Feature suggestion: Support for additional stereochemistry descriptors in SMILES

[gmseabra]

This is feature suggestion.

I run into this while trying to use SMILES to enter the cisplatin molecule. If one tries that:

Chem.MolFromSmiles("[Pt](N)(N)(Cl)(Cl)")

The result is a planar version of what seems to be "trans-platin". The reason seems to be that the SMILES interpreter does not recognize the square-planar geometry in the Pt center.

However, it turns out that the SMILES standard does provide an alternative stereochemistry specification. See here (look for 3.9.6 Other Chiral Configurations) and here (look for 3.3.4 General Chiral Specification), although the implementation of this feature seems to be very rare.

Implementing these alternative symbols would allow to describe the metal coordination better. In this system, for example, cisplatin would be simply: "N[Pt@SP1](Cl)(Cl)N". The standard also includes specifications for trigona-bipyramidal and octahedral geometries.

[greglandrum]

Hi @gmseabra : thanks for the suggestion!
We do plan to, eventually, improve the stereochemistry handling and add support for non-tetrahedral stereochemistry, but that's a bug chunk of work and I don't currently have an estimate for when it will be done.