Running icepyx tests

[rwegener2]

Goal

I'm working on running the tests for icepyx and I'm looking for some guidance. I hit some errors so I'm wondering if these are expected when running the tests outside of Travis and, if they are, how easy it would be for someone to set up a local configuration that would allow them to run the tests successfully. While not necessary, it would be nice to have as someone trying to add to the package without breaking stuff 🙂

Process thus far

The steps I've identified thus far are:

1. clone/fork the icepyx repository
2. create a local enviroment conda create -n icepyx-dev (do we have a recommended version of python?)
3. install the dev requirements doc conda install --file requirements-dev.txt (it's possible steps 2 and 3 could be combined, but I haven't tested this)
4. install a local editable version of icepyx in the development environment pip install -e . from inside the environment
5. navigate to the folder icepyx/tests
6. run all the tests with pytest .

Result

When I do this locally right now on the development branch I get 2 failed tests and 2 errors:

================================================ short test summary info ================================================
FAILED icepyx/tests/test_read.py::test_check_run_fast_scandir[./icepyx/-is2*.py-expect0] - AssertionError: assert (['./icepyx/c...ss_query.py']) == (['./icepyx/c...ss_query.py'])
FAILED icepyx/tests/test_read.py::test_check_run_fast_scandir[./icepyx-bogus_glob-expect2] - AssertionError: assert (['./icepyx/c...st_data'], []) == (['./icepyx/c...x/tests'], [])
ERROR icepyx/tests/test_behind_NSIDC_API_login.py::test_get_custom_options_output - AttributeError: 'Query' object has no attribute '_session'
ERROR icepyx/tests/test_behind_NSIDC_API_login.py::test_download_granules_with_subsetting - AttributeError: 'Query' object has no attribute '_session'
================================= 2 failed, 160 passed, 13 warnings, 2 errors in 45.43s =================================

test_behind_NSIDC_API_login.py errors

For the two test_behind_NSIDC_API_login.py errors I'm wondering if the errors are related to the fact that Travis CI might have stored credentials. Is that the case? If so, any recommendations for users to setup their environment to emulate the test environment?

test_read.py failed tests

Output from one of the test_read.py errors is:

________________________________ test_check_run_fast_scandir[./icepyx/-is2*.py-expect0] _________________________________

dir = './icepyx/', fn_glob = 'is2*.py'
expect = (['./icepyx/core', './icepyx/quest', './icepyx/quest/dataset_scripts', './icepyx/tests'], ['./icepyx/core/is2cat.py', './icepyx/core/is2ref.py', './icepyx/tests/is2class_query.py'])

    @pytest.mark.parametrize(
        "dir, fn_glob, expect",
        [
            (
                "./icepyx/",
                "is2*.py",
                (
                    sorted(
                        [
                            "./icepyx/core",
                            "./icepyx/quest",
                            "./icepyx/quest/dataset_scripts",
                            "./icepyx/tests",
                        ]
                    ),
                    sorted(
                        [
                            "./icepyx/core/is2cat.py",
                            "./icepyx/core/is2ref.py",
                            "./icepyx/tests/is2class_query.py",
                        ]
                    ),
                ),
            ),
            (
                "./icepyx/core",
                "is2*.py",
                ([], ["./icepyx/core/is2cat.py", "./icepyx/core/is2ref.py"]),
            ),
            (
                "./icepyx",
                "bogus_glob",
                (
                    [
                        "./icepyx/core",
                        "./icepyx/quest",
                        "./icepyx/quest/dataset_scripts",
                        "./icepyx/tests",
                    ],
                    [],
                ),
            ),
        ],
    )
    def test_check_run_fast_scandir(dir, fn_glob, expect):
    
        (subfolders, files) = read._run_fast_scandir(dir, fn_glob)
>       assert (sorted(subfolders), sorted(files)) == expect
E       AssertionError: assert (['./icepyx/c...ss_query.py']) == (['./icepyx/c...ss_query.py'])
E         At index 0 diff: ['./icepyx/core', './icepyx/quest', './icepyx/quest/dataset_scripts', './icepyx/tests', './icepyx/tests/expected_outputs', './icepyx/tests/test_data'] != ['./icepyx/core', './icepyx/quest', './icepyx/quest/dataset_scripts', './icepyx/tests']
E         Use -v to get more diff

icepyx/tests/test_read.py:96: AssertionError

Summary

If we land on a standard way of running local tests I'd be happy to add a page to the Contributor docs for how to set it up!