Closed maelle closed 3 weeks ago
For test-arpack I'd need help or pointers to similar tests in the C core / Python interface
I think the one below can be removed because there is a test with subset of attributes
I think the one below can be removed because we added a better test for the hits function in #1449 and the other tests for the to-be-deprecated function names will be removed soon enough.
TODO: Hub and authority scores make little sense for undirected graphs
If the graph is connected, you can verify that HITS and eigenvector centrality return the same result. That would be a good test.
For the one below I will simply use the code from the example
@szhorvat but for hits didn't we already improve the tests that use the hits_scores()
function? If they aren't good enough yet could you please open an issue?
For the comment below I do not know what else to test and how
I wonder what the expectation is supposed to be in:
@szhorvat I am done looking for the TODOs in the test files, and for many of them I'll need some guidance, but there is no hurry.
For the comment below I do not know what else to test and how
I suppose this refers to the fact that the test does not check that they are maximal or maximum. I don't have a quick enough solution for this at the moment. I would leave this as-is for the moment.
How:
Same comment as before. We can check that they are all the same size, but I don't think it's even feasible to check that they are largest/maximum. Checking that they are maximal should be feasible, but still some work. I'd leave it for now.
also how:
You can check that a separator is minimal by:
is_separator
)EDIT: I think we can use is_min_separator()
here directly, which is simpler.
I wonder what the expectation is supposed to be in:
No clue what this is. I suggest removal.
One comment per test file, I'll ask for help once the list is done.