# pytest-dev/pytest-selenium

Switch branches/tags
Nothing to show
52c8acd Sep 6, 2018
4 contributors

### Users who have contributed to this file

641 lines (449 sloc) 20.9 KB

# User Guide

## Quick Start

The pytest-selenium plugin provides a method scoped selenium fixture for your tests. This means that any test with selenium as an argument will cause a browser instance to be invoked. The browser may run locally or remotely depending on your configuration, and may even run headless.

Here's a simple example test that opens a website using Selenium:

def test_example(selenium):
selenium.get('http://www.example.com')

To run the above test you will need to specify the browser instance to be invoked. For example, to run it using Firefox installed in a default location:

pytest --driver Firefox


For full details of the Selenium API you can refer to the documentation.

## Configuration Files

There are a number of options and values that can be set in an INI-style configuration file. For details of the expected name, format, and location of these configuration files, check the pytest documentation.

## Specifying a Base URL

To specify a base URL, refer to the documentation for the pytest-base-url plugin.

Note

By default, any tests using a base URL will be skipped. This is because all tests are considered destructive, and all environments are considered sensitive. See Sensitive Environments for further details.

## Sensitive Environments

To avoid accidental changes being made to sensitive environments such as your production instances, all tests are assumed to be destructive. Any destructive tests attempted to run against a sensitive environment will be skipped.

### Nondestructive Tests

To explicitly mark a test as nondestructive, you can add the appropriate marker as shown here:

import pytest
@pytest.mark.nondestructive
def test_nondestructive(selenium):
selenium.get('http://www.example.com')

### Indicating Sensitive Environments

Sensitive environments are indicated by a regular expression applied to the base URL or any URLs discovered in the history of redirects when retrieving the base URL. By default this matches all URLs, but can be configured by setting the SENSITIVE_URL environment variable, using a :ref:configuration file <configuration-files>, or by using the command line.

An example using a :ref:configuration file <configuration-files>:

[pytest]
sensitive_url = example\.com

An example using the command line:

pytest --sensitive-url "example\.com"


## Specifying a Browser

To indicate the browser you want to run your tests against you will need to specify a driver and optional capabilities.

### Firefox

To run your automated tests with Firefox version 47 or earlier, simply specify Firefox as your driver:

pytest --driver Firefox


For Firefox version 48 onwards, you will need to download GeckoDriver and selenium 3.0 or later. If the driver executable is not available on your path, you can use the --driver-path option to indicate where it can be found:

pytest --driver Firefox --driver-path /path/to/geckodriver


See the GeckoDriver documentation for more information.

#### Configuration

A firefox_options fixture is available to configure various options for Firefox. The following example demonstrates specifying a binary path, preferences, and a command line argument:

import pytest
@pytest.fixture
def firefox_options(firefox_options):
firefox_options.binary = '/path/to/firefox-bin'
firefox_options.set_preference('browser.anchor_color', '#FF0000')
return firefox_options

See the Firefox options API documentation for full details of what can be configured.

You can also use the firefox_preferences and firefox_arguments markers:

import pytest
@pytest.mark.firefox_arguments('-foreground')
@pytest.mark.firefox_preferences({'browser.anchor_color': '#FF0000'})
def test_firefox(selenium):
selenium.get('http://www.example.com')

### Chrome

To use Chrome, you will need to download ChromeDriver and specify Chrome for the --driver command line option. If the driver executable is not available on your path, you can use the --driver-path option to indicate where it can be found:

pytest --driver Chrome --driver-path /path/to/chromedriver


See the ChromeDriver documentation for more information.

#### Configuration

A chrome_options fixture is available to configure various options for Chrome. The following example demonstrates specifying a binary path, adding an extension, and passing an argument to start Chrome in kiosk mode:

import pytest
@pytest.fixture
def chrome_options(chrome_options):
chrome_options.binary_location = '/path/to/chrome'
return chrome_options

See the Chrome options API documentation for full details of what can be configured.

The ChromeDriver supports various command line arguments. These can be passed by implementing a driver_args fixture and returning a list of the desired arguments. The following example specifies the log level:

import pytest
@pytest.fixture
def driver_args():
return ['--log-level=LEVEL']

For a full list of supported command line arguments, run chromedriver --help in your terminal.

