TestProject OpenSDK for Python

TestProject is a Free Test Automation platform for Web, Mobile and API testing.

To get familiar with the TestProject, visit our main documentation website.

TestProject SDK is a single, integrated interface to scripting with the most popular open source test automation frameworks.

From now on, you can effortlessly execute Selenium and Appium native tests using a single automation platform that already takes care of all the complex setup, maintenance and configs.

With one unified SDK available across multiple languages, developers and testers receive a go-to toolset, solving some of the greatest challenges in open source test automation.

With TestProject SDK, users save a bunch of time and enjoy the following benefits out of the box:

• 100% open source and available as a PyPI project.
• 5-minute simple Selenium and Appium setup with a single Agent deployment.
• Automatic test reports in HTML/PDF format (including screenshots).
• Collaborative reporting dashboards with execution history and RESTful API support.
• Always up-to-date with the latest and stable Selenium driver version.
• A simplified, familiar syntax for both web and mobile applications.
• Complete test runner capabilities for both local and remote executions, anywhere.
• Cross platform support for Mac, Windows, Linux and Docker.
• Ability to store and execute tests locally on any source control tool, such as Git.

Getting started

To get started, you need to complete the following prerequisites checklist:

• Login to your account at https://app.testproject.io/ or register a new one.
• Download and install an Agent for your operating system or pull a container from Docker Hub.
• Run the Agent and register it with your Account.
• Get a development token from the Integrations / SDK page.

Installation

The TestProject Python SDK is available on PyPI. All you need to do is add it as a Python module using:

pip3 install testproject-python-sdk

and you're good to go.

Minimum Python version required is 3.6

Test Development

Using a TestProject driver is identical to using a Selenium driver. Once you have added the SDK as a dependency to your project, changing the import statement is enough in most cases.

You can create a TestProject-powered version of a test using Chrome by using the TestProject Chrome driver:

# from selenium import webdriver  <-- replace this import
from src.testproject.sdk.drivers import webdriver

def test_create_a_chrome_driver_instance():
    driver = webdriver.Chrome()
    # Your test code goes here
    driver.quit()

Here's an example of a complete test that is using the Chrome driver from the TestProject SDK:

from src.testproject.sdk.drivers import webdriver

def simple_test():
    driver = webdriver.Chrome()

    driver.get("https://example.testproject.io/web/")

    driver.find_element_by_css_selector("#name").send_keys("John Smith")
    driver.find_element_by_css_selector("#password").send_keys("12345")
    driver.find_element_by_css_selector("#login").click()

    passed = driver.find_element_by_css_selector("#logout").is_displayed()

    print("Test passed") if passed else print("Test failed")

    driver.quit()

if __name__ == "__main__":
    simple_test()

Drivers

The TestProject SDK overrides standard Selenium/Appium drivers with extended functionality.

The examples shown in this document are based on Chrome. The SDK works in the same way for all other supported browsers:

• Firefox
• Safari
• Edge
• Internet Explorer
• Android apps (using Appium)
• iOS apps (using Appium)
• Generic driver (for non-UI tests)

WebDriverWait

In order to use Selenium's WebDriverWait with TestProject SDK all you need to do is import it directly from the SDK classes instead of from selenium's libraries.

An example can be seen here.

Development token

The SDK uses a development token for communication with the Agent and the TestProject platform. To configure your development token for use with the SDK, you have to specify it in an environment variable TP_DEV_TOKEN.

Alternatively, you can pass in your developer token as an argument to the driver constructor:

def test_create_a_chrome_driver_instance():
    driver = webdriver.Chrome(token='YOUR_TOKEN_GOES_HERE')
    # Your test code goes here
    driver.quit()

> If a token is specified in both the driver constructor and in an environment variable, the token in the environment will be the one used.

TestProject Agent

By default, drivers communicate with the local Agent listening on http://localhost:8585. This value can be overridden by setting the TP_AGENT_URL environment variable to the correct Agent address, or pass it in the driver constructor.

Remote Agent

By default, drivers communicate with the local Agent listening on http://localhost:8585.

