moradology / geotrellis-workshop-notebooks

Jupyter notebooks and infrastructure for hands-on workshops

Geotrellis Workshop Scala/Spark Notebooks

This repository contains a collection of Jupyter notebooks to assist in the teaching and learning of Scala and Apache Spark. We've also included a docker image to provide an Almond-based Jupyter environment with some preloaded libraries. The preloaded libraries are

• Spark 3.3.1
• Geotrellis 3.6.3
• Rasterframes 0.11.1

Setup

Run the following commands in a terminal on Mac or Linux to create the docker environment:

docker build -t geotrellis-almond docker
docker run --rm -it -p 8888:8888 -p 4040:4040 geotrellis-almond

If you are using your local ports 8888 or 4040, adjust the -p arguments to the above to be, e.g., -p 9999:8888 to redirect the container's internal port 8888 to your local machine's port 9999.

Upon running the above, a Jupyter splash screen will display in the terminal, and somewhere in that text, you will see a URL of the form

https://127.0.0.1:8888/lab?token=<hexidecimal string>

Copy and paste that into your browser's address bar. If you had to use a different port number instead of 8888, adjust the URL to point to the selected port.

Imports

This Almond-based notebook is built on top of an Ammonite interpreter. Due to the lack of a build environment like sbt, we must rely on the Ammonite import mechanisms to ensure that jars are in place. Specifically, there is special Ivy import syntax that must be used. Also remember that it is both necessary to import the package from Ivy and import the library code itself:

import $ivy.`org.locationtech.geotrellis::geotrellis-vector:3.6.3`
import geotrellis.vector._

Generating a SparkSession

Here is the basic code template for creating a SparkSession:

import $ivy.`org.apache.logging.log4j:log4j-core:2.17.0`
import $ivy.`org.apache.logging.log4j:log4j-1.2-api:2.17.0`
import $ivy.`org.apache.spark::spark-sql:3.3.1`

import org.apache.log4j.{Level, Logger}
Logger.getLogger("org").setLevel(Level.OFF)

import org.apache.spark.sql._

val spark = {
  NotebookSparkSession.builder()
    .master("local[*]")
    .getOrCreate()
}

Note that this uses a custom SparkSession builder, but that builder largely has the same interface as the customary builder.