How would I predict the embedding vector of a new, unseen node with hinSAGE?

[Boraly]

I have a big graph with 7 different node types and 3 different edge types. I used hinSAGE to get an embedding of one node type with the other node types as neighbours information. Now I want to predict the embedding of a new, unseen node that has edges to exisiting nodes in the training graph. How would I predict the embedding vector?
Do I have to add the new node into the exisiting graph and use the model on the whole graph? How would I then specifically extract the vector for the new node?

Thank you in advance!