Use more efficient KNN methods for link prediction

We should implement a proper link prediction inference script. Users can provide a list of nodes and the link prediction inference script should return top K nodes that are likely to connect with the input nodes. We should use FAISS for link prediction.

No file, test, or entry point is named in the issue, so first locate any existing inference scripts and determine how link prediction is represented. Define the FAISS-based inference flow around a supplied list of nodes. Done means the script accepts those nodes and returns the top K nodes most likely to connect to them.

Written by the indexing model from the issue text.