Mixture of normal and AR(p) process?

Questions
3

I get the same error when I specify the shape: here is what I tried (maybe I’m not doing it correctly):

    with pm.Model() as self.model:

        W = np.array([1., 1.])
        w = pm.Dirichlet('w', W)

        intercept = pm.Normal('intercept', mu=-5, sd=5., testval=-5.)
        rhos = pm.Uniform('rhos', lower=-1., upper=1., shape=self.num_lags)

        comp1 = pm.AR.dist(rho=rhos, sd=pm.math.exp(intercept), shape=len(data))
        comp2 = pm.Normal.dist(mu=0., sd=0.01, shape=len(data))

        pm.Mixture('obs', w=w, comp_dists=[comp1, comp2], observed=data)

I still get the same axis 1 is out of bounds [-1, 1) error. A custom likelihood with DensityDist is the next thing I’m going to try, I think.