Moltemplate

The changes in this commit can significantly increase the speed of moltemplate.sh when the user's LT files contain wildcard characters ('*' characters) in any of the LAMMPS coeffs or coeff commands. In particular, simulations using the DREIDING force field ("dreiding.lt", which uses wildcard characters frequently) should be faster to prepare. Some coarse-grained force fields also make frequent use of wildcard characters.

Fixed a serious bug (the number of bond types in the generated data files are incorrect). LAMMPS would crash when run using data files generated by moltemplate. This bug was caused by a typo introduced in version v2.20.0 on 2022-1-11.

Fixed a bug introduced in the previous version (v2.20.0) causing moltemplate to misinterpret blank lines containing nothing but white space (spaces or tabs). Thanks to github user conandolley103 (Tue Le) for reporting this bug!

1) Improved compatibility with ltemplify.py. This required adding better support in moltemplate.sh for atom types and names containing whitespace (which ltemplify.py frequently generates).
2) Improved performance and behavior of dump2data.py and raw2data.py.

This update might break some people's LT files.

1) Now you can omit the bond type in genpoly_lt.py.
2) The "-inherits" argument finally works in genpoly_modify_lt.py.

1) Bugfix: read dump now handles ellipsoid orientation correctly.
2) Corrected the MOLC citation.

Thanks to Otello Roscioni for discovering both problems and correcting them.