### Edge

To use Edge, you will need to download Edge WebDriver and specify Edge for the --driver command line option. If the driver executable is not available on your path, you can use the --driver-path option to indicate where it can be found:

pytest --driver Edge --driver-path \path\to\MicrosoftWebDriver.exe


### Internet Explorer

To use Internet Explorer, you will need to download and configure the Internet Explorer Driver and specify IE for the --driver command line option. If the driver executable is not available on your path, you can use the --driver-path option to indicate where it can be found:

pytest --driver IE --driver-path \path\to\IEDriverServer.exe


### PhantomJS

To use PhantomJS, you will need download it and specify PhantomJS for the --driver command line option. If the driver executable is not available on your path, you can use the --driver-path option to indicate where it can be found:

pytest --driver PhantomJS --driver-path /path/to/phantomjs


See the PhantomJS documentation for more information.

#### Configuration

PhantomJS supports various command line arguments. These can be passed by implementing a driver_args fixture and returning a list of the desired arguments. The following example specifies the log file path:

import pytest
@pytest.fixture
def driver_args():
return ['--webdriver-logfile=phantomjs.log']

For a full list of supported command line arguments, run phantomjs --help in your terminal.

### Safari

To use Safari, you will need to have at least Safari 10 running on OS X El Capitan or later, and selenium 3.0 or later. Once you have these prerequisites, simply specify Safari for the --driver command line option:

pytest --driver Safari


### Selenium Server/Grid

To run your automated tests against a Selenium server or a Selenium Grid you must have a server running and know the host and port of the server.

By default Selenium will listen on host 127.0.0.1 and port 4444. This is also the default when running tests against a remote driver.

To run your automated tests, simply specify Remote as your driver. Browser selection is determined using capabilities. Check the desired capabilities documentation for details of accepted values. There are also a number of browser specific capabilities that can be set. Be sure to also check the documentation for your chosen driver, as the accepted capabilities may differ:

pytest --driver Remote --capability browserName firefox


Note that if your server is not running locally or is running on an alternate port you will need to specify the --host and --port command line options:

pytest --driver Remote --host selenium.hostname --port 5555 --capability browserName firefox


### Sauce Labs

To run your automated tests using Sauce Labs, you must provide a valid username and API key. This can be done either by using a .saucelabs configuration file in the working directory or your home directory, by setting the SAUCELABS_USERNAME and SAUCELABS_API_KEY environment variables, or by using the environment variables as detailed here

Alternatively, when using Jenkins CI declarative pipelines, credentials can be set as environment variables as follows:

environment {
SAUCELABS = credentials('SAUCELABS')
}

For more information, see using environment variables in Jenkins pipelines.

#### Configuration

Below is an example .saucelabs configuration file:

[credentials]
key = secret

#### Running tests

To run your automated tests, simply specify SauceLabs as your driver:

pytest --driver SauceLabs --capability browserName Firefox


See the supported platforms to help you with your configuration. Additional capabilities can be set using the --capability command line arguments. See the test configuration documentation for full details of what can be configured.

#### W3C compliant capabilities

Starting with selenium version 3.11.0 Sauce Labs supports the selenium W3C compliant capabilities. To use the W3C capabilities you have to set the SAUCELABS_W3C environment variable to true and update your :ref:capabilities <capabilities> according to the Sauce labs W3C documentation.

#### Test result links

By default, links to Sauce Labs jobs are only visible to users logged in to the account that ran the job. To make a job visible without having to log in, you can create a link with an authentication token.

This can be configured by setting the SAUCELABS_JOB_AUTH environment variable or by using a :ref:configuration file <configuration-files>

An example using a :ref:configuration file <configuration-files>:

[pytest]
saucelabs_job_auth = token

You can also control the time to live for that link by setting the environment variable or :ref:configuration file <configuration-files>: value to day or hour.

Note that day means within the same day that the test was run, not "24 hours from test-run", likewise for hour

### BrowserStack

To run your automated tests using BrowserStack, you must provide a valid username and access key. This can be done either by using a .browserstack configuration file in the working directory or your home directory, or by setting the BROWSERSTACK_USERNAME and BROWSERSTACK_ACCESS_KEY environment variables.

Alternatively, when using Jenkins CI declarative pipelines, credentials can be set as environment variables as follows:

environment {
BROWSERSTACK = credentials('BROWSERSTACK')
}

For more information, see using environment variables in Jenkins pipelines.

#### Configuration

Below is an example .browserstack configuration file:

[credentials]
key = secret

#### Running tests

To run your automated tests, simply specify BrowserStack as your driver:

pytest --driver BrowserStack --capability browserName Firefox


See the capabilities documentation for additional configuration that can be set using --capability command line arguments.

### TestingBot

To run your automated tests using TestingBot, you must provide a valid key and secret. This can be done either by using a .testingbot configuration file in the working directory or your home directory, or by setting the TESTINGBOT_KEY and TESTINGBOT_SECRET environment variables.

