load.ivy through HTTP/HTTPS proxy for Spark setup

[rvernica]

I am following the Spark setup instructions listed in Spark.ipynb. How can I configure load.ivy to go through an HTTP/HTTPS proxy for donwloading?

Thanks!

[rvernica]

If I set the HTTP/HTTPS proxy in SBT_OPTS before starting Ammonite, the the load.ivy call works from Ammonite:

SBT_OPTS="-Dhttp.proxyHost=... -Dhttp.proxyPort=... -Dhttps.proxyHost=... -Dhttps.proxyPort=..." ./amm
Loading Ammonite Repl...
@ load.ivy("com.github.alexarchambault" % "ammonite-spark_1.3_2.10.5" % "0.3.1-S  NAPSHOT")

@ 

[alexarchambault]

(Initial answer was buggy, now fixed)

You can try editing ~/.ipython/kernels/scala211/kernel.json, and add the options there:

{
  "language" : "scala211",
  "display_name" : "Scala 2.11",
  "argv" : [
    "/path/to/jupyter-scala_2.11.6-0.2.0-SNAPSHOT/bin/jupyter-scala",
    "--quiet",
    "--connection-file",
    "{connection_file}"
  ],
  "env": {
    "JVM_OPT": "-Dhttp.proxyHost=... -Dhttp.proxyPort=... -Dhttps.proxyHost=... -Dhttps.proxyPort=..."
  }
}

I haven't tested these exact options myself, but that's how one can provide java options to the kernel.

[alexarchambault]

Alternatively, you can try to edit the jupyter-scala script directly (its path is in kernel.json above), and add the options at the end, close to the current java options (like -D...).

[rvernica]

Thanks! It worked.

[cequencer]

I noticed this solution worked 3 years ago, but now the jupyter-scala is using couriser to download all the runtime dependencies and artifacts. I have tried the proposed method here, but it no longer works with couriser. I have provided more attempts and clues in issue #189. It will be great if @alexarchambault can revisit this either issue with the new codebase using couriser. Many thanks and sorry for trolling an closed issue from the past.