notebookVB

10PriceScraping/Rohlik/test_login.py

@@ -1,49 +1,99 @@
 """
-Reuse saved browser state (cookies + localStorage) so the Usercentrics cookie
-banner never appears and we stay logged in — same situation as a returning user.
+Reusable login flow for Rohlik.cz:
 
-- If auth_state.json is MISSING: opens a browser, you accept cookies + log in
-  manually, then press Enter to save the state.
-- If auth_state.json EXISTS: loads it and just verifies (no banner, logged in).
+  1. Load saved session (auth_state.json) if it exists.
+  2. Open the site and check whether we're already logged in.
+  3. If yes -> continue.
+  4. If no  -> log in via the JSON API, accept cookies, save the session, continue.
 """
 
+import json
 from pathlib import Path
-from playwright.sync_api import sync_playwright
-from config import BASE_URL, AUTH_STATE_PATH
+from playwright.sync_api import sync_playwright, BrowserContext, Page
+from config import BASE_URL, AUTH_STATE_PATH, ROHLIK_EMAIL, ROHLIK_PASSWORD
 
-auth_path = Path(AUTH_STATE_PATH)
-have_state = auth_path.exists()
+LOGIN_URL = f"{BASE_URL}/services/frontend-service/login"
+
+
+def is_logged_in(page: Page) -> bool:
+    return page.locator('text="Přihlásit se"').count() == 0
+
+
+def accept_cookies(page: Page):
+    """Accept the Usercentrics consent banner via its official JS API."""
+    result = page.evaluate('''async () => {
+        for (let i = 0; i < 20; i++) {
+            if (window.UC_UI && window.UC_UI.isInitialized && window.UC_UI.isInitialized()) break;
+            await new Promise(r => setTimeout(r, 250));
+        }
+        if (window.UC_UI && typeof window.UC_UI.acceptAllConsents === 'function') {
+            await window.UC_UI.acceptAllConsents();
+            await window.UC_UI.closeCMP();
+            return "accepted";
+        }
+        return "UC_UI not available";
+    }''')
+    # Wait for the banner to actually detach from the DOM (close animation ~1s)
+    try:
+        page.wait_for_selector('#usercentrics-cmp-ui', state='detached', timeout=5000)
+    except Exception:
+        pass
+    return result
+
+
+def api_login(context: BrowserContext) -> int:
+    resp = context.request.post(
+        LOGIN_URL,
+        data=json.dumps({"email": ROHLIK_EMAIL, "password": ROHLIK_PASSWORD}),
+        headers={"Content-Type": "application/json", "Accept": "application/json"},
+    )
+    return resp.status
+
+
+def ensure_logged_in(pw) -> tuple[BrowserContext, Page]:
+    auth_path = Path(AUTH_STATE_PATH)
+    have_state = auth_path.exists()
 
-with sync_playwright() as pw:
     browser = pw.chromium.launch(headless=False, args=["--start-maximized"])
-
     ctx_args = {"no_viewport": True}
     if have_state:
         ctx_args["storage_state"] = AUTH_STATE_PATH
     context = browser.new_context(**ctx_args)
     page = context.new_page()
 
-    print(f"Opening {BASE_URL} (state loaded: {have_state}) ...")
+    print(f"1) Opening site (saved session: {have_state}) ...")
     page.goto(BASE_URL, wait_until="domcontentloaded", timeout=60000)
-    page.wait_for_timeout(4000)
+    page.wait_for_timeout(3000)
 
-    banner_present = page.locator('#usercentrics-cmp-ui').count() > 0
-    is_logged_in = page.locator('text="Přihlásit se"').count() == 0
-    print(f"Cookie banner present: {banner_present}")
-    print(f"Logged in: {is_logged_in}")
+    if is_logged_in(page):
+        print("2) Already logged in from saved session — continuing.")
+        return context, page
 
-    if not have_state:
-        print("\n" + "=" * 60)
-        print("No saved state. Accept cookies + log in manually,")
-        print("then press Enter here to save the state.")
-        print("=" * 60)
-        input()
+    print("2) Not logged in — logging in via API ...")
+    status = api_login(context)
+    print(f"   Login API status: {status}")
+
+    page.goto(BASE_URL, wait_until="domcontentloaded", timeout=60000)
+    page.wait_for_timeout(3000)
+
+    print(f"3) Accepting cookies: {accept_cookies(page)}")
+
+    if is_logged_in(page):
         context.storage_state(path=AUTH_STATE_PATH)
-        print(f"Saved state to {AUTH_STATE_PATH}")
-        banner_present = page.locator('#usercentrics-cmp-ui').count() > 0
-        is_logged_in = page.locator('text="Přihlásit se"').count() == 0
-        print(f"  -> banner present now: {banner_present}, logged in: {is_logged_in}")
+        print("4) Logged in and session saved.")
+    else:
+        print("4) Login FAILED — check API status above.")
 
-    print("\nPress Enter to close browser...")
-    input()
-    browser.close()
+    return context, page
+
+
+if __name__ == "__main__":
+    with sync_playwright() as pw:
+        context, page = ensure_logged_in(pw)
+
+        print(f"\n   -> logged in: {is_logged_in(page)}")
+        print(f"   -> cookie banner present: {page.locator('#usercentrics-cmp-ui').count() > 0}")
+
+        print("\nReady to scrape. Press Enter to close browser...")
+        input()
+        context.browser.close()