Alternatively, when using Jenkins CI declarative pipelines, credentials can be set as environment variables as follows:

environment {
TESTINGBOT = credentials('TESTINGBOT')
}

Note that for TestingBot, username corresponds to key and password to secret.

For more information, see using environment variables in Jenkins pipelines.

#### Configuration

Below is an example .testingbot configuration file:

[credentials]
key = key
secret = secret

#### Running tests

To run your automated tests, simply specify TestingBot as your driver:

pytest --driver TestingBot --capability browserName firefox --capability version 39 --capability platform WIN8


See the list of available browsers to help you with your configuration. Additional capabilities can be set using the --capability command line arguments. See the test options for full details of what can be configured.

#### Local tunnel

To run the tests using TestingBot's local tunnel you'll also need to set the --host and --port command line arguments.

### CrossBrowserTesting

To run your automated tests using CrossBrowserTesting, you must provide a valid username and auth key. This can be done either by using a .crossbrowsertesting configuration file in the working directory or your home directory, or by setting the CROSSBROWSERTESTING_USERNAME and CROSSBROWSERTESTING_AUTH_KEY environment variables.

Alternatively, when using Jenkins CI declarative pipelines, credentials can be set as environment variables as follows:

environment {
CROSSBROWSERTESTING = credentials('CROSSBROWSERTESTING')
}

For more information, see using environment variables in Jenkins pipelines.

#### Configuration

Below is an example .crossbrowsertesting configuration file:

[credentials]
key = secret

#### Running tests

To run your automated tests, simply specify CrossBrowserTesting as your driver:

pytest --driver CrossBrowserTesting --capability os_api_name Win10 --capability browser_api_name FF46


Additional capabilities can be set using the --capability command line arguments. See the automation capabilities for full details of what can be configured.

## Specifying Capabilities

Configuration options are specified using a capabilities dictionary. This is required when using a Selenium server to specify the target environment, but can also be used to configure local drivers.

### Command Line Capabilities

Simple capabilities can be set or overridden on the command line:

pytest --driver Remote --capability browserName Firefox


### Capabilities Files

To specify capabilities, you can provide a JSON file on the command line using the pytest-variables plugin. For example if you had a capabilities.json containing your capabilities, you would need to include --variables capabilities.json on your command line.

The following is an example of a variables file including capabilities:

{ "capabilities": {
"browserName": "Firefox",
"platform": "MAC" }
}

### Capabilities Fixtures

The session_capabilities fixture includes capabilities that apply to the entire test session (including any command line or file based capabilities). Any changes to these capabilities will apply to every test. These capabilities are also reported at the top of the HTML report.

import pytest
@pytest.fixture(scope='session')
def session_capabilities(session_capabilities):
session_capabilities['tags'] = ['tag1', 'tag2', 'tag3']
return session_capabilities

The capabilities fixture contains all of the session capabilities, plus any capabilities specified by the capabilities marker. Any changes to these capabilities will apply only to the tests covered by scope of the fixture override.

import pytest
@pytest.fixture
def capabilities(capabilities):
capabilities['public'] = 'private'
return capabilities

### Capabilities Marker

You can add or change capabilities using the capabilities marker:

import pytest
@pytest.mark.capabilities(foo='bar')
def test_capabilities(selenium):
selenium.get('http://www.example.com')

## Common Selenium Setup

If you have common setup that you want to apply to your tests, such as setting the implicit timeout or window size, you can override the selenium fixture:

import pytest
@pytest.fixture
def selenium(selenium):
selenium.implicitly_wait(10)
selenium.maximize_window()
return selenium

## HTML Report

A custom HTML report is generated when the --html command line option is given. By default this will include additional debug information for failures.

### Debug Types

The following debug information is gathered by default when a test fails:

• URL - The current URL open in the browser.
• HTML - The HTML source of the page open in the browser.
• LOG - All logs available. Note that this will vary depending on the browser and server in use. See logging for more details.
• SCREENSHOT - A screenshot of the page open in the browser.

### Capturing Debug

To change when debug is captured you can either set selenium_capture_debug in a :ref:configuration file <configuration-files>, or set the SELENIUM_CAPTURE_DEBUG environment variable. Valid options are: never, failure (the default), and always. Note that always capturing debug will dramatically increase the size of the HTML report.

### Excluding Debug

You may need to exclude certain types of debug from the report. For example, log files can contain sensitive information that you may not want to publish. To exclude a type of debug from the report, you can either set selenium_exclude_debug in a :ref:configuration file <configuration-files>, or set the SELENIUM_EXCLUDE_DEBUG environment variable to a list of the Debug Types to exclude.

For example, to exclude HTML, logs, and screenshots from the report, you could set SELENIUM_EXCLUDE_DEBUG to html:logs:screenshot.