Change "OS" oxygen atoms in silica mol2 file to O

[justinGilmer]

PR Summary:

With the latest updates to mBuild in version 0.9.3, the atom names when
being converted to various file formats or parmed structures now undergo
a different process when their elements are being inferred.

Previously, an element name like OS, like in this silica.mol2 file
would evaulate to O, since OS is not an element.

This usually caused more issues though, because certain elements like
SI would be inferred as sulfur.

This now causes issues with some of our tests that rely on this silica
mol2 file, as OS is now being discovered as osmium.

The OS elements in the mol2 file have been renamed to O.

PR Checklist

---

• Includes appropriate unit test(s)
• Appropriate docstring(s) are added/updated
• Code is (approximately) PEP8 compliant
• Issue(s) raised/addressed?

[mattwthompson]

If anybody were to compile a list of frustrations using biophysics software for materials simulations, the ambiguity of atom types/names/elements in mol2 files would make the cut

[mattwthompson]

Looks like a big :%s/OS/O/g to me, skimming it I didn't find anything else

[justinGilmer]

Thats pretty much what I did, just making sure not to override certain TRIPOS names to keep it as a valid mol2

[codecov]

Codecov Report

Merging #260 into master will not change coverage.
The diff coverage is n/a.

@@           Coverage Diff           @@
##           master     #260   +/-   ##
=======================================
  Coverage   85.76%   85.76%           
=======================================
  Files          10       10           
  Lines        1138     1138           
=======================================
  Hits          976      976           
  Misses        162      162

[ahy3nz]

If anybody were to compile a list of frustrations using biophysics software for materials simulations, the ambiguity of atom types/names/elements in mol2 files would make the cut

"HA" and "CA" are atom names for some amino acids, then the mbuild to_parmed capitalization correction would convert to "Ha" and "Ca", then get tagged by parmed as Hahnium and Calcium, then ruin any sort of atomtyping. This probably just enforces the use of the underscore-virtual-element convention