Neptune PageRank calculation

0

Im working in a Jupyter Notebook and trying to run the pageRank calculation on a substantial graph, however, this will not run due to "Graph does not support graph computer", from research i can see that Amazon Neptune does not currently support the Apache TinkerPop GraphComputer interface. What are my alternatives?

BArcher
asked 2 years ago608 views
1 Answer
0

Amazon Neptune supports the OLTP implementation of Apache TinkerPop. Use of TinkerPop OLAP functions that require withComputer() are not supported (1). If looking to run graph algorithms on top of existing data within Neptune, you can leverage Apache Spark GraphFrames (2). You'll need to extract the data from Neptune to use GraphFrames and this can be done using the recently published Pandas integration (3). A common architecture is to store the PageRank scores back into the graph as a property on the related vertices. As the graph changes, a batch job would be required to extract the data from Neptune and re-run the PageRank calculations as needed.

(1) https://docs.aws.amazon.com/neptune/latest/userguide/gremlin-step-support.html#neptune-gremlin-steps-unsupported (2) https://towardsdatascience.com/graphframes-in-jupyter-a-practical-guide-9b3b346cebc5 (3) https://github.com/awslabs/aws-data-wrangler/blob/main/tutorials/033%20-%20Amazon%20Neptune.ipynb

profile pictureAWS
answered 2 years ago

You are not logged in. Log in to post an answer.

A good answer clearly answers the question and provides constructive feedback and encourages professional growth in the question asker.

Guidelines for Answering Questions