SeleniumBase/help_docs/how_it_works.md

How SeleniumBase Works 👁️

👁️🔎 At the core, SeleniumBase works by extending pytest as a direct plugin. SeleniumBase automatically spins up web browsers for tests (using Selenium WebDriver), and then gives those tests access to the SeleniumBase libraries through the BaseCase class. Tests are also given access to SeleniumBase command-line arguments and SeleniumBase methods, which provide additional functionality.

👁️🔎 pytest uses a feature called test discovery to automatically find and run Python methods that start with test_ when those methods are located in Python files that start with test_ or end with _test.py.

👁️🔎 The most common way of using SeleniumBase is by inheriting BaseCase:

from seleniumbase import BaseCase

Then have your test classes inherit BaseCase:

class MyTestClass(BaseCase):

Here's what a full test might look like:

from seleniumbase import BaseCase

class TestMFALogin(BaseCase):
    def test_mfa_login(self):
        self.open("https://seleniumbase.io/realworld/login")
        self.type("#username", "demo_user")
        self.type("#password", "secret_pass")
        self.enter_mfa_code("#totpcode", "GAXG2MTEOR3DMMDG")  # 6-digit
        self.assert_text("Welcome!", "h1")
        self.highlight("img#image1")  # A fancier assert_element() call
        self.click('a:contains("This Page")')  # Use :contains() on any tag
        self.save_screenshot_to_logs()  # ("./latest_logs" folder for test)
        self.click_link("Sign out")  # Link must be "a" tag. Not "button".
        self.assert_element('a:contains("Sign in")')
        self.assert_exact_text("You have been signed out!", "#top_message")

(See the example, test_mfa_login.py, for reference.)

👁️🔎 Here are some examples of running tests with pytest:

pytest --headless --rs --dashboard --html=report.html -v -n=4
pytest test_mfa_login.py
pytest -m marker2
pytest offline_examples/
pytest -k agent

(See SyntaxFormats for more ways of using SeleniumBase.)

---

✅ No More Flaky Tests!

SeleniumBase methods automatically wait for page elements to finish loading before interacting with them (up to a timeout limit). This means you no longer need random time.sleep() statements in your scripts.

There are three layers of protection that provide reliability for tests using SeleniumBase:

• (1): Selenium's default pageLoadStrategy is normal: This strategy causes Selenium to wait for the full page to load, with HTML content and sub-resources downloaded and parsed.

• (2): SeleniumBase includes methods such as wait_for_ready_state_complete() and wait_for_angularjs(), which run inside other SeleniumBase methods to ensure that it's safe to proceed with the next command.

• (3): SeleniumBase methods automatically wait for elements to be visible and interactable before interacting with those elements.

If you want to speed up your tests and you think the third level of protection is enough by itself, you can use command-line options to remove the first, the second, or both of those first two levels of protection:

• --pls=none --> Set pageLoadStrategy to "none": This strategy causes Selenium to return immediately after the initial HTML content is fully received by the browser.

• --sjw --> Skip JS Waits, which include wait_for_ready_state_complete() and wait_for_angularjs().

(NOTE: Those command-line options were added in SeleniumBase 4.2.0. Using both could lead to a 15% increase in test speed, but it could also lead to flaky tests, so use with caution.)

---