Connectivity table about cmiles

[jchodera]

I'm not sure that's necessary. InChI already also includes a connectivity table encoded as one of its layers, I believe.

[ChayaSt]

@jchodera, as far as I understand, the connectivity in the standard InChI is tautomer-invariant valance bond connectivity so the hydrogens are not explicitly numbered there. The connectivity table will also include the hydrogens and single / double / triple / aromatic bond.

Also, the canonicalization algorithm for the numbering is different than canoncial ordering of OpenEye or RDkit that is used to generate the mapped SMARTS. The connectivity will then have to be mapped onto the map indices in the SMARTS.

[jchodera]

What's the utility of having the complete atomic connectivity table if we have this information already represented in other ways (such as the atom-tagged SMILES)?

[ChayaSt]

Yes, you can get the information from the atom-tagged SMILES so it's not essential.

A better way to phrase this question is should we add this utility to cmiles? And if we do add something like this to cmiles utility, should we also have mapping from and to chemical graphs and SMILES here?
Currently these utilities are in fragmenter. But it might make more sense to have them here.

[ChayaSt]

Addressed by #13