To install the required dependencies, just run the following command on the python directory.

cd python
./upgrade_third_party.sh

Packages will be installed under tests/python/third_party

To run the tests on Windows run the run_tests.bat. Also you need to run it as admin in powershell with unrestricted execution policy (Set-ExecutionPolicy Unrestricted).

To run on Linux or macOS, run the script run_tests.sh.

Add -h to see options. Add --with-coverage to enable coverage.

To run specific test(s), specify module, or module.class or module.class.test:

run_tests.sh test_tank_content.TestValidate.test_valid_path

To run all the tests from a specific file, you can specify both the module or file path:

./run_tests.sh tank_module_test.test_module

./run_tests.sh tank_module_test/test_module.py

To make sure the tank source code can be compiled with Python 3, run the compile_python3.* scripts. Note that the unit tests cannot be executed right now with Python 3.

The tests directory follows the package layout of the tank code, with tests for top level tank modules being at the top level and tests for the tank.platform subpackage living in tests.platform.

Generally there is one test module per module in tank, with the test modules named with the target modules name pre-pended with "test_". It is possible to use more than one test module for a single target module if that module has multiple classes which require substantial tests. In general their is one test case class per method or function, though it is possible to define multiple if there is the need to test the same method with different setups.

There is a tank tests module, tank_test.tank_test_base.py which contains both a base class from which the test cases inherit and a module level setup. This module handles the creation of test fixture data, including studio level and project level configuration.

This is a lighter version of the TankTestBase, which only mocks a ShotGrid connection and does not write anything to disk. If you are testing parts of the API that doesn't need to write to disk, this is a faster option.

Among other features, this module includes:

Module level set up function which determines test data location and sets up studio level directories therein.

Setup method which creates a project, it's project level directories and mocked shotgun entity.

This method copies the config files, test apps and engines from the data directory into the test project.

This method adds a fake entity to the mocked shotgun, creates the entities path in test project and registers that entity with that path in the test project's path cache.

• Create a test module in the appropriate section of the tests area. If the module you wish to test is tank.platform.foo, the test module should be tests.platform.test_foo.
• In this module, import the base class and module setup method: from tank_test.tank_test_base import setUpModule, TankTestBase
• Create a test class inheriting from the TankTestBase class.
• If a setUp other than the base one is needed, be sure to call super(TestClassName, self).setUp() in order to allow the base class to setup the fixtures.