Bug relating to this PF6- still can not get Bad Conformer Id

[teltim]

Relating to this issue PF6- still can not get Bad Conformer Id: #5145
I encounterd the same bug with the recent rdkit

rdkit==2023.3.1

mol=Chem.MolFromSmiles("S(F)(F)(F)(F)F")
mol = Chem.AddHs(mol)
AllChem.EmbedMolecule(mol, randomSeed=42)
conf = mol.GetConformer()
print(conf)

this was not fixed yet?

[15:45:10] UFFTYPER: Unrecognized atom type: S_6+6 (1)
---------------------------------------------------------------------------
ValueError                                Traceback (most recent call last)
Cell In[17], line 14
     12 #AllChem.EmbedMolecule(mol, randomSeed=42)
     13 AllChem.EmbedMultipleConfs(m, numConfs = 10, clearConfs = True, pruneRmsThresh = 0.5)
---> 14 conf = mol.GetConformer()
     15 print(conf)

ValueError: Bad Conformer Id

[greglandrum]

It looks like there's a problem here with the crude approximate bounds which are applied when UFF parameters for an atom cannot be found.

In [24]: m = Chem.MolFromSmiles('S(Cl)(Cl)(Cl)(Cl)(Cl)Cl')

In [26]: ps = rdDistGeom.ETKDGv3()

In [27]: ps.trackFailures = True

In [28]: rdDistGeom.EmbedMolecule(m,ps)
[08:06:29] UFFTYPER: Unrecognized atom type: S_6+6 (0)
Out[28]: -1

In [29]: ps.GetFailureCounts()
Out[29]: (65, 5, 0, 0, 0, 0, 0, 0, 0, 0)

In [30]: ps.useRandomCoords = True

In [31]: rdDistGeom.EmbedMolecule(m,ps)
[08:07:57] UFFTYPER: Unrecognized atom type: S_6+6 (0)
Out[31]: -1

In [32]: ps.GetFailureCounts()
Out[32]: (0, 70, 0, 0, 0, 0, 0, 0, 0, 0)

[teltim]

Thanks for the kind investigation,
Are there any methods that can avoid the error?
there is no method for the time being?

[greglandrum]

Closed in #6379

[teltim]

Closed in #6379

What is the situation of this issue ??

[greglandrum]

As the comment says: it has been fixed. It will be in the next patch release

[teltim]

Thanks for the comment,
Is it possible to download modifiyed version in advance? Best regards.

[greglandrum]

Thanks for the comment, Is it possible to download modifiyed version in advance? Best regards.

The code is in github, so if you are used to doing your own RDKit builds you can try it now. Otherwise it's probably best to wait for the patch release, which probably comes late next week.

[teltim]

Thank you for the comment! it was solved!

[ouielba90]

Is the patch out (which includes such a fix) already?