Should we support SNOwGLoBES materials with mixed binning?

[JostMigenda]

[Raised in a PR comment by @Sheshuk.]

We currently enforce that, for a given material (defined in SNOwGLoBES/channels/channels_{material}.dat, all channels have the same binning.
This assumption is currently violated by the argon_400bins material; which mixes the “usual” SN interaction channels (400 bins, 0.5–100 MeV) with CEνNS (200 bins, 0.5 eV to 10 keV).

In practice, I’m not sure anyone is actually using that, since detectors sensitive to CEνNS are usually far too small to observe a non-zero event rate in the “usual” channels. And even if some detector needed both, there’s the straightforward workaround of using e.g. the argon material for the “usual” channels and argon_coh for CEνNS; and then combine the results from both manually in their analysis as needed. (I suspect it’s preferable to analyse both sets separately anyway—to give just one example, the vastly different energy ranges, rates, etc. would make it difficult to put both in the same plot.)

If this is actually needed in practice and these users would prefer to analyse both data sets together, we should look into whether we can implement this natively rather than require this workaround.