Inefficient Animation rendering

[suryadheeshjith]

Possible inefficiencies due to matplotlib - cartopy and create_animation() function.

See OM4_SE_animation.ipynb Link to notebook and here.

[suryadheeshjith]

The initial implementation provided by the helper functions manually saves each frame as a PNG file, then reads them back into a PIL Image list to create a GIF.

[IamShubhamGupto]

Did some investigating today and even after caching variables, I got a build time exceeding 10 minutes.

a reference to what we did in the l96 demo - https://m2lines.github.io/L96_demo/notebooks/intro_ML_and_NNs.html?highlight=animation

[jbusecke]

Just came across this on X, might be relevant to the animation issue

[IamShubhamGupto]

@suryadheeshjith can we close this?

[suryadheeshjith]

I need to explore more options for rendering before we close this issue!

[suryadheeshjith]

I tried using matplotlib.use("Agg") and it didn't help too much with or without saving temporary images. I also tried using hvplot but plotting a similar animation is much more involved. It would probably be best to keep what we have now and create a new notebook for these newer libraries.

[jbusecke]

I would suggest to keep this issue open since the root problem is not yet addressed and it might affect multiple notebooks. I do not think it is a top priority (we could use labels to indicate priority levels), but closing this indicates that the problem has been solved.