Nice! If alt-text is added to the .png metadata file this way, and the .png file is added to a web page (ie, https://heftieproject.github.io/zarr-benchmarks/), does web browser "know" this is where to find the alt-text, and display it when moused over (or say it when a screen reader is being used)? I assumed that we might have to manually add the alt-text to the markdown file, but if we can semi-automate it like this that would be a bit easier.

Following on from our meeting, out of interest where did you get inspiration for this implementation from? I'm intrigued if adding alt-text to .png files in this way is some sort of standard.

Following on from our meeting, out of interest where did you get inspiration for this implementation from? I'm intrigued if adding alt-text to .png files in this way is some sort of standard.

https://stackoverflow.com/questions/58399070/how-do-i-save-custom-information-to-a-png-image-file-in-python
It may just be if you're using Pillow so probably not standard. I added alt-text to the main PR #79 instead of using this method.

Should we close this in favour of your changes in the other PR then?

We decided this isn't needed, and instead alt-text will be added manually to the markdown files.