In the arviz docs, there is this example using `from_pymc3`

:

```
trace = pm.sample(draws, chains=chains)
prior = pm.sample_prior_predictive()
posterior_predictive = pm.sample_posterior_predictive(trace)
pm_data = az.from_pymc3(
trace=trace,
prior=prior,
posterior_predictive=posterior_predictive,
coords={"school": np.arange(eight_school_data["J"])},
dims={"theta": ["school"], "theta_tilde": ["school"]},
)
```

so pm_data is an InferenceData that is aware of the trace, prior, and posterior_predictive.

But soon `pm.sample`

will return an InferenceData by default. How do I make that InferenceData aware of the prior and posterior_predictive?