right way to seed simulate() for reproducible results

[dionman]

From a bash script, I'm calling twice a python script where following line gets identically instantiated

hawkes_exp_kernels = SimuHawkesExpKernels(adjacency=adjacency, decays=1.,
                                          baseline=baseline, end_time=T,
                                          verbose=False, seed=np.random.seed(ID))

I end up getting simulations of (similar but) different length. What's the right way to call hawkes_exp_kernels.simulate() with fixed seed in order to get reproducible simulations?

[PhilipDeegan]

I think the seed needs to be a constant.

try seed=1337

[dionman]

thanks, this works