Cleanup and fix failing tests

[thinkjrs]

Some of our tests have been failing for a while and there were a few deprecations upstream since I've worked on this client.

Update deprecated tests

• [x] ig audience analysis
• [x] playlist evolution stats

Fix failing tests

Or comment & explain failing tests & remove functionality

  1 tests/test_album.py::test_charts FAILED                                  [  1%]
  2 tests/test_artist.py::test_charts FAILED                                 [ 13%]
  3 tests/test_charts.py::test_amazon_tracks FAILED                          [ 31%]
  4 tests/test_charts.py::test_amazon_albums FAILED                          [ 32%]
  5 tests/test_charts.py::test_cm_score_tracks FAILED                        [ 39%]
  6 tests/test_charts.py::test_cm_score_artists FAILED                       [ 40%]
  7 tests/test_charts.py::test_cm_score_albums FAILED                        [ 42%]
  8 tests/test_charts.py::test_youtube_artists FAILED                        [ 59%]
  9 tests/test_playlist.py::test_evolution FAILED                            [ 80%]
 10 tests/test_playlist.py::test_lists FAILED                                [ 81%]
 11 tests/test_track.py::test_charts FAILED                                  [ 88%]
 12 tests/test_track.py::test_get_track_ids FAILED                           [ 89%]

Cleanup tests

Our tests need a number of new things:

• [x] better test-wide confest.py w/reusable fixtures
• [ ] class-based organization for some tests to minimize boilerplate and upgrade readability

In addition, our current tests need updating: ~- [ ] add modern type hints~

• [ ] more cohesive testing of api combinations
• [x] removal of all magic numbers

[thinkjrs]

test_album.py::test_charts

Working! And..not working! What?

Something's happening to auth as we're getting 403's when using certain dates. However, if sticking to exactly those combinations used in the documentation tests will pass.

I recommend putting a logging.warning in .artist.charts, for now.

What was changed?

The "until" date--we used 2020 in the failing and 2019 in the passing.

The error

Passing!

[thinkjrs]

Added a slowdown and code coverage in commit c6befd10f and de6381bc467 .

[thinkjrs]

Better testing combos

Commentary for more cohesive testing of api combinations, above.

Map key:value pairs

Essentially, we need to grab the possible keys from the chartmetric documentation for each endpoint and then for each key list their possible values.

Randomize the ids used

We should also have a "randomizer" for ids used for testing, controlled from a single place. As in:

pytest.fixture(scope='module')
def getTestData():
    return dict(
         test_artists =  ['artist1cmid', 'artist2cmid'],
        test_tracks = [...],
        test_somethingWithAnID = [...],
    )

pytest.fixture(scope='module')
def data(getTestData):
    return dict(
        test_artist=random.choice(getTestData['test_artists']),
        ...,
    )

[thinkjrs]

Also, coverage:

[thinkjrs]

Okay, so the coverage above included our test coverage for tests, as I forgot to exclude the tests/ dir in the configuration: coverage

Much, much better, and closer to what's expected.