Cores with query atoms may fail to R-group-decompose molecules

[ptosco]

Describe the bug
Attempting to carry out R-group decomposition with cores containing query atoms may result in unexpected failures to decompose molecules that do match the core.

To Reproduce

import rdkit
from rdkit import Chem
from rdkit.Chem import rdRGroupDecomposition

# 1st example
core = Chem.MolFromSmarts("[n]1([*:1])ccccc1")
core

params = rdRGroupDecomposition.RGroupDecompositionParameters()
params.onlyMatchAtRGroups = True
rgd = rdRGroupDecomposition.RGroupDecomposition(core, params)
pyridine = Chem.MolFromSmiles("n1ccccc1")
rgd.Add(pyridine)
-1
N_methyl_pyridinium = Chem.MolFromSmiles("[n+]1(C)ccccc1")
rgd.Add(N_methyl_pyridinium)
0

# 2nd example
core = Chem.MolFromSmarts("[n]1([*:1])ccccc1")
core

params = rdRGroupDecomposition.RGroupDecompositionParameters()
params.onlyMatchAtRGroups = True
rgd = rdRGroupDecomposition.RGroupDecomposition(core, params)
methylcyclohexene = Chem.MolFromSmiles("C=1(C)CCCCC1")
rgd.Add(methylcyclohexene)
-1
methylcyclohexane = Chem.MolFromSmiles("C1(C)CCCCC1")
rgd.Add(methylcyclohexane)
0

Expected behavior
In the 1st example, pyridine should be decomposed, with the R-group being null.
In the 2nd example, methylcyclohexene should be decomposed, with one of the R-groups being null and the other being [1:*]C.

Additional context
Both problems have a common cause.
The RGD algorithm initially adds hydrogens to molecules in order to give them a chance to match core R-groups in case there is no heavy atom available to match them.
However, pyridine and methylcyclohexene will not add any hydrogens at the sp2 nitrogen or carbon, respectively, as they their nitrogens are already all filled.
Addressing this bug requires a fundamental change into how the matching of molecules and cores is done, as adding explicit Hs to the molecules may not work when query atoms are present in the core.