It seems that when we replace a non-leaf node with nothing in a rule, an error is raised rather than removing all the downstream nodes. The hack for now is to create a singleton type and return that object (so we leave a mark of when a part of the graph was removed due to senescence or pruning). We need to figure out which cases are triggering this bug and make sure that it works as intended.
It seems that when we replace a non-leaf node with
nothingin a rule, an error is raised rather than removing all the downstream nodes. The hack for now is to create a singleton type and return that object (so we leave a mark of when a part of the graph was removed due to senescence or pruning). We need to figure out which cases are triggering this bug and make sure that it works as intended.