…esorbed species

…ut file

… the input file"

This reverts commit 92fad47.
also "fig adjacencyListGroup in input" which was a fixup.

The functionality has been available for 2 years since
ReactionMechanismGenerator#2185
so adding a new syntax for doing it would be confusing and unnecessarry.

SurfaceChargeTransfer and ArrheniusChargeTransfer are not yet defined
(on this branch). Revert this commin once they are.

For now it just tests Arrhenius types, but should eventually
work on others.

maximumSurfaceSites and maximumSurfaceBondOrder

Doesn't make a real difference, just easier for a human 
to parse the code without making errors, when the "if" 
blocks are in the same order as the documentation and examples.

If you happen to pick the first of the list then the
or or method was faster. But if you pick
the last of the list (or not in the list) then the
in tuple method is faster.

Also changed some lists to sets for consistency,
and because they're a touch faster.

See pylint https://pylint.pycqa.org/en/latest/user_guide/messages/refactor/consider-using-in.html

(Change requested in code review)

Also switched a format to an fstring.

They describe some kinetics comments that were moved in 
921d425
and removed in 
22e6a13

Must be left over from the PEP8-ify project.

The only way r would be 'None' is if a symmetrical reaction (reactants == products)
was found in a library, and hit this check in check_for_existing_reaction that's called by 
make_new_reaction. But that suggests some other problem. So I doubt this test is often needed.
But Matt added it for something.

I can't see any way it would be False, so now we check it's not None.
(as requested in code review)

Unfortunately we need to revert the commit titled
"Remove charge transfer types from average_kinetics (to be reverted)"
at some point in the not so distant future, and 
doing this different averaging scheme now would mess 
that up.