Hi Minudika, thank you for showing up early and expressing an interest in participating in GSoC this year with Zeppelin project!
Its great that you already have relevant experience creating an interpreter for Zeppelin, building it, etc. If that work is opensource, could you please also share the link here? On the Beam interpreter project - AFAIK there is no such implementation yet, so the objective, or the big goal I would suggest is: to build 2-3 Zeppelin notebooks, each showcasing some data analysis using Beam\Dataflow API over a real dataset, using Flink\Spark executors. I would imagine, that would include multiple steps like: - create plain java interpreter - create Beam interpreter (using plain java one + Beam SDK + executor on top of Flink\Spark, including Interpreter features like "job progress monitoring", "job cancellation" and auto-suggestions in UI) - create example notebooks showcasing data processing uing Beam interpreter, over some real dataset First steps would include a small research on feasibility of such plan of course, and may be a simple CLI PoC (Java REPL, Beam SDK + Flink\Spark executor). What do you think? --- Alex On Sun, Feb 21, 2016, 11:58 Minudika Malshan <minudika...@gmail.com> wrote: > Hi all, > > I am looking forward to contribute in this year GSOC. > I have a basic understand in writing interpreters for zeppelin and have > already implement an interpreter for WSO2 siddhi which is a complex event > processing language. > Could you please give me more details about the requirement for > implementing a new interpreter for Apache beam? > It would be a great help for me to getting familiar with this > implementation. > > Thank you! > > Minudika Malshan > Undergraduate > Department of Computer Science and Engineering > University of Moratuwa. >
