Saving layers as svg fails with text labels and color images

[Micro-Sandworms]

🐛 Bug Report

When saving layers as svg, the program crashes if one of the layers contains a color image. In addition, text labels are not saved.
The former can be solved by changing the code at the layer_to_xml.py, function
def extrema_image(image, meta):
"""Compute the extrema of an image layer, accounting for transforms."""
coords = np.array([[0, 0], list(image.shape)]) # This does not work for color images
#coords = np.array([[0, 0], list(image.shape[:2])]) # This works
return extrema_coords(coords, meta)

The later, I don't see how to solve, and I appreciate your help in advance.

💡 Steps to Reproduce

import napari
import numpy as np

# Create a viewer
viewer = napari.Viewer()

# Create a color noise image
noise_image = np.random.random((100, 51, 3))
# noise_image = np.random.random((100, 51))  # this one works to save the image as is not a color image

# Add the noise image layer first
viewer.add_image(noise_image, name='noise_image')

# Add a layer with a single point
points = np.array([[100, 200]])
viewer.add_points(points, size=10, face_color='red', name='points_layer')

# Add a text layer
text = ['Sample Text']
viewer.add_points(points, size=0, text=text, name='text_layer')

# Set the zoom level to 0.3
viewer.camera.zoom = 0.7

# Start the Napari event loop
napari.run()

💡 Expected Behavior

No response

🌎 Environment

napari: 0.5.5
Platform: Windows-10-10.0.19045-SP0
Python: 3.11.11 | packaged by conda-forge | (main, Dec 5 2024, 14:06:23) [MSC v.1942 64 bit (AMD64)]
Qt: 5.15.2
PyQt5: 5.15.11
NumPy: 1.24.0
SciPy: 1.15.0
Dask: 2024.12.1
VisPy: 0.14.3
magicgui: 0.10.0
superqt: 0.7.1
in-n-out: 0.2.1
app-model: 0.3.1
psygnal: 0.11.1
npe2: 0.7.7
pydantic: 2.10.4

OpenGL:

• GL version: 4.6.0 - Build 30.0.101.1692
• MAX_TEXTURE_SIZE: 16384
• GL_MAX_3D_TEXTURE_SIZE: 2048

Screens:

• screen 1: resolution 1920x1200, scale 1.0
• screen 2: resolution 1680x1050, scale 1.0
• screen 3: resolution 1707x960, scale 1.0

Optional:

• numba: 0.60.0
• triangle: 20250106
• napari-plugin-manager: 0.1.4

Settings path:

• C:\Users\alvarez\AppData\Local\napari\Visualization_Gene_Expression_d453efcbacc35064a38aa632634bb2eff004ecff\settings.yaml
  Plugins:
• napari: 0.5.5 (81 contributions)
• napari-console: 0.1.3 (0 contributions)
• napari-svg: 0.2.0 (2 contributions)

💡 Additional Context

No response

[jni]

Thanks for the issue @Micro-Sandworms! I've transferred it to napari-svg which is its own repo.

Are you able to contribute a fix for the image issue? That would be much appreciated!

I don't know all that much about svg but I imagine it's not too hard to add text, so it would be a matter of (a) looking for any text annotations on the layer, and (b) saving those in the appropriate part of the xml.

[Micro-Sandworms]

Thank you, Juan. As I mentioned, replacing the function 'extrema_image' in 'layer_to_xml.py' with

def extrema_image(image, meta):
    """Compute the extrema of an image layer, accounting for transforms."""
    coords = np.array([[0, 0], list(image.shape[:2])])
    return extrema_coords(coords, meta)

It worked for me, but I am not in a position to foresee possible negative consequences down the line. Still, I made a commitment with this suggestion.
Regarding the text labels, I have no clue how to solve the issue. I'm sorry.

[jni]

@Micro-Sandworms oh thanks, I more meant, would you like to make a pull request to the repo? Here's a good guide for how to do that, if you are not familiar with the process: https://www.dataschool.io/how-to-contribute-on-github/

[Micro-Sandworms]

I made two pull requests. One is for a possible solution regarding the color images, and the other is for the text labels.