Smith: more informative legend entries

[dweindl]

Use more informative legend entries and prettify some labels.

Previously:

🤔

Also fixes some misleading dataset IDs (compare https://link.springer.com/article/10.1186/1752-0509-7-41).

[dilpath]

Looks good, thanks! I prefer tex-style formatting e.g. "H_$2$O_$2$" for consistency across the benchmark problems.

[dweindl]

I prefer tex-style formatting e.g. "H_$2$O_$2$" for consistency across the benchmark problems.

Please specify somewhere, where tex should be used and where unicode can be used. If unicode does the job, I prefer that, but I can change it if necessary.

[dilpath]

I prefer tex-style formatting e.g. "H_$2$O_$2$" for consistency across the benchmark problems.

Please specify somewhere, where tex should be used and where unicode can be used. If unicode does the job, I prefer that, but I can change it if necessary.

A possible issue is that there might be a visualization issue if some parts are unicode and other parts are tex. e.g. subscript 2 might look different depending on the choice. Currently, I am only aware of ASCII or tex-style labels in the benchmark collection, but I might be wrong... I didn't think about it until now, but if you agree with the possible issue, then I would choose tex since it's more flexible, and can add this to the docs.

[dweindl]

There are pros and cons for either. If you use tex, it will use a different font for that part. This is fine for actual math, but I don't find it very helpful for some biochemical species. There are already some examples where µ is used instead of $\mu$, because $\mu\mathrm{M}$ doesn't look too pretty.

[dweindl]

Cons for unicode: The respective glyphs might not be supported by all fonts. In that case, better a different latex font than just rendering tofu.