husakm

Can I generate OBDotNet.dll, 64 version for 3.1.1 by recompilin the OB source code in Visual Studio 2017 ?

I am afraid you must go to the source code of the CIF import in OB and investigate yourself.

OB use an database of HM symbols. This database cover only a subset of possible symbols (all standard, some non standard). It does not for sure cover all possible symbols...

Maybe the molecule is protonized ? OB fails to assign correct atom types e.g. for cycle with protonized N ...

I can confirm both 6.5.0 and 6.7.0 works OK if i put in the directory the `libiomp5md.dll` from the` xtb-6.5.0-windows-x86_64.zip` file. The 6.5. reports "Input file format not suitable for...

I apology for miss-information. It sounds like my `benzene.coord` file was wrong and the reason of the crash. I had re-created the file from data on your www and it...

Sure, I will post a salt CIF file demonstrating the problem. Give me 2 days, I am out of office ..., Generaly the issue is not CIF specific, the Kekulization...

To make the problem simple and clearly visible, I had separated the issue form CIF file usage and I will demonstrate it on .XYZ file. [KAXWOV_cation_xyz.txt](https://github.com/openbabel/openbabel/files/15184837/KAXWOV_cation_xyz.txt) The file in attachment...

The problem is clearly demonstrated on the image in attachment. The Dreading atom types on the right part of the image are correctly assigned by a commercial engine (Materials Studio)...

So you suggest: A) Correct the OpenBabel engine (implement charged fragments support on our side) or B) Switch from OpenBabel to RDkit, witch alredy have the required functionality ? Generaly...