cimm-kzn/cgrtools

CGRs, molecules and reactions manipulation

cgrtools's Issues

Converting SMILES string into molecule container

Hello! I wonder if it is possible to convert individual SMILES string into a molecule container object? Without working with files

Inchis of single atoms read from xyz files different than those generated by openbable.

I am trying to use your tool to generate inchi strings. This is my way around of installing the obenbabel. Here is the script I wrote:

with cgr.files.XYZRead(file='h.xyz') as f:
    xyz = f.read()
xyz = xyz[0]
rdkitMol = cgr_utils.to_rdkit_molecule(data =xyz)
inchi = rdkit.Chem.inchi.MolToInchi(rdkitMol)

As you can see I go from the cgr molecule object into rdkit mol and then do the conversion to inchi. The problem is that for, e.g. a single hydrogen atom xyz file:

1

H 0.0 0.0 0.0

the produced inchi is as follows: InChI=1S/H/q-1 where it should rather be InChI=1S/H
I also noticed that in the MoleculeContainer for atomic hydrogen, the molecular_charge property was set to -1. Should not it rather be 0?

Similarly for the following xyz input:

1

F 0.0 0.0 0.0

the corresponding inchi is: InChI=1S/FH/h1H/p-1 as compared to InChI=1S/F
Similarly to hydrogen example, the molecular_charge property is also -1.

I wonder if there are any settings I could try while parsing the xyz files or converting to the rdkit mol object so that the resulting molecules would produce the expected inchi strings.

Just to say that for all larger molecules I have converted with my script the produced inchis were identical to those from openbabel.

not all SMILES lines are read in a reaction SMILES file

The following script doesn't translate all lines to the rdf format:

#!/usr/bin/env python3

import sys
from CGRtools import SMILESRead
from CGRtools import RDFWrite

input_fn = sys.argv[1]
count = 0

with SMILESRead(input_fn) as input:
    with RDFWrite(sys.stdout) as output:
        for d in input:
            output.write(d)
            count += 1

print('read %d' % count, file = sys.stderr)

it would be nice a CGR_from_reaction method

currently, you need to use the tilde operator, but this is not very easy to find / explicit
when you are new to this package.
Also, there could be a to_CGR method in the reaction object/container.