Data type of posterior samples

[tomasstolker]

Hi @semaphoreP,

Thanks for maintaining this really helpful package!

I ran into the following error for pds70b:

line 222, in _newton_solver_wrapper
    eanom = _kepler._c_newton_solver(manom, ecc, tolerance=tolerance, max_iter=max_iter)
  File "orbitize/_kepler.pyx", line 18, in orbitize._kepler._c_newton_solver
ValueError: Buffer dtype mismatch, expected 'double_t' but got 'float'

The issue seems to be that the posterior samples that are read with get_chains have the numpy.float32 datatype while the orbitize function requires numpy.float64. For several other companions that I checked the datatype was already numpy.float64 so the error did not occur.

A simple solution would be adding dtype=np.float64 to post = np.array(hf.get('post')) in get_chains.