scipy.stats.nbinom.logpmf(0,1,1) returns nan. Correct value is 0

[pbrod]

One solution is to replace _logpmf with:

def _logpmf(self, x, n, p):
    coeff = gamln(n + x) - gamln(x + 1) - gamln(n)
    return coeff + n*log(p) + special.xlog1py(x, -p)

Another issue is that according to wikipedia it is only defined for n>0, but here it is defined for n>=0.
Perhaps consider replacing _argcheck with

def _argcheck(self, n, p):
        return (n > 0) & (p >= 0) & (p <= 1)

[rgommers]

Fixed by gh-4049.