Fix q column in tidy_ distribution functions

[spsanderson]

Currently the "quantile" plot will produce the same line for all simulations of a given tidy_ distribution because q is calculated incorrectly. It uses a single simulation instead of a zero one scaling of the current y value from the random variate function.

Example:

tidy_poisson(.num_sims = 3) %>%
  tidy_autoplot(.plot_type = "quantile")

Currently for at least tidy_poisson() the q value is calculated as such:

dplyr::mutate(q = list(stats::qpois(qs, lambda = lambda)))

It should be:

dplyr::mutate(q = list(stats::qpois(tidy_scale_zero_one_vec(unlist(y)), lambda = lambda)))

Which would give the following: