Store analytic bilby priors when `--disable_prior_sampling` is passed

If I run the following command line,

$ summarypages --webdir ./webpage --samples bilby.json --labels prior_sampling --multi_process 4 --gw --nsamples_for_prior 1000 --disable_expert

The analytic bilby priors are stored in the pesummary metafile:

>>> import h5py
>>> f = h5py.File("webpage/samples/posterior_samples.h5", "r")
>>> f["prior_sampling"]["priors"]["analytic"]
<HDF5 group "/one/priors/analytic" (108 members)>
>>> f.close()

However, if the --disable_prior_sampling is provided,

$ summarypages --webdir ./webpage --samples bilby.json --labels no_prior --multi_process 4 --gw --disable_prior_sampling --disable_expert

The analytic bilby priors are not stored in the pesummary metafile:

>>> import h5py
>>> f = h5py.File("webpage/samples/posterior_samples.h5", "r")
>>> f["no_prior"]["priors"]["analytic"]
...
KeyError: "Unable to open object (object 'analytic' doesn't exist)"
>>> f.close()

This seems like a design flaw. I think the analytic priors should always be saved if they are available in the bilby json file.