Ensure correct stereo consistency (Fix #812) #813

johnmay · 2022-02-03T10:41:08Z

Fixes #812. This was quite a tricky one and very subtle. The minimal example is */C=C/C>> |$R$|, essentially due to AtomContainer2 AtomRef/BondRef the stereo gets in an inconsistent state. To handle these these implementations now remap stereo chemistry to their internal references when adding stereo.

A side-effect is addStereo() is now more robust and so some tests needed updating. There was also a slight miss-use of the connected components in SDG (my own doing) to split a single molecule. In this case we need to check if we split in the middle of stereochemistry. This also exposes another issue with Sgroups which I will open an issue for.

…s added to a molecule it's atom/bond components are remapped to the internal AtomRef/BondRef of the molecule in question. This ensure any future actions such as swapping and atom for another work correctly. This is now more robust and we get an error if we try to add stereochemistry that includes atom/bonds not in this molecule - this makes sense.

… function in the structure diagram generator. This is a bit of a missue but reasonable - we now need check if the stereochemistry has not be split between two molecules. We should also do the same of Sgroups (issue openned).