Problem / Goal

As we will use plotting for diagnostics of the model and different runs, we want to have much more certainty that they are working as expected. Elements of the plots can be misleading, such as colour scale choices, and there are a range of options for differencing both in methodology and strategy (e.g. per frame or precomputed) making quickly checking manually difficult.

Would be better to have a set of tests that check with dummy data that everything is calculated correctly and the plots are made as expected. While unit tests are important, e2e tests are the automated version of "running it and seeing if it does what i think it will do". Fixtures in pytest are very useful for this purpose.

What does good look like?

• Add fixtures for the plotting
• Use dummy data
• e2e tests on the plotting pipeline

Related context

• Related Add more tests #62