You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
While it does its job, it became bloated and slow. Much of this boils down to the fact that special cases are handled all across the code.
what should be changed
Porphyrin handling is the main pain point, mostly because we start fragmentations on metals (and not on something like a cycle-based criterion). If we do not do this handling, the net-embedding won't be meaningful, also the building blocks won't be meaningful.
Move some immutable properties to the graph and do not keep track of them through the building blocks. For instance, branching indices are - to our current definition - a property of the graph that is shared between both building blocks. Instead of keeping track of them in both building blocks, we might just leave them on the original MOF
Make it easier to customize the clustering approach -> make it also work for COFs
Openbabel dependency -> Mostly to extract SMILES based on the graph, we should be able to implement this ourselves
Consider switching to graphtool or the rust variant of networkx
what should remain the same
Having the collections of SBU molecules works great
The text was updated successfully, but these errors were encountered:
While it does its job, it became bloated and slow. Much of this boils down to the fact that special cases are handled all across the code.
what should be changed
what should remain the same
The text was updated successfully, but these errors were encountered: