lihaoyi/Ammonite

Rock-solid shell scripting

Matthew de Detrich

@mdedetrich

for some reason the coursier variable is no longer visible?

ahhh its coursierapi now

Matthew de Detrich

@mdedetrich

Hmm actually I am having a problem here, for some reason Authentication is not visible in the coursier api

Matthew de Detrich

@mdedetrich

In fact it seems to be impossible to add a repository to interp.repositories() that requires authentication?

Brian Maso

@bmaso

I literally had a dream last night of being able to plot out visualizations in a scala REPL (a la a jupyter notebook with matplot). With regard to that, can anyone offer a rough sketch of what modifications would be necessary in Ammonite to support outputting an image as the response to a line of code in Ammonite?

Sören Brunk

@sbrunk

@bmaso perhaps not exactly the same but you can use Ammonite with support for embedded plots in an actual Jupyter notebook via https://almond.sh

Brian Maso

@bmaso

@sbrunk Thanks. I hadn't realized jupyter-scala had advanced somewhat -- last time I had tried to use it it, it was a bit krufty.

Anton Kulaga

@antonkulaga

Stupid question: how one can pre-compile ammonite script without running it?
In my docker container I have a script with one @main() method, I want to precompile when I build docker container so the user will not wait much, but I do not want the main function to be run.

ReiReiRei

@ReiReiRei

Hi all, is it any workaround for lihaoyi/Ammonite#763 ?

Alexandre Archambault

@alexarchambault

@ReiReiRei The workaround described in the issue should work with the latest releases of Ammonite.

ReiReiRei

@ReiReiRei

@alexarchambault the problem is I am using spark 2.3.1/scala 2.11 and almond-spark The highest version for scala 2.11 almond-spark is 0.6.0 and uses old version of ammonite

hmf

@hmf

I have asked this question in the Mill channel but I think it is more of a Ammonite question. How can I get the path of a class define in the a script (defined in an external file script)? I need to add this to a classpath for dynamic class loading. TIA

ReiReiRei

@ReiReiRei

@lihaoyi @alexarchambault so does it mean that ammonite does not support most stable version of spark correctly? beacase spark is officially still on 2.11 scala

Alexandre Archambault

@alexarchambault

@hmf Not 100% sure I understand your question, but repl.sess.frames.flatMap(_.classloader.inMemoryClasses) gives a list of class names / class byte code generated by the repl.

@ReiReiRei I'm afraid so, as it stands currently. 2.11 is getting quite old, 2.12 has been out for almost 3 years now(!)

ReiReiRei

@ReiReiRei

@alexarchambault but a lot of spark libs a based on only 2.11 spark ) hadoop distributions( cloudera hortonworks does not support spark based on scala 2.12 yet)

is it possible to say to old version of ammonite not to add sources to classpath ? something like this ?

interp.resolutionHooks += { fetch =>
  fetch.withClassifiers(fetch.classifiers.filter(_ != "sources"))
}

hmf

@hmf

@alexarchambault Thanks for the feedback. Don't think that will do - I am running a Mill script. I need to get hold of the .class file of an object (extends PostModifier) and place it on the classpath for dynamic loading by another library (MDoc). Is this doable?

trepidacious

@trepidacious

Is it possible to disable the "extra" printing in the ammonite repl (e.g. imports, defining variables etc.)? I'm using it from almond and it's great, but I don't need the extra output, just the results of println and explicitly sending plots etc. to the worksheet.

Crosson David

@dacr

Hi, any idea why I get java.lang.ClassNotFoundException: scala.jdk.CollectionConverters$ issues with amm-2.12-1.7.4 and not with amm-2.13-1.7.4, and not at all within the test (sbt +test) of my project which cross-compiles 2.11, 2.12, 2.13 (https://github.com/dacr/drools-scripting) using scala-collection-compat library.

10:46 $ amm-2.12-1.7.4 drools-hello-world.sc 
Compiling (synthetic)/ammonite/predef/interpBridge.sc
Compiling (synthetic)/ammonite/predef/sourceBridge.sc
Compiling (synthetic)/ammonite/predef/frontEndBridge.sc
Compiling (synthetic)/ammonite/predef/DefaultPredef.sc
Compiling /home/dcr/notes/scripts/drools-hello-world.sc
drools$minushello$minusworld$HelloTest:
Drools
ammonite.$file.drools$minushello$minusworld$HelloTest$ *** ABORTED ***
...
10:48 $ amm-2.13-1.7.4 drools-hello-world.sc 
Compiling (synthetic)/ammonite/predef/interpBridge.sc
Compiling (synthetic)/ammonite/predef/sourceBridge.sc
Compiling (synthetic)/ammonite/predef/frontEndBridge.sc
Compiling (synthetic)/ammonite/predef/DefaultPredef.sc
Compiling /home/dcr/notes/scripts/drools-hello-world.sc
drools$minushello$minusworld$HelloTest:
Drools
- should say hello

I'm not sure it is a bug, if it is I'll fill a dedicated issue, the executed code is the following : https://gist.github.com/dacr/6921d569fd33182da358d6a8e383aa0a :

import $ivy.`org.scalatest::scalatest:3.0.8`
import $ivy.`fr.janalyse::drools-scripting:1.0.1`

import org.scalatest._, org.scalatest.OptionValues._
import fr.janalyse.droolscripting._

object HelloTest extends FlatSpec with Matchers {
  "Drools" should "say hello" in {
    val drl =
      """package testdrools
        |rule "hello" when
        |then
        |  insert("HELLO WORLD");
        |end
        |""".stripMargin
    val engine = DroolsEngine(drl)
    engine.fireAllRules()
    engine.getObjects.headOption.value shouldBe "HELLO WORLD"
  }
}
HelloTest.execute()

May be because ammonite is using an old release of ("org.scala-lang.modules" %% "scala-collection-compat") "2.0.0" while the latest is "2.1.2" ?

Crosson David

@dacr

Yes I've just checked it and it was coming from this old release :)
Pull request sent. lihaoyi/Ammonite#1011

nsram

@nsram_twitter

Just started using Li Haoyi's work. Needed to load jars from a local folder due to limited internet access. The following worked.

@  val jarpath = pwd / 'dev / 'jar  // local folder with jars
@  ls! jarpath | interp.load.cp

Brian Maso

@bmaso

:thumbsup: That's pretty cool

nsram

@nsram_twitter

On ubuntu, I am not seeing the color syntax highlighting for error stack trace (its all red). Colors are working normally otherwise. Amm 1.7.1

Milan Satala

@msatala

Hi guys,
ammonite-sshd doesn't run predef when running inside docker container.
I'm running ammonite embedded in play server. When I start play via runProd and connect to ammonite via ssh I get this

Compiling /home/milan/dev/traceability/traceability/server/playserver/target/universal/stage/(console)
Setting front-end. // My println in predef
Welcome to the Ammonite Repl 1.7.4

However when I run the same play server inside alpine-java docker container I only get this:

Password: 
Welcome to the Ammonite Repl 1.7.4

Any ideas?

Glen Marchesani

@fizzy33

anyone know of a simple way to use ammonite as a server side scripting language ?

Olivier Mélois

@Baccata

@fizzy33 can you refine your question ? ammonite provides you with a way to access a repl over ssh, and you can can embed it in a server : http://ammonite.io/#RemoteREPL , but that might not be what you need

@msatala chances your docker container doesn't see your predef. It's likely a docker problem, not an ammonite problem.

Abhishek Srivastava

@abhishes_twitter

I am having a problem with ammonite with international characters. Try to copy and paste this string into ammonite (on macOS).
大学院小学高校入試高校受験
Now move the cursor to the beginning of the line and try to move the cursor to the end. you will see that the cursor gets stuck at the character 校. Even Cmd+E doesn't work and it doesn't take you to the end of the line.

This doesn't happen in Scala REPL or Bash. only in ammonite it gets stuck

Emmanuel Nhan

@enhan

Hi there, I am using sbt 1.3.2 and I try to use ammonite 1.7.4 with a commandAlias. It works, but the supershell feature of sbt makes it almost unusable. Is there a way to deactivate supershell for just one task ? I tried Test / useSuperShell := falsebut it did not work. Does someone has an alternative workaround than just calling sbt with --supershell=false ?

Abhishek Srivastava

@abhishes_twitter

I had the following two lines in my predef.sc

import coursier.MavenRepository
interp.repositories() ++= Seq(MavenRepository("https://repository.mulesoft.org/nexus/content/repositories/public/"))

These used to work fine. today I upgraded my ammonite to 1.7.4 and now its broken. Error

predef.sc:17: not found: value coursier
import coursier.MavenRepository

Abhishek Srivastava

@abhishes_twitter

OK. I was able to resolve the issue of MavenRepository. The updated code is like this

import coursierapi.MavenRepository
interp.repositories() ++= Seq(
        MavenRepository.of("https://repository.mulesoft.org/nexus/content/repositories/public/")
)

Now it works fine.

Zsolt Szilagyi

@esgott

@abhishes_twitter I have the same issue, but I also need authentication. Couldn't manage to make it work yet.

Natalia Tisseyre

@supperslonic

Is it possible to have 2 ScalaInterpreters in the single JVM process who share same classLoader? I ran into ClassCastException when trying to deserialize/serialize some instances created by a different classLoader: almond-sh/almond#452

Haifeng Li

@haifengl

I develop a small DSL in scala. It works in Scala's REPL. But in Ammonite, I get the following error:

@ val formula = "class" ~ 
cmd2.sc:1: value ~ is not a member of String
val formula = "class" ~
                      ^
Compilation Failed

All my classes and implicit conversation are already imported.

The same expression works in Scala REPL

scala> val formula = "class" ~
formula: smile.data.formula.FormulaBuilder = FormulaBuilder(Some(class),ListBuffer())

Haifeng Li

@haifengl

Never mind. I find the root cause. It is on my end.

phrmoy

@phrmoy

hello all, thanks for such a great package!

I am running into "org.apache.spark.sql.AnalysisException: Unable to generate an encoder for inner class ammonite.$sess.cmd29$Helper$SimpleTuple without access to the scope that this class was defined in.
Try moving this class out of its parent class.;" and google searches are not too helpful at the moment

can anyone help me get through this?

Just mapping values of a dataset to

case class SimpleTuple(id: Int, desc: String) {
    def >(other:SimpleTuple) = this.id > other.id
}

phrmoy

@phrmoy

So here's what I had to do to make this work

case class SimpleTuple4(id: Int, desc: String) {
    def >(other:SimpleTuple4) = this.id > other.id
}
org.apache.spark.sql.catalyst.encoders.OuterScopes.addOuterScope(this)
textFile.map(line => SimpleTuple4(line.split(" ").size,line)).reduce((a, b) => if (a > b) a else b)

all within the same cell

is this the current workaround? This is Spark 2.4.4

Sören Brunk

@sbrunk

@phrmoy yes, on Spark master this is included by default, but it seems it's not part of 2.4. See https://github.com/apache/spark/pull/23607/files#diff-28a2ffae3ceff1bd5646567163c3a00b

phrmoy

@phrmoy

Well, it's great to know that this will be included in Spark at some point. Thanks for sharing.

Li Haoyi

@lihaoyi

@/all Ammonite 1.8.0 is out http://ammonite.io/#1.8.0, with an OS-Lib upgrade

Where communities thrive

People

Repo info

Activity