Playwright captcha and contexts

Author: kamath

examples/e2e/test_playwright.py

@@ -0,0 +1,37 @@
+import pytest
+from playwright.sync_api import Playwright, sync_playwright
+from dotenv import load_dotenv
+import os
+from browserbase import Browserbase
+
+from .. import (
+    BROWSERBASE_API_KEY,
+    playwright_basic,
+    playwright_captcha,
+    playwright_contexts,
+)
+
+bb = Browserbase(api_key=BROWSERBASE_API_KEY)
+load_dotenv()
+
+SKIP_CAPTCHA_SOLVING = os.getenv("SKIP_CAPTCHA_SOLVING", "false").lower() == "true"
+
+
+@pytest.fixture(scope="session")
+def playwright():
+    with sync_playwright() as p:
+        yield p
+
+
+def test_playwright_basic(playwright: Playwright):
+    playwright_basic.run(playwright)
+
+
+def test_playwright_captcha(playwright: Playwright):
+    if SKIP_CAPTCHA_SOLVING:
+        pytest.skip("Skipping captcha solving")
+    playwright_captcha.run(playwright)
+
+
+def test_playwright_contexts(playwright: Playwright):
+    playwright_contexts.run(playwright)

examples/e2e/test_playwright_basic.py

@@ -0,0 +1,61 @@
+from playwright.sync_api import Playwright, sync_playwright, ConsoleMessage
+
+
+from examples import (
+    BROWSERBASE_API_KEY,
+    BROWSERBASE_PROJECT_ID,
+    BROWSERBASE_CONNECT_URL,
+    bb,
+)
+
+DEFAULT_CAPTCHA_URL = "https://www.google.com/recaptcha/api2/demo"
+OVERRIDE_TIMEOUT = 60000  # 60 seconds, adjust as needed
+
+
+def run(playwright: Playwright):
+    # Create a session on Browserbase
+    session = bb.sessions.create(project_id=BROWSERBASE_PROJECT_ID)
+    assert session.id is not None
+    assert session.status == "RUNNING", f"Session status is {session.status}"
+
+    # Connect to the remote session
+    connect_url = (
+        f"{BROWSERBASE_CONNECT_URL}?sessionId={session.id}&apiKey={BROWSERBASE_API_KEY}"
+    )
+    chromium = playwright.chromium
+    browser = chromium.connect_over_cdp(connect_url)
+    context = browser.contexts[0]
+    page = context.pages[0]
+
+    captcha_solving_started = False
+    captcha_solving_finished = False
+
+    # Browserbase logs messages to the console to indicate when captcha solving has started and finished
+    # We can track these messages to know when the captcha solving has started and finished
+    def handle_console(msg: ConsoleMessage):
+        nonlocal captcha_solving_started, captcha_solving_finished
+        if msg.text == "browserbase-solving-started":
+            captcha_solving_started = True
+            page.evaluate("window.captchaSolvingFinished = false;")
+        elif msg.text == "browserbase-solving-finished":
+            captcha_solving_finished = True
+            page.evaluate("window.captchaSolvingFinished = true;")
+
+    page.on("console", handle_console)
+
+    page.goto(DEFAULT_CAPTCHA_URL, wait_until="networkidle")
+    page.wait_for_function(
+        "() => window.captchaSolvingFinished === true", timeout=OVERRIDE_TIMEOUT
+    )
+
+    assert captcha_solving_started, "Captcha solving did not start"
+    assert captcha_solving_finished, "Captcha solving did not finish"
+
+    page.close()
+    browser.close()
+    print("Done!")
+
+
+if __name__ == "__main__":
+    with sync_playwright() as playwright:
+        run(playwright)

