[BUG] geom_swim_label() Introduces Scale Imbalance

[rsh52]

Expected Behavior

In a perfect world, the order of geom_swim_label() / geom_swim_point() wouldn't matter and the legend would render either way they are called.

Current Behavior

Due to geom_swim_label()s use of guides() to force emojis where glyphs typically reside, ordering does matter. For the time being, this is low priority and may not be addressed with the current setup of ggplot and ggnewscale.

Here is an example of two working ggswim setups:

# swim point call before label call
patient_data |>
  ggplot() +
  geom_swim_point(data = infusion_events,
                  mapping = aes(x = time_from_initial_infusion, y = pt_id, colour = infusion_type)) +
  geom_swim_label(data = end_study_events,
                  mapping = aes(x = time_from_initial_infusion, y = pt_id,
                                label_vals = end_study_label, label_names = end_study_name))

# swim label before point with new_scale_color()
patient_data |>
  ggplot() +
  geom_swim_label(data = end_study_events,
                  mapping = aes(x = time_from_initial_infusion, y = pt_id,
                                label_vals = end_study_label, label_names = end_study_name),
                  label.size = NA, fill = NA, size = 5) +
  new_scale_color() +
  geom_swim_point(data = infusion_events,
                  mapping = aes(x = time_from_initial_infusion, y = pt_id, colour = infusion_type))

How to Reproduce the Bug:

And here is one that fails, where geom_swim_label() is called before geom_swim_point() and without new_scale_color():

patient_data |>
  ggplot() +
  geom_swim_label(data = end_study_events,
                  mapping = aes(x = time_from_initial_infusion, y = pt_id,
                                label_vals = end_study_label, label_names = end_study_name),
                  label.size = NA, fill = NA, size = 5) +
  geom_swim_point(data = infusion_events,
                  mapping = aes(x = time_from_initial_infusion, y = pt_id, colour = infusion_type))

Current Solution

In commit ec13553 we reintroduced try_ggswim() which will detect this type of error and return it with slightly more context and suggestion of how to fix it.

Failure Logs

The typical error without try_ggswim():

Error:
✖ `[[<-.data.frame`(`*tmp*`, i, value = list(end_study_label = c("✅", "❌", "⚠️"))) replacement has 3 rows, data has 5

And the updated error with try_ggswim():

Error:
✖ `[[<-.data.frame`(`*tmp*`, i, value = list(end_study_label = c("✅", "❌", "⚠️"))) replacement has 3 rows, data has 5
! Scale misalignment detected when rendering ggswim.
ℹ This can often be resolved by calling `new_scale_color()` or reordering point and label geoms.

Update 2024-09-03: This only seems to work for certain cases. It does not always capture or trigger try_ggswim(). Fixed in recent commit.

Checklist

Before submitting this issue, please check and verify below that the submission meets the below criteria:

• The issue is atomic
• The issue description is documented
• The issue title describes the problem succinctly
• Developers are assigned to the issue
• Labels are assigned to the issue