Source Edit

Stirling's formula for the gamma function gamma(x) = sqrt(2 pi) x^(x-.5) exp(-x) ( 1 + 1/x P(1/x) )

For float32: .028 < 1/x < .1 relative error < 1.9e-11

Imports

Procs

func gamma[F: SomeFloat](x: F): F: CPython math.gamma-like, with error message more detailed.
Example:

template chkValueErr(arg) = doAssertRaises ValueError: discard gamma arg chkValueErr NegInf chkValueErr 0.0 # Currently +-0.0 result is not consistent with CPython; # assert NegInf == 1.0/gamma(-180.5)
Source Edit
func gamma[T: SomeFloat](x: T; res: var T): GammaError: a more error friendly version of gamma Source Edit
func rGamma[F: SomeFloat](x: F): F {....raises: [].}: behaviors like R's gamma except for this without any warning.
Example:

from std/math import isNaN assert isNaN rgamma NegInf assert Inf == 1/rgamma(-180.5) # never returns -0.0
Source Edit
func stdlibJsGamma[F: SomeFloat](x: F): F {....raises: [].}: behaviors like @stdlib-js/gamma.js Source Edit