I have a multi-project build.sbt file. I would like to assemble the jar for just one of the projects. Currently, I do the following:
$ sbt
project analysis
assembly
...
exit
I would like to save a few steps and assemble the jar for the project "analysis" from the command line. Is there a way to do this?
Thanks.
You can use sbt without its REPL:
$ sbt analysis/assembly
Related
is there a way to build/compile all configurations at once? I have a project that has a Dev configuration in addition to the default Compile and Test configuration, and i am looking for a command or a setting in my build.sbt that would allow me to compile/package all 3 configurations at once.
Basically looking for a way to avoid having to do these 3 commands to build the entire source tree:
sbt compile
sbt dev:compile
sbt test:compile
When I use sbt from IntelliJ it is able to do this on building the project, but I am looking to do this from the command line.
First, you can run multiple tasks with a single sbt invocation:
sbt compile dev:compile test:compile
Second, you could define an alias in your build which does what you want:
addCommandAlias("compileAll", "; compile; dev:compile; test:compile")
Then, just run sbt compileAll.
I want to build subproject in Spark with sbt. I found this example and it works
$ ./build/sbt -Phive -Phive-thriftserver (build)
sbt (spark)> project hive (switch to subproject)
sbt (hive)> testOnly *.HiveQuerySuite -- -t foo ( run test case)
However, I tried the following but it does not build but quit
./build/sbt -mllib
I do not know how does the author figure out -Phive -Phive-thriftserver. I cannot find this in Spark source code.
I just want to do the exact same thing as the example but with a different subproject.
This is not asking how to use projects to print out all available projects.
Specify the project scope:
./build/sbt mllib/compile
refer to: http://www.scala-sbt.org/0.13/docs/Scopes.html
I have a self-contained application in SBT. My data is stored on HDFS (the hadoop file system).How can I get a jar file to run my work on another machine.
The directory of my project is the following:
/MyProject
/target
/scala-2.11
/MyApp_2.11-1.0.jar
/src
/main
/scala
If you don't have any dependencies then running sbt package will create a jar will all your code.
You can then run your Spark app as:
$SPARK_HOME/bin/spark-submit --name "an-app" my-app.jar
If your project has external dependencies (other than spark itself; if it's just Spark or any of it's dependencies, then the above approach still works), then you have two options:
1) Use the sbt assembly plugin to create an uper jar with your entire class-path. Running sbt assembly will create another jar which you can use in the same way as before.
2) If you only have very few simple dependecies (say just joda-time), then you can simply include them into your spark-submit script.
$SPARK_HOME/bin/spark-submit --name "an-app" --packages "joda-time:joda-time:2.9.6" my-app.jar
Unlike Java, in Scala, the file’s package name doesn’t have to match the directory name. In fact, for simple tests like this,
you can place this file in the root directory of your SBT project, if you prefer.
From the root directory of the project, you can compile the project:
$ sbt compile
Run the project:
$ sbt run
Package the project:
$ sbt package
Here is link to understand:
http://alvinalexander.com/scala/sbt-how-to-compile-run-package-scala-project
Is there a way to create (generate) maven project from existing scala sbt project via some script, e.g. automatically?
Execute make-pom from sbt:
sbt make-pom
You'll find the pom in target/scala-2.11/
Since most IDEs are only able to import Maven projects, I'd like to generate a POM.xml from an SBT managed project, is there a better way to do it?
Did you try make-pom from sbt?
It generates basic POM for the current project at the ./target.
To customize generation, you can override pomExtra, pomIncludeRepository and pomPostProcess at the project definition.
There is a very direct way provided by SBT. You can use the below command where your SBT file exixts:
sbt makePom
This will generate the .pom file in the target folder you can search that and rename to pom.xml and keep that file in the location and run mvn clean compile install to get full out of it.
Vasil's answer is correct, but for Eclipse and IDEA you can generate IDE metadata more directly using plugins. For IDEA https://github.com/mpeltonen/sbt-idea, and for Eclipse https://github.com/musk/SbtEclipsify .