Selenium wrapper

The class FuncSeleniumMixin has some Selenium/full browser specific methods, as well as the Common WebTest/Selenium API.

class django_functest.FuncSeleniumMixin[source]

Configuration

These are class attributes and class methods that determine how the browser will be set up and run, and defaults for how some methods behave. It is easiest to override the class attribute, but the class method exists also for more involved needs. Note that most of these are used when setUpClass is called, and most of the related methods are therefore classmethods. Some are called within setUp, and some are used later when different methods are called. For these latter ones, setting self.<attribute> within the test body will be effective.

It is usually a good idea to set up a FuncSeleniumMixin sub-class in your project, to be used as a base-class for your tests, and set these configuration values on it.

Values that apply after test setup can be changed.

browser_window_size

If set, should be a tuple containing (width, height) in pixels. This will be used inside setUp to set the browser window size. Defaults to None.

default_timeout

Controls most Selenium timeouts, defaults to 10 (seconds).

display

Controls whether browser window is displayed or not, defaults to False.

Note that PhantomJS is always invisible, regardless of this value.

driver_name

Controls which Selenium ‘driver’ i.e. browser will be used. Defaults to "Firefox". You can also use "Chrome" if Chrome and chromedriver are installed, and "PhantomJS" if PhantomJS is installed.

auto_scroll_by_default

Controls whether submit() will automatically attempt to scroll to the element in order to click it.

Defaults to True.

This default behaviour is usually desirable, but depending on your styling, scrolling may not work, and Selenium may still be able to interact with the controls.

scroll_method

Defines the method to be used for scrolling. Defaults to "auto" which uses the most reliable known method.

Other values:

legacyWindowScrollTo — a legacy method based on window.scrollTo

display_browser_window()[source]

classmethod. Returns boolean that determines if the browser window should be shown. Defaults to display.

get_browser_window_size()[source]

Returns browser_window_size by default.

get_default_timeout()[source]

classmethod. Returns the time in seconds for Selenium to wait for the browser to respond etc. Defaults to default_timeout.

get_driver_name()[source]

classmethod. Returns the driver name i.e. the browser to use. Defaults to driver_name.

get_page_load_timeout()[source]

classmethod. Returns the time in seconds for Selenium to wait for the browser to return a page. Defaults to page_load_timeout.

page_load_timeout

Controls Selenium timeouts for loading page, defaults to 20 (seconds).

get_webdriver_options()[source]

Returns options to pass to the WebDriver class. Defaults to {}. This can be used to pass capabilities, firefox_binary or firefox_options if you are using the Firefox driver, for example.

Other attributes and methods

click(css_selector=None, xpath=None, text=None, text_parent_id=None, wait_for_reload=False, double=False, scroll=True, window_closes=False, expect_alert=True)[source]

Clicks the button or control specified by the CSS selector e.g.:

self.click("input.default")

Alternatively, xpath or text can be provided as keyword arguments, instead of a CSS selector e.g.:

self.click(xpath='//a[contains(text(), "kitten")]')
self.click(text="kitten")

Additionally, text_parent_id can be used in combination with text to limit the search to descendent elements of the one with the supplied id.

This method will attempt to scroll the window to make the element visible if scroll=True is passed (the default) - this is usually necessary for browsers to click controls correctly.

If double=True is passed, a double click will be performed. Note, this will simply be two clicks, like a user would, rather than the Selenium double_click action chain, which doesn’t actually trigger single click events.

If you are expecting the click to produce a standard browser “alert” or “confirm” dialog box, you should pass expect_alert=True.

See also the notes in submit() regarding wait_for_reload and window_closes (noting that the default values are different).

accept_alert()[source]

Chooses “OK” (or similar) on standard, browser-provided alert/confirm dialogue.

dismiss_alert()[source]

Chooses “Cancel” (or similar) on standard, browser-provided confirm dialogue.

execute_script(script, *args)[source]

Executes the supplied Javascript in the browser and returns the results.

If you need to pass arguments, you can receive them in the script using arguments e.g.:

self.execute_script("return arguments[0] + arguments[1];", 1, 2)

Arguments and return values are serialized and deserialized by Selenium.

hover(css_selector)[source]

Peform a mouse hover over the element specified by the CSS selector.

is_element_displayed(css_selector)[source]

Returns True if the element specified by the CSS selector is both present (see is_element_present()) and visible on the page (e.g. does not have display: none;), False otherwise.

save_screenshot(dirname='./', filename=None)[source]

Saves a screenshot of the browser window. By default, it is saved with a filename that includes a timestamp and the current test being run, into the current working directory, but this can be overridden by passing in a directory path and/or a filename. The full filename of the screenshot is returned.

set_window_size(width, height)[source]

Sets the browser window size to the specified width and height in pixels.

For PhantomJS browser, this sets the document size - there isn’t really a window.

switch_window(handle=None)[source]

Switches the browser window that has focus.

If there are only 2 windows, it can work out which window to switch to. Otherwise, you must pass in the window handle as the handle kwarg.

The method returns a tuple of (old_window_handle, new_window_handle) which can be used in subsequent calls to switch_window.

wait_for_page_load()[source]

Waits until the page has finished loading. You may want to override this to add extra things if a page has specific requirements.

wait_until(callback, timeout=None)[source]

Waits until the callback returns True, with a timeout that defaults to the default_timeout. The callback must accept a single parameter which will be the Selenium driver instance (which you may not need to use).

If the timeout is reached, a Selenium TimeoutException will be raised.

Example:

self.wait_until(lambda driver: self.is_element_present("#my-id"))

Often it is useful to wait until a specific assertion passes. You can do this conveniently with the provided method assertion_passes which converts an assertion method (plus arguments) into a callable that returns True on success. For example:

self.wait_until(self.assertion_passes(self.assertTextPresent, "Hello!", within="#my-id"))

This will call self.assertTextPresent("Hello!", within="#my-id") repeatedly until it passes i.e. no assertion is raised, or the timeout is reached.

wait_until_loaded(css_selector)[source]

Waits until an element matching the CSS selector appears.

Missing something?

The above may not be enough for your needs. In that case:

  • the self._driver attribute on the test class instance contains the Selenium driver/browser instance which you can use for lower level interactions.

  • feel free to open a ticket for feature requests, or to contribute back your utilities.