Interchain contacts

[gezmi]

Hi,

very nice tool, thanks for doing this.

There is one thing, that I missed from the implementation: AF3 (and AF2) masks interchain contacts from the template, not taking into account relative orientations. With AF2, we explored this option and saw that enabling it (for peptide-protein complexes) was needed to be able to score them. I am curious how you handled this problem, as as far as I understood, the interchain contacts are still masked in AF3score.

[gezmi]

I created a pull request that helps with this