bug fix in SpacegroupAnalyzer

[gpetretto]

There is a problem in the get_conventional_standard_structure method of the SpacegroupAnalyzer.

The code below demonstrates the issue:

from pymatgen.symmetry.analyzer import SpacegroupAnalyzer
from pymatgen.analysis.structure_matcher import StructureMatcher
from pymatgen.ext.matproj import MPRester

s = MPRester().get_structure_by_material_id("mp-761986")

sm = StructureMatcher(ltol=0.0001, stol=0.0001, angle_tol=0.1, primitive_cell=False, scale=False)

spga = SpacegroupAnalyzer(s)
sc = spga.get_conventional_standard_structure()
s_ref = spga.get_refined_structure()

print(sm.fit(s, s_ref))
print(sm.fit(s, sc))

Output:

True
False

This comes from the fact that in the triclinic case the reference structure can be modified by the call to struct.get_reduced_structure("LLL"), but the lattice object is not updated accordingly.

[coveralls]

Coverage decreased (-0.6%) to 83.011% when pulling 381b843 on gpetretto:bugfix into ce2fb4e on materialsproject:master.

[mkhorton]

Thanks for fixing @gpetretto -- would it be possible to add a quick test case to catch this too? Thank you