                print(json.dumps(json.loads(value), ensure_ascii=False, indent=2))
            except Exception as exc:
                print("probe failed: " + str(exc))
    return 0


if __name__ == "__main__":
    sys.exit(main())