Does nebula graph support random walk?
Not for now.
Nebula doesn’t support algorithms yet. If you want to run algorithm, you can use spark-connector to read nebula data to DataFrame, and then call Spark MLLib’s algorithm.
sparkSession = SparkSession.builder() .appName("appName") .setMaster("xxx") .getOrCreate() df = sparkSession.read .nebula("127.0.0.1:45500", "spaceName", "100") .loadEdgesToDF("edge", "*")
How to call Spark MLLib’s algorithm , please see spark’s document.
I’d +1 the usefulness of a random walk ability.
As the number of vertices go to billions, it becomes impractical to employ neural network embeddings.
There are approximations based on selecting only the vertices we care about and embedding the distribution of randomly sampled reachable nodes around each.
You can use Spark MLLIB’s random walk algorithm.