Closed sstemann closed 3 months ago
@sstemann @MarkDWilliams I looked into this and it seems to be a result of the ARS merging because the aux graph aux_graph_420_creative_DTD_option_group_0
cannot be found in the auxiliary_graphs
field. Looking at the subgraph it seems there are several very general nodes which may be blocked. According to the ARS logic this would remove the entire aux graph.
Yes, the current behavior is to remove the whole support graph if it contains blocked nodes to avoid sending possibly disjointed graphs after removing specific nodes. However, I can change this behavior to selectively remove nodes and edges and leave the rest of the graph.
@MarkDWilliams we should coordinate to run a test for the behavior you described. Please reach out when you have some time.
i can't retest this while NCATSTranslator/Relay#681 is happening
i cant replicate this, i'm seeing all of the ARAX support in their inferred edges in the UI. so i'm closing this.
I don't know where this issue is happening at but here are the steps:
In Test UI > What drugs may treat > multiple sclerosis: https://arax.ncats.io/beta/?r=157f5fc2-1748-459f-b381-724d0c353ab9
View result for Teriflunomide - three "lookup paths"
View this same result in the ARAX GUI: https://arax.ncats.io/beta/?r=157f5fc2-1748-459f-b381-724d0c353ab9
One treats edge is [sg] which has an edge support graph - this graph has a predicate "predisposes" which i think might be updated to "predisposes_to_condition" in Biolink Model - so maybe this is an obsolete issue? NCATSTranslator/Feedback#330