I would like to use az.plot_dist_comparison() to visually compare the prior and the posterior of a variable. az.plot_dist_comparison() uses the posterior samples and prior samples from an az.InferenceData object, passed as an argument. I have an az.InferenceData object, with posterior samples, created via a call to pm.sample(). And I have prior samples, created via a call to pm.sample_prior_predictive().
But the prior samples are not within the az.InferenceData object. Instead pm.sample_prior_predictive() returns the samples as a dictionary with variable names as keys and numpy arrays as values. How do I take that dict and add it as the prior on the az.InferenceData object? Or maybe there is a simpler path, some way to tell pm.sample_prior_predictive() to add the prior samples to an existing az.InferenceData object?