Closed bannanc closed 5 years ago
@jmaat did a quick review on this she said over slack:
A couple comments: Your logic looks sound within the code. The Doc string was thorough and easy to understand, and the code was very clean. Hope this helps!
So I'm going to go ahead and merge this.
Initially I copied the "scoring function" from smirky where a set of SMIRKS are scored based on how they type a list of molecules compared to a reference (in this case user specified clusters).
However, I realized this is sort of a waste of time. In SMIRKY we wanted scores when the SMIRKS patterns didn't perfectly match the reference. Here however, we only care about SMIRKS either matching or not matching the reference.
Therefore, I've replaced the "scoring" with a criteria where if a current type matches more than one reference type (or vice versa) then the assignments do not match.
This is the same criteria
smirksify
was using before, but before it was checking for a score = 1.0 and now it just checks for "do these match".