Allow both 'auto' and None as texture_format when choosing VisPy node

[Czaki]

Description

In #6411, we missed these lines:

napari/napari/_vispy/layers/image.py

Lines 25 to 32 in 4f4c063

	if (
	texture_format == 'auto'
	and 'texture_float' not in get_gl_extensions()
	):
	# if the GPU doesn't support float textures, texture_format auto
	# WILL fail on float dtypes
	# https://github.com/napari/napari/issues/3988
	texture_format = None

These were added in #3990 because 'auto' has a slightly different meaning in VisPy than really fully auto: it chooses the default texture dtype for the input NumPy array's dtype. For float arrays, this is float. However, if your OpenGL implementation doesn't support float textures, VisPy will raise.

Instead, passing texture_format=None when creating the ImageNode tells VisPy to transform the data to whatever texture format it sees fit.

In #6411, when getting the VisPy node for a given dtype, we only check for texture_format != 'auto' as the "catch-all" texture format. But, in fact, if we are on a machine that doesn't support float32 textures, by this point in the code the format has been changed to None, incorrectly triggering these lines:

napari/napari/_vispy/layers/image.py

Lines 59 to 73 in 89f8194

	if (
	res.texture_format != "auto"
	and dtype is not None
	and _VISPY_FORMAT_TO_DTYPE[res.texture_format] != dtype
	):
	# it is a bug to hit this error — it is here to catch bugs
	# early when we are creating the wrong nodes or
	# textures for our data
	raise ValueError(
	trans._(
	"dtype {dtype} does not match texture_format={texture_format}",
	dtype=dtype,
	texture_format=res.texture_format,
	)
	)

and causing a ValueError.

This PR fixes that by also checking for None in that same clause.

This PR also adds a test by monkeypatching the function that checks for float32 texture support.

References

#3988
#3990

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Comparison is base (4f4c063) 92.32% compared to head (d3ed995) 92.27%.
Report is 3 commits behind head on main.

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #6652      +/-   ##
==========================================
- Coverage   92.32%   92.27%   -0.05%     
==========================================
  Files         605      605              
  Lines       54227    54231       +4     
==========================================
- Hits        50064    50044      -20     
- Misses       4163     4187      +24     

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

[psobolewskiPhD]

Possibly related: https://forum.image.sc/t/opening-image-in-napari-linux/91950?u=psobolewskiphd

[Czaki]

Not possibly, but it is exactly this error.

[imagesc-bot]

This pull request has been mentioned on Image.sc Forum. There might be relevant details there:

https://forum.image.sc/t/opening-image-in-napari-linux/91950/3

[leopold-franz]

Thank you so much for adding a fix! I will try it now.

[leopold-franz]

I just tried running it (unfortunately i do not have direkt desktop access so I had to use x11 forwarding & xquartz to test this) and im running into this error now:

(follicle-tracker) labadmin@tph31-52-59:~$ napari                 
libGL error: No matching fbConfigs or visuals found              
libGL error: failed to load driver: swrast                 
WARNING: QOpenGLWidget: Failed to create context             
WARNING: composeAndFlush: QOpenGLContext creation failed                                                                                                                                
18:09:28 : WARNING : MainThread : composeAndFlush: QOpenGLContext creation faile                                                                                                                                                                                       
WARNING: composeAndFlush: makeCurrent() failed                                                                                                                                          
18:09:28 : WARNING : MainThread : composeAndFlush: makeCurrent() failed                                                                                                                 
WARNING: composeAndFlush: makeCurrent() failed                          
18:09:28 : WARNING : MainThread : composeAndFlush: makeCurrent() failed

[psobolewskiPhD]

That looks like a driver issue ☹️

Also, have you gotten napari to work via X11 and xquartz? I always get OpenGL issue and blank screen.
I've had better luck with Xpra, as our docker container uses, and web client.

[Czaki]

OpenGL does not work well with x11 screen forwarding. To use napari remote, I use nomachine/anydesk.

[imagesc-bot]

This pull request has been mentioned on Image.sc Forum. There might be relevant details there:

https://forum.image.sc/t/error-running-napari-on-hpc-could-not-load-qt-platform-plugins/92012/10