Agent URL (host and port), can be also provided explicitly using this constructor using agent_url parameter:

driver = webdriver.Chrome(agent_url='http://remote-ip:port')

NOTE: By default, the agent binds to localhost. In order to allow the SDK to communicate with agents running on a remote machine (On the same network), the agent should bind to an external interface. For additional documentation on how to achieve such, please refer [here](https://docs.testproject.io/testproject-agents/testproject-agent-cli#start)

Remote (Cloud) Driver

By default, TestProject Agent communicates with the local Selenium or Appium server. In order to initialize a remote driver for cloud providers such as SauceLabs or BrowserStack, a custom capability cloud:URL should be set, for example:

SauceLabs

def driver():
    chrome_options = ChromeOptions()
    chrome_options.set_capability("cloud:URL", "https://{USERNAME}:{PASSWORD}@ondemand.us-west-1.saucelabs.com:443/wd/hub")
    driver = webdriver.Chrome(chrome_options=chrome_options, projectname="Examples")
    yield driver
    driver.quit()

BrowserStack

def driver():
    chrome_options = ChromeOptions()
    chrome_options.set_capability("cloud:URL", "https://{USERNAME}:{PASSWORD}@hub-cloud.browserstack.com/wd/hub")
    driver = webdriver.Chrome(chrome_options=chrome_options, projectname="Examples")
    yield driver
    driver.quit()

Reports

By default, the TestProject SDK reports all executed driver commands and their results to the TestProject Cloud. This allows us to create and display detailed HTML reports and statistics in your project dashboards.

Reports can be completely disabled using this driver constructor:

def test_disable_reporting():
    driver = webdriver.Chrome(disable_reports=True)
    # no reports will be created for this test
    driver.quit()

Implicit project and job names

The SDK will attempt to infer Project and Job names when you use pytest or unittest. For example:

• when using pytest, tests in the my_tests.py module in the e2e_tests/chrome package will be reported with a project name e2e_tests.chrome and job name my_tests.
• when using unittest, tests in the my_tests.py module in the e2e_tests/chrome package will be reported with a project name chrome and job name my_tests.

Examples using inferred project and job names:

• pytest
• unittest

Explicit project and job names

Project and Job names can be also specified explicitly using this constructor:

def test_specify_project_and_job_names_in_driver_constructor():
    driver = webdriver.Chrome(projectname='My custom project', jobname='My custom job')
    # Your test code goes here
    driver.quit()

or using the @report decorator:

from src.testproject.decorator import report

@report(project='My project', job='My job')
def test_specify_project_and_job_name_in_decorator():
    driver = webdriver.Chrome()
    # Your test code goes here
    driver.quit()

Examples using explicitly specified project and job names:

• pytest
• unittest

Reporting extensions

Reporting extensions extend the TestProject SDK reporting capabilities by intercepting unit testing framework assertion errors and reporting them as failed steps.

This functionality can be added by decorating your test method with the @report_assertion_errors decorator.

This decorator has an optional boolean argument 'screenshot' that will decide if failed assertions will include screenshots in the report.

from src.testproject.decorator import report_assertion_errors

@report_assertion_errors
def test_automatically_report_assertion_error():
    driver = webdriver.Chrome()
    assert 1 == 2  # This assertion will be reported automatically as a failed step
    driver.quit()

@report_assertion_errors(screenshot=False)
def test_automatically_report_assertion_error_without_screenshots():
    driver = webdriver.Chrome()
    assert 1 == 2  # This assertion will be reported automatically as a failed step and no screenshot will be taken
    driver.quit()

Here is a working example for pytest, and here is one for unittest.

Please make sure to follow the advice given here to ensure correct test name reporting.

Test reports

Automatic test reporting

Tests are reported automatically when a test ends or when the quit() command is called on the driver. This behavior can be overridden or disabled (see the Disabling Reports section below).

In order to determine whether a test has ended, the call stack is inspected, searching for the current test method. When the test name is different from the latest known test name, it is concluded that the execution of the previous test has ended. This is supported for both pytest and unittest.

To override the inferring of the test name and specify a custom test name instead, you can use the @report decorator:

from src.testproject.decorator import report

@report(test='My test name')
def test_specify_test_name_in_decorator():
    driver = webdriver.Chrome()
    # Your test code goes here
    driver.quit()

Here is a complete example using automatic reporting.

