Reconstruct marginalised parameters

Closes #324 (closed)

This extends the previous functions to reconstruct the distance posterior when using a marginalised likelihood to allow reconstruction of all marginalised parameters.

Distance and phase work very nicely.

The time reconstruction is a little problematic, I think this is because the times at which the likelihood is sampled are too coarse and for each sample, there is only one time with any support. There is also some buggy behaviour when reconstructing distance and time, this may be due to the fact that the time marginalised likelihood does not incorporate the change in the response function between the beginning of the segment and the sample time.

bilby/core/sampler/__init__.py

@@ -182,7 +182,7 @@ def run_sampler(likelihood, priors=None, label='label', outdir='outdir',
             result.injection_parameters = conversion_function(
                 result.injection_parameters)
 
-    result.samples_to_posterior(likelihood=likelihood, priors=priors,
+    result.samples_to_posterior(likelihood=likelihood, priors=result.priors,
                                 conversion_function=conversion_function)
     if save == 'hdf5':
         result.save_to_file(extension='hdf5')