Questions/problems:
-[ ] Is this the right place in the code for this to live?
-[ ] Why doesn't this show up automatically in the GUI dropdown list of colormaps? Is there another place I need to update code?
-[ ] Why does the shade of blue or red seem to vary when applied to float data? (Maybe we leave this question for a future PR?)

Minimal reqproducible example:

from napari.utils.colormaps.colormap_utils import vispy_or_mpl_colormap

hilo_colormap = vispy_or_mpl_colormap('gray')
hilo_colormap.colors[0] = [0, 0, 1, 1]   # blue (lowest pixel values)
hilo_colormap.colors[-1] = [1, 0, 0, 1]  # red  (highest pixel values)

import napari
import skimage.data
viewer = napari.Viewer()
camera = skimage.data.camera()
viewer.add_image(camera, colormap=hilo_colormap)
# or if you have this PR branch checked out, you can do...
# viewer.add_image(camera, colormap='hilo')

Screenshot:

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 92.43%. Comparing base (f195690) to head (a9fbe3b).

@@            Coverage Diff             @@
##             main    #6906      +/-   ##
==========================================
- Coverage   92.45%   92.43%   -0.02%     
==========================================
  Files         614      614              
  Lines       55164    55169       +5     
==========================================
- Hits        51001    50998       -3     
- Misses       4163     4171       +8     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

Closing, in favor of the approach suggested by melonora:

I checked this and this does not seem to be correct. We had the same approach, but due to conversion to uint8 in vispy for a 16 bit image for example the intensity values are binned. This means that if you have intensity values of both 0 and 1 in your 16 bit image, both will be blue, which should not be the case.

A shader is required

See #6878 (comment)