Manual test reporting

To report tests manually, you can use driver.report().test():

def test_report_test_manually():
    driver = webdriver.Chrome()
    # Your test code goes here
    driver.report().test(name='My test name', passed=True)
    driver.quit()

Reporting steps

Steps are reported automatically for every driver commands that is executed. If this feature is disabled, or you would like to add steps manually, you can use driver.report().step():

def test_report_step_manually():
    driver = webdriver.Chrome()
    # Your test code goes here
    driver.report().step(description='My step description', message='An additional message', passed=False,
                         screenshot=True, element=element_search_criteria_object, inputs=dict_of_input_parameters,
                         outputs=dict_of_output_parameters)
    driver.quit()

Here is a complete example using manual test reporting of tests and steps.

Step settings

Step settings allow controlling driver default execution and reporting behavior such as:

• Default timeout.
• Sleep duration Before/After step execution.
• Screenshot capturing logic.
• Execution result inversion.

Here is an example on how to take a screenshot upon any driver command executed:

def test_use_step_settings():
    driver = webdriver.Chrome()
    # Using StepSettings for the whole test.
    driver.step_settings = StepSettings(screenshot_condition=TakeScreenshotConditionType.Always)
    # Your test code goes here - all driver commands will use the defined step_settings
    driver.quit()

Single step settings override

For convenience we can also use the StepSettings inside a 'with' compound statement called DriverStepSettings.

Here is an example on how a single step can be used with different step settings. By default, screenshots are taken on step failures only, the following example demonstrates how to override this behavior and take a screenshot when a step passes:

def test_use_single_step_settings():
    driver = webdriver.Chrome()
    # A single step we want to run with an overriding StepSettings.
    with DriverStepSettings(driver, StepSettings(screenshot_condition=TakeScreenshotConditionType.Success)):
    driver.get("https://example.testproject.io/web/")  # Screenshot will be taken only if step passes.

    driver.get("https://example.testproject.io/web/")  # Screenshot will be taken only if step fails (default).

Disabling reports

If reports were not disabled when the driver was created, they can be disabled or enabled later. However, if reporting was explicitly disabled when the driver was created, they cannot be enabled later.

Disable all reports

The following will temporarily disable all reporting:

def test_temporarily_disable_all_reporting_then_reenable_it_later():
    driver = webdriver.Chrome()
    driver.report().disable_reports(True)
    driver.find_element_by_id('your_element_id').click()  # This statement will not be reported
    driver.report().disable_reports(False)
    driver.quit()

Disable automatic test reports

The following will disable automatic reporting of tests. All steps will end up in a single test report, unless tests are reported manually using driver.report().test():

def test_disable_automatic_test_reporting():
    driver = webdriver.Chrome()
    driver.report().disable_auto_test_reports(True)
    # Tests will not be reported automatically from here on
    driver.quit()

Disable driver command reports

The following will disable driver command reporting, which results in the reporting of tests that will have no steps, unless reported manually using driver.report().step():

def test_disable_automatic_reporting():
    driver = webdriver.Chrome()
    driver.report().disable_command_reports(True)
    # From here on, driver commands will not be reported automatically
    driver.quit()

Disable driver command redaction

When driver commands are being reported, the SDK will, by default, redact the values typed into sensitive elements by replacing the actual text with three asterisks (***) in the report. Elements are considered sensitive if they:

• have an attribute type with value password (all browsers and platforms)
• are of type XCUIElementTypeSecureTextField (iOS / XCUITest only)

This redaction of sensitive commands can be disabled, if desired:

def test_disable_driver_command_report_redaction():
    driver = webdriver.Chrome()
    driver.report().disable_redaction(True)
    # From here on, driver commands will not be redacted
    driver.quit()

If no test name is specified using the decorator, the test method name will be used as the test name in the report.

The importance of using quit()

Even more so than with regular Selenium- or Appium-based tests, it is important to make sure that you call the quit() method of your TestProject driver object at the end of every test that uses the TestProject SDK.

Upon calling quit(), the SDK will send all remaining report items to the Agent, ensuring that your report on the TestProject platform is complete.

Tip for pytest users: use a pytest fixture to ensure that quit() is called at the end of the test, even when an error occurred during test execution:

import pytest

@pytest.fixture
def driver():
    driver = webdriver.Chrome()
    yield driver
    driver.quit()

def test_using_pytest_fixture(driver):
    driver.get("https://example.testproject.io/web")

Tip for unittest users: use the setUp() and tearDown() methods for driver creation and destroying:

import unittest

class ChromeTest(unittest.TestCase):

    def setUp(self):
        self.driver = webdriver.Chrome()

    def test_using_unittest_setup_and_teardown(self):
        driver.get("https://example.testproject.io/web")

    def tearDown(self):
        self.driver.quit()

Cloud and Local Report

By default, the execution report is uploaded to the cloud, and a local report is created, as an HTML file in a temporary folder.

