Labeled coords and dims in hierarchical group setting with different sizes

Questions v5
12

Ah, I failed to adapt the per-participant offset. Try this?

    adaptation_offset = pm.Normal('a_offset', 
                                  mu=0, sigma=1,  
                                  dims=("participants"))
    adaptation = pm.Deterministic("adaptation", 
                                  mu_a[exp_idx] + (adaptation_offset[participant_idx] * sigma_a[exp_idx])
                                  )

So now there is a mu_a per experiment, a sigma_a per experiment, and an offset per participant. Each gets selected according to the appropriate index. I think.