Tracks id not displaying

🐛 Bug

There is an unanswered question on the image.sc forum about track IDs not displaying. Indeed, when using the simple example from the docs, the IDs cannot be displayed and an error appears in the terminal which seems related to something deeper in vispy.

To Reproduce

1. Run the demo code:

import napari
from skimage import data

hubble_image = data.hubble_deep_field()

tracks_data = [
    [1, 0, 236, 0],
    [1, 1, 236, 100],
    [1, 2, 236, 200],
    [1, 3, 236, 500],
    [1, 4, 236, 1000],
    [2, 0, 436, 0],
    [2, 1, 436, 100],
    [2, 2, 436, 200],
    [2, 3, 436, 500],
    [2, 4, 436, 1000],
    [3, 0, 636, 0],
    [3, 1, 636, 100],
    [3, 2, 636, 200],
    [3, 3, 636, 500],
    [3, 4, 636, 1000]
]

viewer = napari.view_image(hubble_image, name='image')
viewer.add_tracks(tracks_data, name='tracks')

1. Try to activate the track IDs through the tick box. No IDs are displayed.

This is the error trace that appears in the terminal (but doesn’t make napari crash):

Traceback (most recent call last):
  File "/Users/gw18g940/miniconda3/envs/napari-tutorial/lib/python3.9/site-packages/vispy/app/backends/_qt.py", line 897, in paintGL
    self._vispy_canvas.events.draw(region=None)
  File "/Users/gw18g940/miniconda3/envs/napari-tutorial/lib/python3.9/site-packages/vispy/util/event.py", line 453, in __call__
    self._invoke_callback(cb, event)
  File "/Users/gw18g940/miniconda3/envs/napari-tutorial/lib/python3.9/site-packages/vispy/util/event.py", line 471, in _invoke_callback
    _handle_exception(self.ignore_callback_errors,
  File "/Users/gw18g940/miniconda3/envs/napari-tutorial/lib/python3.9/site-packages/vispy/util/event.py", line 469, in _invoke_callback
    cb(event)
  File "/Users/gw18g940/miniconda3/envs/napari-tutorial/lib/python3.9/site-packages/vispy/scene/canvas.py", line 218, in on_draw
    self._draw_scene()
  File "/Users/gw18g940/miniconda3/envs/napari-tutorial/lib/python3.9/site-packages/vispy/scene/canvas.py", line 277, in _draw_scene
    self.draw_visual(self.scene)
  File "/Users/gw18g940/miniconda3/envs/napari-tutorial/lib/python3.9/site-packages/vispy/scene/canvas.py", line 315, in draw_visual
    node.draw()
  File "/Users/gw18g940/miniconda3/envs/napari-tutorial/lib/python3.9/site-packages/vispy/scene/visuals.py", line 99, in draw
    self._visual_superclass.draw(self)
  File "/Users/gw18g940/miniconda3/envs/napari-tutorial/lib/python3.9/site-packages/vispy/visuals/visual.py", line 605, in draw
    v.draw()
  File "/Users/gw18g940/miniconda3/envs/napari-tutorial/lib/python3.9/site-packages/vispy/scene/visuals.py", line 99, in draw
    self._visual_superclass.draw(self)
  File "/Users/gw18g940/miniconda3/envs/napari-tutorial/lib/python3.9/site-packages/vispy/visuals/visual.py", line 451, in draw
    self._program.draw(self._vshare.draw_mode,
  File "/Users/gw18g940/miniconda3/envs/napari-tutorial/lib/python3.9/site-packages/vispy/visuals/shaders/program.py", line 102, in draw
    Program.draw(self, *args, **kwargs)
  File "/Users/gw18g940/miniconda3/envs/napari-tutorial/lib/python3.9/site-packages/vispy/gloo/program.py", line 498, in draw
    raise RuntimeError('All attributes must have the same size, got:\n'
RuntimeError: All attributes must have the same size, got:
<VertexBuffer size=15 last_dim=1>: 15
<VertexBuffer size=48 last_dim=4>: 48
<DataBufferView on <VertexBuffer size=48 last_dim=None> at offset=0 size=48>: 48
<DataBufferView on <VertexBuffer size=48 last_dim=None> at offset=8 size=48>: 48
<VertexBuffer size=48 last_dim=1>: 48
<VertexBuffer size=48 last_dim=3>: 48

Expected behavior

There’s little documentation about the show ID feature (it’s never used in the documentation) but from discussions in the PR implementing the layer, it looks like it should display the IDs.

Environment

napari: 0.4.12
Platform: macOS-10.15.7-x86_64-i386-64bit
System: MacOS 10.15.7
Python: 3.9.7 (default, Sep 16 2021, 08:50:36) [Clang 10.0.0 ]
Qt: 5.15.2
PyQt5: 5.15.6
NumPy: 1.20.3
SciPy: 1.7.2
Dask: 2021.11.1
VisPy: 0.9.3

OpenGL:
- GL version: 2.1 INTEL-14.7.20
- MAX_TEXTURE_SIZE: 16384

Screens:
- screen 1: resolution 1440x900, scale 2.0

Plugins:
- cellpose-napari: 0.1.4
- console: 0.0.4
- scikit-image: 0.4.12
- svg: 0.1.5

Additional context

[brisvag]

Nope, doesn’t work, you’re right. I reverted to using two instances.

[quantumjot]

I just had chance to try git-bisect on this, using examples/tracks_3d_with_graph.py. The bad commit is:

5f1525c394606bfc5193a837f55f1c38cdcbeac7 is the first bad commit
commit 5f1525c394606bfc5193a837f55f1c38cdcbeac7
Author: Lorenzo Gaifas <[e-mail]>
Date:   Tue Sep 28 22:24:47 2021 +0200

    `_vispy` module cleanup (#3333)

so it looks like something got broken during the vispy module cleanup.