Return e raised to the power x, minus 1. Here e is the base of natural logarithms. For small floats x, the subtraction in exp(x) - 1 can result in a significant loss of precision; the expm1() function provides a way to compute this quantity to full precision:
math.expm1(x)>>> from math import exp, expm1
>>> exp(1e-5) - 1 gives result accurate to 11 places
1.0000050000069649e-05
>>> expm1(1e-5) result accurate to full precision
1.0000050000166668e-05