compute_pvals passing invalid value to log_GP_pmf

[ypinto31]

When computing p-values, get an error saying an invalid value is being passed to sp.special.gammaln(x + 1) in the log_GP_pmf() function in the genpois module. Tried troubleshooting and found that rows from input file are being processed in a random order and the function fails at different rows each time. x, the variable that is being passed to the function is of type: numpy.int64.

RuntimeWarning: invalid value encountered in log

• sp.special.gammaln(x + 1)

[laserson]

Unfortunately I don't have the bandwidth to debug this. If you are so inclined I'd review a PR if it's small. That said, I'm assuming that something like log(0) is being evaluated and that probably gets propagated as NaNs, in which case this may not be problematic?