How do I set a different prior distribution for each Numpyro Component

Question

I'm building mixture model in Numpyro, and would like to set each component with a different prior. So far I have this where n == the number of components in the mixture.

def model(data, n):
    weights = numpyro.sample('weights', dist.Dirichlet(concentration=jnp.ones(n)))
    with numpyro.plate('components', n):
        alpha = numpyro.sample('alpha', dist.LogNormal(loc=a, scale=b))
        beta = numpyro.sample('beta', dist.LogNormal(loc=c, scale=d))

    with numpyro.plate('data', len(data)):
        assignment = numpyro.sample('assignment', dist.Categorical(weights))
        numpyro.sample('obs', dist.Beta(alpha[assignment], beta[assignment]), obs=data)

However what if for each component n, I'd like to sample the alpha/beta parameters from a different LogNormal distribution (unique a,b,c,d parameters here)? I've tried creating an array of distributions and passing that to 'obs' but a shape related issue was returned (data shape versus component shape). I looked into numpyro plate_stack however that doesn't seem to fit what I need.