moderngl on Jupyter Notebook or Google Colab

[AK82]

Hi
Is moderngl working on Jupyter Notebook or Google Colab, and if so can you please post an example?
Thank you.

[einarf]

Personally I have very little knowledge about Jupyter Notebook or Google Colab, so I woudn't be able to give a concrete answer. Maybe someone else here knows?

[stuaxo]

The easiest way to get an image to appear as Jupyter cell output is by encoding it as a PNG, and passing that data to it's Image class.

# < insert code to get png data from ModernGL here. >
display(Image(data=png_data))

You can register data types so this happens automatically.

You could register types for moderngl.Texture and whatever else it makes sense to see in the cell.

There is some work to package this all up as an extension, this would probably be a seperate project - e.g moderngl-jupyter.

VisPy have (or had?) GL rendering in Notebooks a while ago, via their Gloo library, so if you need it now that may be an option for you.

Here's a Jupyter extension I made for cairo to display cairo Surfaces and Contexts, I lean heavily on the PngFormatter so I can just pass through png data to make this work.
https://github.com/stuaxo/cairo-jupyter/blob/master/extension/cairo_jupyter/__init__.py#L46

There is info on writing Jupyter extensions here.
https://jupyter-notebook.readthedocs.io/en/stable/examples/Notebook/Distributing%20Jupyter%20Extensions%20as%20Python%20Packages.html

[einarf]

Thank you @stuaxo for answering the question. I'll close this for now. The issue still be useful in search results in the future.

[stuaxo]

I'll also point to Vispys GLIR spec. Which encodes a subset of opengl over the wire, which they use to for webgl targets.

If somebody liked moderngl syntax a lot more than Gloos they could target GLIR with a moderngl - like API.
http://api.vispy.org/en/latest/gloo.glir.html