Closed danielhollas closed 11 months ago
Attention: 2 lines
in your changes are missing coverage. Please review.
Comparison is base (
a1b5134
) 79.39% compared to head (528e74b
) 79.98%.
:umbrella: View full report in Codecov by Sentry.
:loudspeaker: Have feedback on the report? Share it here.
Good point @danielhollas indeed I do not find anymore cases that justify openbabel so from my side we can remove it and if I find some new problematic cases we can think about how to handle them.
I tested an old molecule I had that gave me problems but also with openbabel seems not to work.
I add it here as another problematic case not working at present:
[H][C]1C2=CC=CC3=C2C(C4=C1C=CC5=C4C6=C(C7=C([C]([H])C8=CC=CC9=CC=CC7=C98)C=C%10)C%10=CC=C6C=C5)=CC=C3
a correct (not best conformer) interpretation of this helicoidal molecule is provided by Mathematica:
As discussed in https://github.com/aiidalab/aiidalab-docker-stack/issues/388 we want to remove openbabel dependency from AWB because it cannot be installed by pip. We use it here only as a fallback to RDKit in SmilesWidget.
In my experience, it is not needed once the RDkit SMILES generation is more stable. @cpignedoli I would be interested if you have any SMILES that needed Openbabel. Please open a separate issue if you have any that still don't work for the current code. Thanks!
This also includes a little cleanup of the canonicalization code.