Offline support

Packaging up your entire application for offline deployment

You can use ml-gradle to package up your application such that it can be deployed using ml-gradle in an offline or "disconnected" environment.

For an example of this, please see the build.gradle file in the disconnected example project.

Packaging up ml-gradle for offline usage

If you'd like to use ml-gradle in an environment where you can't connect to repositories like jcenter and mavenCentral, you can copy some of what's in the disconnected-project build.gradle file to package up a zip of ml-gradle and its dependencies and then reference those from the filesystem in your build.gradle file. Alternatively, you can build an "uber" jar - see the next section below.

To do this, you'll need to be in a "connected" environment to begin so you can download ml-gradle dependencies. You'll also need access to Gradle and a build.gradle file to edit.

First, let's declare ml-gradle as a regular application dependency instead of a buildscript dependency (the version number of ml-gradle doesn't matter, this will work with any version):

repositories {
  mavenCentral()
}
configurations {
  mlgradle
}
dependencies {
  mlgradle "com.marklogic:ml-gradle:4.5.0"
}

Next, let's make a task that copies ml-gradle and all of its dependencies to a directory (any directory is fine here):

task copyMlGradleDependencies(type: Copy) {
  from configurations.mlgradle
  into "build/ml-gradle-dependencies"
}

You now have everything you need to use ml-gradle in an offline environment. You'll probably want to package these jars up into a zip/tar and copy it to your offline environment - you can of course use Gradle for that too (see the docs on Zip for more info - note that this will create the file under build/distributions by default):

task buildMlGradleZip (type: Zip, dependsOn: ["copyMlGradleDependencies"]) {
  from "build/ml-gradle-dependencies"
  archiveName "ml-gradle-package.zip"
}

Once you have your ml-gradle dependencies in your offline environment, you'll then apply the Gradle plugin in a slightly different manner (this is shown in the disconnected-project above) - assuming we've copied our jars to a directory named "ml-gradle-dependencies":

buildscript {
  dependencies {
    classpath fileTree(dir: "ml-gradle-dependencies", include: "*.jar")
  }
}
apply plugin: "com.marklogic.ml-gradle"

You should be in business now - no repositories needed, you have everything you need for ml-gradle in that directory of jars.

If you're using MarkLogic's Data Hub Framework, then this gist may be helpful to you as well.

Creating an uber jar of ml-gradle

If you'd prefer to have a single jar containing all ml-gradle dependency classes, as opposed to a zip or tar of all the dependency jars, toss the following into a build.gradle file:

plugins {
	id "java"
}

repositories {
	mavenCentral()
}

dependencies {
	compile "com.marklogic:ml-gradle:4.5.0"
}

jar {
	from {
		configurations.compile.collect { it.isDirectory() ? it : zipTree(it) }
	}
}

Then run:

gradle jar

This will produce an uber jar under ./build/libs. You can then copy that to another project and reference it in that project's build.gradle file:

buildscript {
	dependencies {
		classpath fileTree(dir: "path/to/uber-jar-directory", include: "*.jar")
	}
}
apply plugin: "com.marklogic.ml-gradle"