Issue in CDX → SMILES Conversion Using Indigo.NET (v1.30.1)

[allasairadha]

We have been using the Indigo.NET library (v1.5.0) in our production workflows for some time. In that version, the iterateCDXFile(filepath) method was working as expected and generated correct SMILES for structures stored in CDX files.
After upgrading to the latest Indigo.NET version (v1.30.1), we are seeing incorrect SMILES being generated for the same CDX files, using the same code path.

Please see the attached structure in CDX file that reproduces the issue.
For example, for the following structure in a CDX file:
SMILES previously generated (v1.5.0): NC(C)C(=O)O (correct)
SMILES currently generated (latest version):
N*(*)C(=O)O |$;CH;CH3;;;$|
The new SMILES output appears chemically incorrect and seems to include placeholder atoms / attachment points instead of the expected explicit atoms.
We are using the following approach in C#:
foreach (IndigoObject obj in indigo.iterateCDXFile(cdxPath))
{

    string smiles = obj.smiles();
    // processing...

}
Could you please advise on:
Any recommended workaround or settings to restore the previous behavior.

Thank you for your support.

[AlexanderSavelyev]

please attach the cdx file

[allasairadha]

please attach the cdx file
amino.zip

[allasairadha]

Any suggestions to get normal smiles and not CXSMILES?

[AlexanderSavelyev]

cdx and cdxml files are supported by the standard loadMolecule function. Here is the python code, but Indigo.NET should work the same

import indigo
import os

indigo = indigo.Indigo()

print(indigo.version())

mol = indigo.loadMoleculeFromFile(os.path.join(os.path.dirname(__file__),
                                  'data', 'amino.cdx'))
print(mol.smiles())

1.36.1.0-g9d855db19-x86_64-linux-gnu-11.2.1
NC(C)C(=O)O