lmk if you think this is too much to do on the user's behalf. i think its nice though since they can always override if they don't like what it gives.

I like it. I like the combination of do nice things on the user's behalf, while giving them the power to override easily. It's also pretty clear where the code slots in, so it's easy to remove as needed.

Any reason not to use d3.lab(color).l >= 60 or similar here instead? Feels simpler than implementing YIQ and should be at least as effective.

I broke out the yiq function separately… I think you could now inline this function into the widget above, and also consider using d3.lab instead.

Sounds great. I'll let Ian reply regarding if there was a reason to do YIQ instead of d3.lab instead since he implemented that part. Looks like he was adopting the approach from this notebook: https://observablehq.com/@vsaarinen/text-color-based-on-luminance

Looking at the notebook comparing luminance only vs yiq; I do think yiq is preferable. I see more "misses" with d3.lab vs yiq... even when I tweak the cut-offs for d3.lab. I think d3.lab would be totally fine, but if there is a better option available, why not just go with it?