Add `res_lnorm_hurdle()` to get residuals for log-normal hurdle model

[joethorley]

[nehill197]

Note from the brms package:
"The density of a hurdle distribution can be specified as follows. If x = 0 set f(x) = θ. Else set f(x) = (1 - θ) * g(x) / (1 - G(0)) where g(x) and G(x) are the density and distribution function of the non-hurdle part, respectively."