Fix Kuiper test

[Azhag]

Heya!

Was trying to get Kuiper 2 sample test going, and saw a few mistakes which I fixed:

• Setup.py wasn't installing kuiper_table.npy
• _kuiper_lookup wasn't converted from Matlab properly, you had an off by one error that triggered for N>501.

Thanks for the package, really useful!

cheers Loic

[fabiansinz]

Hey Loic,

thanks for contributing. We haven't had a pull request in a while and we need to update our .travis.yml for automated testing. Once that is fixed, I will ask you to update your pull request. Thanks!

Fabian

[fabiansinz]

Can you merge the latest updates and update you pull request? We fixed the travis file.

[Azhag]

Done! But I didn't check the unit tests and they fail now, I'll have a look and see if I introduced a bug.

[fabiansinz]

That'd be great. Thank you.

[coveralls]

Coverage remained the same at 90.341% when pulling b33d7b10b0497a23a8107474b4385baaeae9cbe1 on Azhag:master into 41878f98a8a403334e02cf8a7bbdee8a55375859 on circstat:master.

[coveralls]

Coverage remained the same at 90.341% when pulling b33d7b10b0497a23a8107474b4385baaeae9cbe1 on Azhag:master into 41878f98a8a403334e02cf8a7bbdee8a55375859 on circstat:master.

[Azhag]

Ok fixed it, actually your test was wrong, which my commit fixed :P

Made a correct test now, with an N that really isn't in the Kuiper Table.

Added a test to make sure that results with angles in [0, 2pi] work in [-pi, pi], as I wasn't sure if the library supported it :)

[coveralls]

Coverage remained the same at 90.341% when pulling a6b8138ee43fd10c0cf993b34e88c2b31be44855 on Azhag:master into 41878f98a8a403334e02cf8a7bbdee8a55375859 on circstat:master.

[fabiansinz]

Thank you!