RNG is not seeded

[mfisherlevine]

np.random.seed() is never called, nor is an RNG instantiated and passed around, leading to non-deterministic results for the same input.

e.g. from d568533 run python runExtractor.py --verbose a few times and see both very different run times and different RuntimeWarnings (especially in the Start ChromaticPSF1D polynomial fit... and Calibrating order 1 spectrum... stages), as well as different final results sometimes.

[mfisherlevine]

Places with random used:

spectractor/simulation/image_simulation.py:261: noisy += np.random.normal(scale=self.read_out_noise)
spectractor/extractor/psf.py:1241: bgd_std = float(np.std(np.random.poisson(bgd)))
spectractor/extractor/psf.py:1448: bgd_std = float(np.std(np.random.poisson(bgd)))
spectractor/fit/fitter.py:68: [np.random.uniform(self.p[i] - 0.02 * self.p[i], self.p[i] + 0.02 * self.p[i], self.nwalkers)
spectractor/fit/fitter.py:70: self.start[self.start == 0] = 1e-5 * np.random.uniform(0, 1)