Add p-values/significance to ALL boxplots (pairwise) in facet_wrapped plot

Question

I have a facet_wrapped plot with 5 facets, that have different numbers of boxplots in them.

I wanna pairwise compare all of them if they differ siginficantly from each other, ideally already giving the p-value IN the graph.

My code:

data.csv %>%
  group_by(method, year) %>%
  ggplot(aes(crop, H, fill = crop)) +
  geom_boxplot() +
  labs(x = "", y = "Shannon Index per field") +
  scale_fill_manual(
    values = c(
      "antiquewhite4", "burlywood4", "cornsilk2",
      "cornsilk", "antiquewhite3", "antiquewhite2"
    )
  ) +
  facet_wrap(~method, scales = "free_x") +
  stat_compare_means(aes(group = crop), label = "p.signif")

What I tried:

• stat_compare_means -> doesn't seem to work, gives the wrong values (I think it compares ALL barleys with ALL whetas and not only the ones IN one of the facets).

• stat_pvalue_manual(stat.test) -> gives me errors, doesn't find object stat.test no matter how I put the object before

What I need: for it to work without making every facet a single boxplot graph and then adding all together later thanks a lot for your help!!!