Do we think there is any reasonable way to add SQL Interfaces to GraphFrames algorithms?

I have some experience wrapping dataframes functionality as sql functions but I think ,for graphframes, the parser would probably need modification as well to support some new kind of Binary Node and a syntax to describe the graph since it is comprised of two tables.

Some prior work I've done on Sedona: https://github.com/apache/sedona/blob/master/spark/common/src/main/scala/org/apache/spark/sql/sedona_sql/expressions/PhysicalFunction.scala#L98