Correctly sets the diagonal of the Burden matrix with charge values. Fixes bug 3347528
Looks good to me. Funny bug!
I'll apply it later.
Rajarshi, one unit test value changes:
Testcase: testAromaticity(org.openscience.cdk.qsar.descriptors.molecular.BCUTDescriptorTest): FAILED
element 2 does not match expected:<-0.18066486607184457> but was:<-0.21175815206924314>
junit.framework.AssertionFailedError: element 2 does not match expected:<-0.18066486607184457> but was:<-0.21175815206924314>
If that is OK, please provide a patch.
The current patch is applied and pushed.
The problem is not in the BCUT descriptor - instead, the partial charges calculted for the two forms of the molecule (kekule and aromatic) differ - hence the difference. If the charges are meant to differ, then I can update the new BCUT descriptor values to go away.
But I think partial charges should not be representation dependent
This can be closed (I think it's already applied) - the failing test went away with the new SMILES code giving a kekule structure. It will come back again though with the new patches to keep the aromatic specification from SMILES. As Rajarshi says it's the charges which correctly give different values.
Log in to post a comment.
Sign up for the SourceForge newsletter:
You seem to have CSS turned off.
Please don't fill out this field.