cedricsimar

Similarly to @gvanderheide I've used successfully a workaround based on np.random.choice code published on https://github.com/numpy/numpy/blob/76a76c78d1b049126153e81b0a9d137fa3e4947b/numpy/random/mtrand/mtrand.pyx#L1194: Given the probability distribution `probability` and the sampling size `size`, the code is `cumulative_distribution =...