Incorporate numba with scipy.special and numba-stats

[deanm0000]

There may be some potentially significant speed improvements by running code that is compiled. At first glance, it seems that there doesn't exist a fit method in numba-stats so that might be a significant hurdle and not all the distributions are available that way so this is kind of a place holder or an aspirational enhancement.

[erdogant]

Interesting. I have been looking into it but the number of available distributions is limited. In this line, I again tried to use numba but it keeps giving the same error since forever: https://github.com/numba/numba/issues/3250

Then I looked into numba-scipy but I could not find the documentation about distributions. Maybe it is not supported.

[deanm0000]

numba-scipy opens up the scipy.special namespace so that numba is aware of them.

[erdogant]

I just created a new release where I added parallelization to the bootstrap function which significantly speeds up the results. To enable it, simply set n_jobs=-1 during initialization. However, when using multiple cores, it sometimes randomly results in a RuntimeWarning: invalid value encountered. I am not sure why this happens but if you set it back to n_jobs=1, it does not show any RuntimeWarnings.

Regarding, scipy-special, I am not sure how these functions match with scipy.stats?

[erdogant]

This issue is related to #49 which is recently solved!