2

I know databricks offers the possibility to simply convert notebooks into "production-grade" rest APIs.

Is there a similar functionality for open source notebooks like Zeppelin, Scala-Notebook or Jupiter Notebook or hue-notebook? It would be great if the solution would support sparkR.

Thanks a lot

Thomas K
  • 39,200
  • 7
  • 84
  • 86
Georg Heiler
  • 16,916
  • 36
  • 162
  • 292

1 Answers1

2

Perhaps you're looking for a Spark REST server solutions, like:

?

Tagar
  • 13,911
  • 6
  • 95
  • 110
  • Yes and no. Currently we are working to implement our models using spark-jobserver. However, it does not "talk" R and still our notebooks need to be re-written. So some sort of automatic solution would be favourable. – Georg Heiler Apr 08 '16 at 06:54
  • Got it. Livy recently added R support, see https://github.com/cloudera/livy/pull/96 It might be only in master branch though. – Tagar Apr 08 '16 at 15:34
  • I would prefer some integrated approach like https://www.lightbend.com/blog/scala-and-spark-notebook-the-next-generation-data-science-toolkit – Georg Heiler Apr 11 '16 at 06:49
  • do you know if livy is the only option which supports sparkR? – Georg Heiler Apr 19 '16 at 19:17
  • Yep, I think so.. You'll have to use Livy from master branch thought to have SparkR. – Tagar Apr 19 '16 at 19:39