MOD: out of range circ-median-ci

[miki998]

Hi,
Thanks for the implementation of pycircstat2, continuing on the former pycircstat.
I quickly got started and stumbled upon a small problem when instantiating a Circle (I unfortunately can't really share the bug), I suggested a very tiny patch for that feel free to check it out and see if it makes sense.

[huangziwei]

Thank you very much! Is it possible to add a test with some artificial data for this special case (to tests/test_descriptive/test_circ_median_ci())?

[huangziwei]

@miki998 Thank you very much again, and Happy New Year! I'll take a look after the New Year break.

[huangziwei]

If the data is as simple as this, we don't need to add it as a dataset, but simply put, alpha = np.arange(18) within the test_circ_median_ci(). Moreover, I tried this but it doesn't break the original code. Maybe I didn't make it clear enough before: can you provide a case of artificial data which breaks the original code?

[huangziwei]

so I think about this (a lot), and conclude that this out-of-range error is extremely unlikely to happen with real data with ci>=0.95. Until we can find a test case where this really break, I am going to close this for now.