Add ability to name scales

[eliocamp]

From #48

A possible fix is to change the function to: new_scale("alpha", "some.name") instead of only adding the _new (or leaving it as the default behavior) and then using that name in the guides will probably work

[g-pacheco]

Hello @eliocamp,

I am sorry but I am not sure if I could not understand well what you suggested. Was it something like this? With this command I get the following error:

Error in new_scale("alpha", "Layka") : unused argument ("Layka")

  ggplot() +
  geom_raster(data = Salinity_df_r, aes(x = x, y = y, fill = Present.Surface.Salinity.Mean)) +
  scale_fill_viridis_c(option = "G", direction = -1) +
  geom_sf(data = Global, fill = "#fff7f3", colour = "#000000") +
  new_scale("alpha", "Layka") +
  geom_star(data = Coors_Grouped, aes(x = Longitude, y = Latitude, starshape = Group, fill = BioState),
            size = 6) +
  coord_sf(xlim = c(-11, 31), ylim = c(67, 49.75), expand = FALSE) +
  scale_starshape_manual(values = Shapes) +
  scale_x_continuous(breaks = seq(-10, 32, by = 10)) +
  scale_y_continuous(breaks = c(50, 58, 66)) +
  annotation_north_arrow(location = "br", which_north = "false", style = north_arrow_fancy_orienteering,
                         height = unit(2.5, "cm"), width = unit(3, "cm"),
                         pad_x = unit(2.75, "in"), pad_y = unit(.4, "in")) +
  annotation_scale(location = "br", line_width = 2, text_cex = 1.35, style = "ticks", pad_y = unit(.3, "in")) +
  theme(panel.background = element_blank(), 
        panel.border = element_rect(colour = "black", size = .5, fill = NA),
        panel.grid.major = element_line(color = "#000000", linetype = "dashed", size = .00005),
        plot.margin = margin(t = .005, b = .005, r = .2, l = .2, unit = "cm"),
        legend.background = element_rect(size = .15, color = "#000000", fill = "#fff7f3", linetype = "dotted"),
        legend.key = element_blank(),
        legend.position = c(.142, .8675),
        axis.title = element_blank(),
        axis.text = element_text(colour = "#000000", size = 15, face = "bold"),
        axis.ticks = element_line(color = "black", size = .5)) +
  guides(fill = guide_colourbar(title = "Salinity", title.theme = element_text(size = 16, face = "bold"),
                                label.theme = element_text(size = 14, face = "bold"), label.position = "top",
                                barwidth = 12.3, barheight = 1, order = 1, frame.linetype = 1, frame.colour = "#000000",
                                ticks.colour = "#ffffff", direction = "horizontal", reverse = TRUE, even.steps = TRUE,
                                draw.ulim = TRUE, draw.llim = TRUE),
         starshape = guide_legend(title = "Sampling Localities", title.theme = element_text(size = 16, face = "bold"),
                                  label.theme = element_text(size = 14),
                                  override.aes = list(starshape = Shapes, size = 4), order = 2))

Thanks in advance, George.

[eliocamp]

Yes, this is an idea that's not implemented.

In fact, after I opened the issue I realised that is not as easy to do because new_scale() renames the previous entries.

[g-pacheco]

I hope you will manage to find a solution. It would be great to have this option.

[eliocamp]

In the meantime, take a look at relayer, which already implements this but with a different syntax.

https://github.com/clauswilke/relayer

[rsh52]

@eliocamp Great work on this package, it's helping out a lot with a plotting package I'm developing that requires separation of layers.

Wanted to check in on this and see if there's any plans to address, it would be very helpful for our workflow to have a renaming function. I realize how tricky this is in the context of ggplot's layering system, unfortunately we won't be using relayer since it isn't actively maintained.

[eliocamp]

This is not easy/possible right now due to how ggnewscale works. It might be possible after a future rewrite which will be possible once this ggplot2 issue is resolved via this PR.

[rsh52]

Understood, I'll sub to both of those. Hopefully this is in the cards, would be incredibly useful and helpful! Thank you.