examples/playwright_captcha.py

@@ -0,0 +1,127 @@
+from playwright.sync_api import Playwright, sync_playwright, Browser, Cookie
+from browserbase.types.session_create_params import (
+    BrowserSettings,
+    BrowserSettingsContext,
+)
+import time
+from typing import Optional
+
+from examples import (
+    BROWSERBASE_API_KEY,
+    BROWSERBASE_PROJECT_ID,
+    BROWSERBASE_CONNECT_URL,
+    bb,
+)
+
+CONTEXT_TEST_URL = "https://www.browserbase.com"
+SECOND = 1000
+
+
+def add_hour(date: float) -> int:
+    return int((date + 3600) * 1000) // SECOND
+
+
+def find_cookie(browser: Browser, name: str) -> Optional[Cookie]:
+    default_context = browser.contexts[0]
+    cookies = default_context.cookies()
+    return next((cookie for cookie in cookies if cookie.get("name") == name), None)
+
+
+def run(playwright: Playwright):
+    context_id = None
+    session_id = None
+    test_cookie_name = None
+    test_cookie_value = None
+
+    # Step 1: Creates a context
+    context = bb.contexts.create(project_id=BROWSERBASE_PROJECT_ID)
+    assert context.id is not None
+    context_id = context.id
+
+    uploaded_context = bb.contexts.retrieve(id=context_id)
+    assert uploaded_context.id == context_id
+
+    # Step 2: Creates a session with the context
+    session = bb.sessions.create(
+        project_id=BROWSERBASE_PROJECT_ID,
+        browser_settings=BrowserSettings(
+            context=BrowserSettingsContext(id=context_id, persist=True),
+        ),
+    )
+
+    assert (
+        session.context_id == context_id
+    ), f"Session context_id is {session.context_id}, expected {context_id}"
+    session_id = session.id
+
+    # Step 3: Populates and persists the context
+    print(f"Populating context {context_id} during session {session_id}")
+    connect_url = (
+        f"{BROWSERBASE_CONNECT_URL}?sessionId={session_id}&apiKey={BROWSERBASE_API_KEY}"
+    )
+    browser = playwright.chromium.connect_over_cdp(connect_url)
+    page = browser.contexts[0].pages[0]
+
+    page.goto(CONTEXT_TEST_URL, wait_until="domcontentloaded")
+
+    now = time.time()
+    test_cookie_name = f"bb_{int(now * 1000)}"
+    test_cookie_value = time.strftime("%Y-%m-%dT%H:%M:%S.000Z", time.gmtime(now))
+    browser.contexts[0].add_cookies(
+        [
+            {
+                "domain": ".browserbase.com",
+                "expires": add_hour(now),
+                "name": test_cookie_name,
+                "path": "/",
+                "value": test_cookie_value,
+            }
+        ]
+    )
+
+    assert find_cookie(browser, test_cookie_name) is not None
+
+    page.goto("https://www.google.com", wait_until="domcontentloaded")
+    page.go_back()
+
+    assert find_cookie(browser, test_cookie_name) is not None
+
+    page.close()
+    browser.close()
+
+    time.sleep(5)
+
+    # Step 4: Creates another session with the same context
+    session = bb.sessions.create(
+        project_id=BROWSERBASE_PROJECT_ID,
+        browser_settings=BrowserSettings(
+            context=BrowserSettingsContext(id=context_id, persist=True)
+        ),
+    )
+    assert (
+        session.context_id == context_id
+    ), f"Session context_id is {session.context_id}, expected {context_id}"
+    session_id = session.id
+
+    # Step 5: Uses context to find previous state
+    print(f"Reusing context {context_id} during session {session_id}")
+    connect_url = (
+        f"{BROWSERBASE_CONNECT_URL}?sessionId={session_id}&apiKey={BROWSERBASE_API_KEY}"
+    )
+    browser = playwright.chromium.connect_over_cdp(connect_url)
+    page = browser.contexts[0].pages[0]
+
+    page.goto(CONTEXT_TEST_URL, wait_until="domcontentloaded")
+
+    found_cookie = find_cookie(browser, test_cookie_name)
+    print(found_cookie)
+    assert found_cookie is not None
+    assert found_cookie.get("value") == test_cookie_value
+
+    page.close()
+    browser.close()
+
+
+if __name__ == "__main__":
+    with sync_playwright() as playwright:
+        run(playwright)