The plotting tests are parametrized by all of the pre-defined figure sizes. This makes sense from an automated testing perspective, but the PDF files generated by the plotting tests are also sometimes used for manual visual inspection to check for regressions/bugs that are hard to detect programmatically.
Some of the smaller figure sizes are too small for some of the test plots, so the plots end up squished. This is expected behavior but it looks wrong and clutters the process of manually inspecting the PDFs. We should figure out a sensible way to filter the PDFs to avoid this. This filtering is probably best implemented in the pytest fixture that saves the PDFs generated by the tests (which are otherwise ephemeral).
The plotting tests are parametrized by all of the pre-defined figure sizes. This makes sense from an automated testing perspective, but the PDF files generated by the plotting tests are also sometimes used for manual visual inspection to check for regressions/bugs that are hard to detect programmatically.
Some of the smaller figure sizes are too small for some of the test plots, so the plots end up squished. This is expected behavior but it looks wrong and clutters the process of manually inspecting the PDFs. We should figure out a sensible way to filter the PDFs to avoid this. This filtering is probably best implemented in the pytest fixture that saves the PDFs generated by the tests (which are otherwise ephemeral).