No extension is an island

[ukemi]

Last Friday while browsing models from the legacy data load on development, we noticed that the model for MGI:98297 had a lot of nodes that represented cells and anatomical structures and appeared to not be connected to any other nodes of the graph. These types of node should only be from annotation extensions and should always have properties connecting them to 'core annotations'.

[ukemi]

Note @dustine32, the model is here:
http://noctua-dev.berkeleybop.org/editor/graph/gomodel:ca36fa49-bf5d-4b2c-90b8-dd1b34014ae6

[dustine32]

@ukemi Interesting. I'll try generating this gene's model through the newest code to see if this is still occurring. Thanks!

[dustine32]

@ukemi Oh wow, we need a new load of models into noctua-dev. I'll start the run tomorrow but just keep yelling at me until it's done!

Looks like this has been handled (since the last load in late Sept 2019!) in the code:

http://68.181.46.18:8910/editor/graph/gomodel:885bcfca-fea7-4240-aee9-e028572b27b9

The GO_REL relation (GOREL:0001002) is even used now.

[ukemi]

Ah. Ok @dustine32 this makes sense to me. Sorry, I wasn't aware of the status on noctua-dev and I thought that when I checked before, things all looked ok. I won't yell, but I may bug you now and then. :)