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
{{ message }}
Repository navigation
AtomContactSet uses Pair<Atom> which is directed (atom1 -> atom2) #1153
Atom contacts are saved in AtomContactSet as a Pair without enforcing atom ordering.
This way, a Contact c1 (atom1, atom2) is not considered equal to another Contact c2(atom2, atom1).
The fix is to ensure atom ordering as described in this comment.
However, this fix will impose behaviour change, which needs to be studied thoroughly first.
Thanks for reporting. At least the new docs that went in with #1147 make it clear what the behaviour is now. One thing that is important to consider: in intra-chain contacts the order is defined in a certain way (and that's an issue as reported above), but not in the inter-chain case.
One special case of inter-chain that is worth remembering is inter-chain contacts of chains that are symmetry mates: in there the contacts are not necessarily symmetric (they are only symmetric in isologous interfaces, i.e. with C2 symmetry) but they are identified with same serial numbers and same chain ids.
Activity
josemduarte commentedon Sep 4, 2026
Thanks for reporting. At least the new docs that went in with #1147 make it clear what the behaviour is now. One thing that is important to consider: in intra-chain contacts the order is defined in a certain way (and that's an issue as reported above), but not in the inter-chain case.
One special case of inter-chain that is worth remembering is inter-chain contacts of chains that are symmetry mates: in there the contacts are not necessarily symmetric (they are only symmetric in isologous interfaces, i.e. with C2 symmetry) but they are identified with same serial numbers and same chain ids.
aalhossary commentedon Sep 6, 2026
OK What about adding new method overloads that take a parameter (
directed)?