correct merging of [ atomtypes ] when bondtypes are used within.

[zetadin]

Merging [ atomtypes ] with _merge_FF_files() in alchemy.py was retaining only the first 6 columns.
However when [ atomtypes ] also had a bond type or an atom number column, the final epsilon column would get cut off.
Eg.
c3 c3 0.00000 0.00000 A 3.39967e-01 4.57730e-01
would become
c3 c3 0.00000 0.00000 A 3.39967e-01
when we want
c3 0.00000 0.00000 A 3.39967e-01 4.57730e-01

This change parses the input [ atomtypes ], recognizes whether these optional columns are present, and outputs only the required columns (ie no bond type and no atom number) into the merged file.

There is also a new unit test to verify functionality on both GAFF and OPLS atomtypes formats.
CGENFF ligands tend to directly include the charmm ff and have empty [ atomtypes ] sections, so testing on them is pointless.