Oh, yay, tests are running on pull requests now! One of the things that I included in #214 was an update to our CI/CD pipeline.

Previously, if I created a branch in demiangomez/Parallel.GAMIT, and requested a PR against master, tests would run (as expected). However, if I cloned Parallel.GAMIT (i.e., into espg/Parallel.GAMIT) and then opened a pull request against the demiangomez/Parallel.GAMIT master branch, the unit tests wouldn't run-- this is why we didn't see the failing unit test in #214 until after you approved and rebased-merged.

Now, with us calling our test workflow on any pull_request_target actions, we'll run tests on PR when we (and others) use proper git hygiene by not working in master and cloning the repo...

@demiangomez this can get merged. The reason this is failing is because the test suite is running the current code inside of the pgamit master branch:

def test_max_clust_expansion(min_clust, max_clust, neighbors, overlap):
        """Test algorithmic guarantee of `over_cluster`
    
        Verify that expanded cluster size is under (<=, less than or equal to):
        [initial cluster size + (neighbors * overlap)]"""
    
        data = gen_variable_density_clusters()
        clust = BisectingQMeans(min_size=min_clust, opt_size=max_clust,
                                init='random', n_init=50, algorithm='lloyd',
                                max_iter=8000, random_state=[42]
        clust.fit(data)
    
        OC = over_cluster(clust.labels_, data, metric='euclidean',
                          neighbors=neighbors, overlap_points=overlap)

while this PR is modifying the unit test that is failing:

    OC = over_cluster(clust.labels_, data, metric='euclidean', 
                      neighbors=neighbors, overlap_points=overlap,
                      method='dynamic')

The pull_request_target workflow uses the current master branch of demiangomez/Parallel.GAMIT for running tests-- the reason is because if it used the PR codebase for tests, then the PR submitter could change the test code to do something annoying/nefarious (i.e., a bot could submit PR's that would modify the test actions to send spam email or otherwise hijack the git runner image).

We'll see real-time test results for PR's that don't modify the test suites... and for anything that does add a new test, we'll see it delayed. That said, the 'on push' workflow will still show it correctly as passing inside of the cloned repo where the PR is coming from (i.e., this PR does pass tests in my repo since my repo runs tests using code from itself (espg/Parallel.GAMIT), rather than from the upstream demiangomez/Parallel.GAMIT master branch).