Ah, this time the tests saved me; I was a little too eager with just one of the fixes.

Yes you were right! At least the one I was able to test, in the weighted price calculation. It turned out the code path wasn't used before due to an unrelated bug, which I now also fixed. I will say that the whole function seems pretty fragile...

For the CO2 emissions calculations, the code path you pointed out is also never taken! Hence why I didn't find the mistake. It turns out that in sector-coupled perfect foresight networks there aren't any co2-emitting generators anyway; all emissions go the the co2 atmosphere store.

Great! Thanks for looking into it and sorry that the script is so buggy...

test failures seem unrelated, I will merge given previous commit CI ran through.

No worries, happy to help and anyway it's so much better to have anything at all to get started with than having to do everything scratch!