Fix random seed when drawing from a `target`

[aotownsend]

Currently, every time we draw(...) from a target, the parameters (redshift, t0, etc.) are randomly drawn from the functions described in the model. There is no way to fix the random seed to reproduce your results. This would be great to implement for fine-tuning parameters distributions.

There is a way to fix the random seed in the final DataSet...

DataSet.from_targets_and_survey(targets, survey, seed = seed)

but this should also be extended to the target populations.