Unexpected behavior in CrystalNN: very low coordination numbers when adsorbates are present

[Andrew-S-Rosen]

I have the following structure of CO on a Rh surface. When using CrystalNN() on a surface Rh atom without an adsorbate, I get a coordination number of 9. When getting the coordination number of the surface Rh with CO bound to it, I get a coordination number of 1. One would expect the coordination number of the Rh with CO on it to be 1 greater than its neighboring sites. I'm not sure that this is a bug, but it's definitely unexpected! I have to dig into the root cause. @mkhorton, you might be interested.

from pymatgen.core import Structure
from pymatgen.analysis.local_env import CrystalNN
s = Structure.from_file('test.cif')
cnn = CrystalNN()
cn1 = cnn.get_cn(s,0) # Rh surface site with CO (returns 1)
cn2 = cnn.get_cn(s,1) # Rh surface site without CO (returns 9)

test.cif.txt