MolSym fails to create SymText for O2

[nlk36701]

The qcmol MolSym Molecule:
O 0.00000000 0.00000000 -1.13383568
O 0.00000000 0.00000000 1.13383568

Traceback (most recent call last):
File "run.py", line 64, in
job.run_nosym()
File "/home/vulcan/nlk36701/github/QSymm/SORHF/rhf.py", line 30, in run_nosym
self.symtext = molsym.Symtext.from_molecule(qcmol)
File "/home/vulcan/nlk36701/github/another/steef_7/MolSym/molsym/symtext/symtext.py", line 63, in from_molecule
return Symtext(mol, rotate_to_std, reverse_rotate, pg, symels, atom_map, None, irreps, irrep_mat)
File "/home/vulcan/nlk36701/github/another/steef_7/MolSym/molsym/symtext/symtext.py", line 45, in init
self.get_character_table()
File "/home/vulcan/nlk36701/github/another/steef_7/MolSym/molsym/symtext/symtext.py", line 90, in get_character_table
cc.append(multiply(self.mult_table, sidx2, sidx, inverse(self.mult_table, sidx2)))
File "/home/vulcan/nlk36701/github/another/steef_7/MolSym/molsym/symtext/multiplication_table.py", line 160, in inverse
for i in range(mtable.shape[0]):
AttributeError: 'NoneType' object has no attribute 'shape'

[sgoodlett]

This bug is associated with a bigger issue: what to with linear point groups. We still don't have well defined objects like multiplication tables for linear point groups which is why the bug is thrown. Let's make this issue about enhancing MolSym's capabilities with linear molecules! The first question is: what kind of information do we need to properly handle the symmetries of linear molecules? Then we can design data structures for that.

[sgoodlett]

Putting this issue under #35