Filter results by SMIRKS or other chemical descriptors

[mattwthompson]

No description provided.

[mattwthompson]

In #44 I have a proposal that adds a couple of functions that return lists of IDs which can then be fed through to the various MoleculeStore.get_x methods. The idea is something like

In [20]: store.get_dde('openff-2.1.0', molecule_ids=store.filter_by_smirks("[#6]#[#6]"))
Out[20]:
[DDE(qcarchive_id=37016444, force_field='openff-2.1.0', difference=-5.333306021926276e-06),
 DDE(qcarchive_id=37016445, force_field='openff-2.1.0', difference=-0.6762625367734985),
 DDE(qcarchive_id=37016446, force_field='openff-2.1.0', difference=nan)]

In [21]: from yammbs.checkmol import ChemicalEnvironment

In [22]: store.get_rmsd('openff-2.1.0', molecule_ids=store.filter_by_checkmol(ChemicalEnvironment.Alkyne))
Out[22]:
[RMSD(qcarchive_id=37015996, force_field='openff-2.1.0', rmsd=0.7047817362519956),
 RMSD(qcarchive_id=37016945, force_field='openff-2.1.0', rmsd=0.022691976546048032),
 RMSD(qcarchive_id=37016286, force_field='openff-2.1.0', rmsd=0.03265649619028413),
 RMSD(qcarchive_id=37016444, force_field='openff-2.1.0', rmsd=0.009567076814051558),
 RMSD(qcarchive_id=37016445, force_field='openff-2.1.0', rmsd=0.009781303169628252),
 RMSD(qcarchive_id=37016446, force_field='openff-2.1.0', rmsd=0.00955735978797403)]

In [23]: [store.get_smiles_by_molecule_id(store.get_molecule_id_by_qcarchive_id(id_)) for id_ in [37015996, 37016945, 37016286]
    ...: ]
Out[23]:
['[H:15][c:3]1[c:5]([c:9]([c:13]([c:10]([c:6]1[H:18])[H:22])[C:1]#[C:2][c:14]2[c:11]([c:7]([c:4]([c:8]([c:12]2[H:24])[H:20])[H:16])[H:19])[H:23])[H:21])[H:17]',
 '[H:7][C:3]#[C:1][C:2]([H:4])([H:5])[H:6]',
 '[H:8][C:3]#[C:1][C:2](=[C:4]([H:6])[H:7])[H:5]']

I'm not sure why the results are different; the molecules missing from the first query sure like like alkynes to me:

[mattwthompson]

It would be worth adding atom tags to the SMIRKS pattern above (even though I'm pretty sure the toolkit just passes them through as SMARTS to RDKit/OEChem?). Otherwise the funkiness might be due to aromaticity

[mattwthompson]

@ntBre noticed that this comparison has two differences - DDE vs RMSD in addition to Checkmol/SMARTS filtering. Being consistent with either produces results one expects. Thanks Brent!

In [3]: store = MoleculeStore("yammbs/_tests/data/ch.sqlite")

In [4]: store.get_dde('openff-2.1.0', molecule_ids=store.filter_by_checkmol(ChemicalEnvironment.Alky
   ...: ne))
Out[4]:
[DDE(qcarchive_id=37016444, force_field='openff-2.1.0', difference=-5.333306021926276e-06),
 DDE(qcarchive_id=37016445, force_field='openff-2.1.0', difference=-0.6762625367734985),
 DDE(qcarchive_id=37016446, force_field='openff-2.1.0', difference=nan)]

In [5]: store.get_dde('openff-2.1.0', molecule_ids=store.filter_by_smirks("[#6]#[#6]"))
Out[5]:
[DDE(qcarchive_id=37016444, force_field='openff-2.1.0', difference=-5.333306021926276e-06),
 DDE(qcarchive_id=37016445, force_field='openff-2.1.0', difference=-0.6762625367734985),
 DDE(qcarchive_id=37016446, force_field='openff-2.1.0', difference=nan)]

In [6]: store.get_rmsd('openff-2.1.0', molecule_ids=store.filter_by_smirks("[#6]#[#6]"))
Out[6]:
[RMSD(qcarchive_id=37015996, force_field='openff-2.1.0', rmsd=0.7047817362519956),
 RMSD(qcarchive_id=37016945, force_field='openff-2.1.0', rmsd=0.022691976546048032),
 RMSD(qcarchive_id=37016286, force_field='openff-2.1.0', rmsd=0.03265649619028413),
 RMSD(qcarchive_id=37016444, force_field='openff-2.1.0', rmsd=0.009567076814051558),
 RMSD(qcarchive_id=37016445, force_field='openff-2.1.0', rmsd=0.009781303169628252),
 RMSD(qcarchive_id=37016446, force_field='openff-2.1.0', rmsd=0.00955735978797403)]

In [7]: store.get_rmsd('openff-2.1.0', molecule_ids=store.filter_by_checkmol(ChemicalEnvironment.Alk
   ...: yne))
Out[7]:
[RMSD(qcarchive_id=37015996, force_field='openff-2.1.0', rmsd=0.7047817362519956),
 RMSD(qcarchive_id=37016945, force_field='openff-2.1.0', rmsd=0.022691976546048032),
 RMSD(qcarchive_id=37016286, force_field='openff-2.1.0', rmsd=0.03265649619028413),
 RMSD(qcarchive_id=37016444, force_field='openff-2.1.0', rmsd=0.009567076814051558),
 RMSD(qcarchive_id=37016445, force_field='openff-2.1.0', rmsd=0.009781303169628252),
 RMSD(qcarchive_id=37016446, force_field='openff-2.1.0', rmsd=0.00955735978797403)]