Besides the implementation in statsmodels as described in @Josef's answer, the E-test is also implemented in the SciPy library, with the function scipy.stats.poisson_means_test.

Usage example:

import scipy.stats as stats
count1, n1, count2, n2 = 0, 100, 3, 100
res = stats.poisson_means_test(count1, n1, count2, n2)
res.statistic, res.pvalue
(-1.7320508075688772, 0.08837900929018157)