[Open Babel] Re: Modified bond order perception

[Geoff H. <ghu...@ws...>]

Hello again,

You could probably say I've been AWOL from development on OB. I have a 
pile of things on my desk, but today I was waiting for some supplies to 
arrive and had a bit of free time.

I think the current code should be more reliable at bond-typing:
* I added back a pass that checks rings if they're "flat" and thus types 
anything with 2 or 3 single bonds as probably sp2. This helps with 5 and 
6-membered aromatic systems with odd bond angles (e.g. pyrrole)
* I fixed a few small bugs in the assignment phase -- it ensures we 
don't "go over" the maximum bond valence for an element.
* I added a few functional group recognition bits.
* It won't add a multiple bond to a nitrogen if that would make it 
tetravalent.

It still doesn't have an aromatic perception phase, which is the biggest 
hurdle, but I haven't had the time to finish that.

In the meantime, I see quite a bit of improvement. Please let me know if 
I've (re)introduced bugs as far as PDB reading. I'm hoping this will be 
good enough for a release and I can get the last bit finished over the 
holidays.

-Geoff