adjacent_res_smarts requires more atoms in adjacent_mol than capping residues (NME, ACE, phosphate) can provide

[rwxayheee]

Hi, I was trying to register a phosphate capping group (similar to NME, ACE in proteins) as a new chem template. Just realized that this can be reproduced by protein examples.

To reproduce -
Protein structure 2G5Z is a green fluorescent protein (GFP) variant. Chain A is part of the beta-barrel and finishes with NME.

As the input, a stripped PDB file for chain A:
2g5z_A.pdb.txt

Command:

mk_prepare_receptor.py --macromol 2g5z_A.pdb --skip_gpf -o 2g5z

Output:

Traceback (most recent call last):
  File "/Users/amyhe/micromamba/envs/mk_dev/bin/mk_prepare_receptor.py", line 7, in <module>
    exec(compile(f.read(), __file__, 'exec'))
  File "/Users/amyhe/Desktop/0_forks/Meeko/scripts/mk_prepare_receptor.py", line 424, in <module>
    chorizo = LinkedRDKitChorizo.from_prody(
  File "/Users/amyhe/Desktop/0_forks/Meeko/meeko/linked_rdkit_chorizo.py", line 798, in from_prody
    chorizo = cls(
  File "/Users/amyhe/Desktop/0_forks/Meeko/meeko/linked_rdkit_chorizo.py", line 699, in __init__
    padded_mols = self._build_padded_mols(self.residues, bonds, padders)
  File "/Users/amyhe/Desktop/0_forks/Meeko/meeko/linked_rdkit_chorizo.py", line 1253, in _build_padded_mols
    padded_mol, mapidx = padders[link_label](
  File "/Users/amyhe/Desktop/0_forks/Meeko/meeko/linked_rdkit_chorizo.py", line 1990, in __call__
    raise RuntimeError(
RuntimeError: length of hits must be 1, but it's 0

I think the error occurred when it's trying to add the "5-term" padding to the normal residue that's adjacent to NME, when trying to find expected adjacent_res_smarts O=C[C:11][N:12] from its adjacent_mol that is NME.

Is it safe to make adjacent_res_smarts shorter? Or do we always check atoms in 1-4 range? If adjacent_res_smarts can't be shortened, then the capping groups might need to be handled differently.

The quickest walkaround would be to un-register the capping residues from chem templates.

[rwxayheee]

closing becuase #185 is merged