Plugin#

Note

The --slow option conflicts in marker name with --run-slow provided by pytest-astropy. If you have both ci-watson and pytest-astropy installed, you need to provide both option flags to enable tests marked as slow. See spacetelescope/ci_watson#83 ,

The plugin portion of ci_watson contains:

  • --slow option and @pytest.mark.slow decorator to run or skip tests that are resource intensive. What counts as resource intensive is up to the author of the test.

  • --bigdata option and @pytest.mark.bigdata decorator to run or skip tests that require intranet (Artifactory, Central Storage, etc) access. Additional setup is required for these tests to run successfully (see Big Data). It is up to the author of the test to perform such setup properly.

  • --env option and envopt fixture to set the test environment to "dev" or "stable". This plugin only sets the value. It is up to the author of the test to use this environment setting properly.

  • _jail fixture to enable a test to run in a pristine temporary working directory. This is particularly useful for pipeline tests.

  • resource_tracker and log_tracked_resources fixtures to track memory and runtime and log them in the junit XML results file.

Configuration Options#

inputs_root/results_root - The ‘bigdata’ remote repository name/local data root directory for testing input/output files. Setting the value of either option will make it availble to tests via the pytestconfig fixture. Test code can then obtain the name of the artifactory repository/local data root directory to use when accessing locations needed for running tests.

Note

If used, these values should appear in either pytest.ini OR the appropriate section in pyproject.toml, not both.

Example configuration within pyproject.toml:

[tool.pytest.ini_options]
inputs_root = my_data_repo
results_root = my_results_repo

Example configuration within pytest.ini:

[pytest]
inputs_root = my_data_repo
results_root = my_results_repo

The value(s) defined in the pytest configuration file may be accessed as a list by test code via the pytestconfig fixture which must be passed in as an argument to the test method or function that will use the value.

Example of accessing configuration values within test code itself:

def test_important_thing(pytestconfig):
    setup_cfg_inputs_root = pytestconfig.getini('inputs_root')[0]
    assert setup_cfg_inputs_root == 'my_data_repo'

From within a fixture or a test class the configuration values must be accessed using a slightly different approach:

import pytest
inputs_root = pytest.config.getini('inputs_root')[0]
On this page