Input_file_handling_robustness

[tbody-cfs]

This PR makes the input file handling routines more flexible and robust.

1. Lets you not specify any grid entry in the input file. You can now run a scalar case.
2. Let's you not specify any algorithm entry in the input file. This makes a Algorithm.empty which does nothing at all (neater than handling no algorithms as some special case).
3. As a consequence of allowing scalar cases, adds a minor modification to sanitize_variable which works around scalar-valued coordinates (see Multiple methods fail on a DataArray with scalar dimension pydata/xarray#1709). This has the pleasing side-effect of getting the correct dimension name in the regression JSON.
4. Adds cfspopcon as an entry point, equivalent to popcon. Now you can either poetry run popcon or poetry run cfspopcon — these do the same thing.
5. Lets you pass dictionaries to input_file_handling instead of just filepaths, in case you want to programmatically set input values.
6. Runs ruff on tests to tidy this up

[hassec]

Looks good, just some small comments.

Regarding the Alogrithm.empty(), I'm not sure if I follow how it would be used. If you have a case that doesn't specify any algorithms, what would you do with that case?

[tbody-cfs]

Algorithm.empty() is useful if you want to set up a dataset based on an input file. Often useful in a development step, where you want to load up a case and then develop an algorithm in a jupyter notebook.

[hassec]

Just one more tiny comment, otherwise looks great 👍