Opt.Output -> Opt.separate

[stestoll]

This PR addresses #176 and implements new settings for Opt.Output

Opt.Output = 'components';
Opt.Output = 'sites';
Opt.Output = 'orientations';
Opt.Output = 'transitions';

Support for 'separate' is removed.

[cetait]

I've looked through the changes in this PR in a bit more detail and I think it is very useful. A couple of comments:

• The third output in pepper containing only transitions for an Opt.separate = 'components' simulation, e.g. for isotopologues, might be confusing. Maybe it's worth including a field listing the isotopologues as well and some indication what the subspectra represent for the specific case in the out structure?
• Using Opt.separate = 'transitions' for chili currently gives an "Unknown value" error. It might be better to replace this error with a more explicit message mentioning that separate transitions aren't implemented for chili (as currently done in garlic for example).
• It might be useful to have an example using Opt.separate = 'orientations' as well.
• Could an option for returning separate spectra for the different nuclei in salt be a useful option?
• Opt.separate is not yet documented for saffron.
• In general, clearer information on what the different subspectra represent could be useful (e.g. isotopologues given explicitly in the out structure).

[cetait]

The detailed information in the info structure is great. I've added the specific combination of isotopes, since that information is easily accessible within compisoloop now and could be useful for the user too. This is in a separate output field present only if there are different isotopologues, I'm not sure about the fieldname.