At the end of execution, the report is uploaded to the cloud and SDK outputs to the console/terminal the path for a local report file:

Execution Report: {temporary_folder}/report.html

This behavior can be controlled, by requesting only a LOCAL or only a CLOUD report.

When the Agent is offline, and only a _cloud_ report is requested, execution will fail with appropriate message.

Via a driver constructor:

driver = webdriver.Chrome(report_type=ReportType.LOCAL)

Control Path and Name of Generated Report

By default, the local reports name is the timestamp of the test execution, and the path is the reports directory in the agent data folder.

The SDK provides a way to override the default values of the generated local reports name and path.

Via driver constructor:

driver = ChromeDriver(chrome_options=ChromeOptions(), report_name="Python Local report", report_path="/my_executions/reports);

Logging

The TestProject Python SDK uses the logging framework built into Python. The default logging level is INFO and the default logging format is %(asctime)s %(levelname)s %(message)s, which results in log entries formatted like this:

13:37:45 INFO Using http://localhost:8585 as the Agent URL

If you wish, you can override the default log configuration:

• For pytest users, it is recommended to provide alternative values in your pytest.ini
• Users of unittest can override the configuration by setting the TP_LOG_LEVEL and / or TP_LOG_FORMAT environment variables, respectively, to the desired values

See this page for a list of accepted logging levels and look here for more information on how to define a custom logging format.

Behave Support

The SDK also supports automatic reporting of Behave features, scenarios and steps using the @behave_reporter decorator.

It will disable the reporting of driver commands and automatic reporting of tests. Instead, it will report:

• A test for every scenario in a feature file
• All steps in a scenario as steps in the corresponding test
• Steps are automatically marked as passed or failed, to create comprehensive living documentation from your specifications on TestProject Cloud.

To enable Behave feature reporting, in your environment.py decorate one or more of the following methods:

• method used to initialize your driver (usually before_all or before_feature to store the driver in the behave context)

• after_step

• after_scenario

  Storing the driver in the context provides direct access to the driver throughout the program such as in the step implementations.

@behave_reporter
def before_all(context):
    context.driver = webdriver.Chrome(projectname="Behave BDD")


@behave_reporter
def after_step(context, step):
    pass


@behave_reporter
def after_scenario(context, scenario):
    pass

By default, screenshots are taken only when step fail in your test, if you would like to change the behavior to always take a screenshot, pass the screenshot argument as True in your decorator.

@behave_reporter(screenshot=True)
def after_step(context, step):
    pass

Addon Proxy

One of the greatest features of the TestProject platform is the ability to execute a code written by someone else. It can be your account colleagues, writing actions that you can reuse, or TestProject community users creating addons and solving common automation challenges.

To get started, download a source file with the proxy class for the Action(s) you want to execute. It can be done by navigating to the Addons page, opening an addon, and clicking on the Proxy link at the bottom left corner of the popup.

Now, let's pretend that one of your colleagues coded and uploaded an Addon naming it - Example Addon. To use it in your test, download its proxy source file, add it to your project and invoke the actions using the following driver method:

driver.addons().execute()

That expects an instance of the ActionProxy class. For example:

# Use Addon proxy to invoke 'Clear Fields' Action
driver.addons().execute(ClearFieldsAction())

Another example using 'Type Random Phone' Action:

# Use Addon proxy to invoke 'Type Random Phone' Action
# Notice how the action parameters are provided using an action proxy convenience method
driver.addons().execute(TypeRandomPhoneAction("1", 10), by=By.CSS_SELECTOR, by_value="#phone")

Refer to the Addon Proxy Test for complete example source.

Examples

Here is a list of all examples for the different drivers that are supported by this SDK:

Web

• Chrome test
• Firefox test
• Safari test
• Edge test
• Internet Explorer test
• Addon Proxy Test

Android

• Android native test
• Android native app
• Web test on mobile Chrome

iOS

• iOS native test
• iOS native app
• Web test on mobile Safari

License

The TestProject Python SDK is licensed under the LICENSE file in the root directory of the project source tree.