Tests running multiple configurations + some cleanup

We could probably save some time by running a few configurations in parallel. I hope to do a few things here:

• Allow a few configurations to be run at once
• Get rid of warnings because tools are instanced a few times under the same name. Probably just means mangling the tool names more carefully.
• Use pathlib to handle configuration file paths and input file paths.
• Add a unit test

While I was doing this I did some other cleanup:

• Migrated a few more scripts to use the common command line parser
• Use std::filesystem for paths in C++ code
• Made the usage printout for test-dumper fill possible commands automatically
• Enable more flake8 tests, make jobs fail if flake8 fails

I'm going to ask @mguth and @pgadow if they can take a quick look. We're a bit short staffed so I'm allowing "self approval" for MRs right now, but if you can take a look in the next few hours and try to spot obvious issues that might help.