painyeph / fishersexacttest Goto Github PK

Dear painyeph

I found that for very few contingency tables, your implementation fails:

a, b, c, d = (2, 3, 0, 2)
fisher.test1t(a, b, c, d)  #0.5238095238095246
scipy.stats.fisher_exact([[a, b], [c, d]], 'two-sided')[1]  # 1.0

Hey painyeph

Thanks a lot for your great work!

I lazily wanted to see how much faster your tool would be using numba / Cython. The result is quite impressive, see my repo!

https://github.com/MrTomRod/fast-fisher

f_python is your version, f_compiled is mine.

I added the MIT licence. Hope that's ok for you.

The only downside is that numba is added as a dependency.

The newest version is a Cython module with almost the same performance as numba. The only dependency is now that python-devel must be installed.

Shall I make a pull request?

Best,
MrTomRod