Allow replacing h parameter with CMB BAO scale theta_s/theta_MC

[pmcdonal]

Planck constrains it so well that you mostly live on this subspace and it is useful to parameterize that explicitly... (of course, maybe this is already possible, in which case I'd suggest putting it in fisher_planck.ipynb as an example...)

[adematti]

this is unfortunately not yet possible... this requests modifying cosmoprimo. A dirty fix would be to pass theta_cosmomc to camb directly, but that would work only for camb (resp. 100 * theta_star for class). The best solution would be to adapt camb's theta_cosmomc method to Python, and solve for h, given theta_cosmomc, Omega_b (or omega_b), and Omega_cdm (or omega_cdm). (Happy to take contributions on this :))