Develop and take a look at AWS Glue 5.0 jobs domestically utilizing a Docker container

AWS Glue is a serverless knowledge integration service that means that you can course of and combine knowledge coming via completely different knowledge sources at scale. AWS Glue 5.0, the most recent model of AWS Glue for Apache Spark jobs, offers a performance-optimized Apache Spark 3.5 runtime expertise for batch and stream processing. With AWS Glue 5.0, you get improved efficiency, enhanced safety, help for the following era of Amazon SageMaker, and extra. AWS Glue 5.0 lets you develop, run, and scale your knowledge integration workloads and get insights quicker.

AWS Glue accommodates varied growth preferences via a number of job creation approaches. For builders preferring direct coding, Python or Scala growth is offered utilizing the AWS Glue ETL library.

Constructing production-ready knowledge platforms requires strong growth processes and steady integration and supply (CI/CD) pipelines. To help various growth wants—whether or not on native machines, Docker containers on Amazon Elastic Compute Cloud (Amazon EC2), or different environments—AWS offers an official AWS Glue Docker picture via the Amazon ECR Public Gallery. The picture permits builders to work effectively of their most popular surroundings whereas utilizing the AWS Glue ETL library.

On this publish, we present tips on how to develop and take a look at AWS Glue 5.0 jobs domestically utilizing a Docker container. This publish is an up to date model of the publish Develop and take a look at AWS Glue model 3.0 and 4.0 jobs domestically utilizing a Docker container, and makes use of AWS Glue 5.0 .

Accessible Docker pictures

The next Docker pictures can be found for the Amazon ECR Public Gallery:

• AWS Glue model 5.0 – ecr.aws/glue/aws-glue-libs:5
  

AWS Glue Docker pictures are suitable with each x86_64 and arm64.

On this publish, we use public.ecr.aws/glue/aws-glue-libs:5 and run the container on an area machine (Mac, Home windows, or Linux). This container picture has been examined for AWS Glue 5.0 Spark jobs. The picture incorporates the next:

To arrange your container, you pull the picture from the ECR Public Gallery after which run the container. We display tips on how to run your container with the next strategies, relying in your necessities:

• spark-submit
• REPL shell (pyspark)
• pytest
• Visible Studio Code

Stipulations

Earlier than you begin, make it possible for Docker is put in and the Docker daemon is working. For set up directions, see the Docker documentation for Mac, Home windows, or Linux. Additionally just remember to have at the least 7 GB of disk area for the picture on the host working Docker.

Configure AWS credentials

To allow AWS API calls from the container, arrange your AWS credentials with the next steps:

1. Create an AWS named profile.
2. Open cmd on Home windows or a terminal on Mac/Linux, and run the next command:

PROFILE_NAME="profile_name"

Within the following sections, we use this AWS named profile.

Pull the picture from the ECR Public Gallery

In the event you’re working Docker on Home windows, select the Docker icon (right-click) and select Swap to Linux containers earlier than pulling the picture.

Run the next command to tug the picture from the ECR Public Gallery:

docker pull public.ecr.aws/glue/aws-glue-libs:5

Run the container

Now you’ll be able to run a container utilizing this picture. You may select any of following strategies primarily based in your necessities.

spark-submit

You may run an AWS Glue job script by working the spark-submit command on the container.

Write your job script (pattern.py within the following instance) and reserve it underneath the /local_path_to_workspace/src/ listing utilizing the next instructions:

$ WORKSPACE_LOCATION=/local_path_to_workspace
$ SCRIPT_FILE_NAME=pattern.py
$ mkdir -p ${WORKSPACE_LOCATION}/src
$ vim ${WORKSPACE_LOCATION}/src/${SCRIPT_FILE_NAME}

These variables are used within the following docker run command. The pattern code (pattern.py) used within the spark-submit command is included within the appendix on the finish of this publish.

Run the next command to run the spark-submit command on the container to submit a brand new Spark software:

$ docker run -it --rm 
    -v ~/.aws:/residence/hadoop/.aws 
    -v $WORKSPACE_LOCATION:/residence/hadoop/workspace/ 
    -e AWS_PROFILE=$PROFILE_NAME 
    --name glue5_spark_submit 
    public.ecr.aws/glue/aws-glue-libs:5 
    spark-submit /residence/hadoop/workspace/src/$SCRIPT_FILE_NAME

REPL shell (pyspark)

You may run a REPL (read-eval-print loop) shell for interactive growth. Run the next command to run the pyspark command on the container to begin the REPL shell:

$ docker run -it --rm 
    -v ~/.aws:/residence/hadoop/.aws 
    -e AWS_PROFILE=$PROFILE_NAME 
    --name glue5_pyspark 
    public.ecr.aws/glue/aws-glue-libs:5 
    pyspark

You will notice following output:

Python 3.11.6 (primary, Jan  9 2025, 00:00:00) [GCC 11.4.1 20230605 (Red Hat 11.4.1-2)] on linux
Sort "assist", "copyright", "credit" or "license" for extra info.
Setting default log stage to "WARN".
To regulate logging stage use sc.setLogLevel(newLevel). For SparkR, use setLogLevel(newLevel).
Welcome to
      ____              __
     / __/__  ___ _____/ /__
    _ / _ / _ `/ __/  '_/
   /__ / .__/_,_/_/ /_/_   model 3.5.2-amzn-1
      /_/

Utilizing Python model 3.11.6 (primary, Jan  9 2025 00:00:00)
Spark context Net UI accessible at None
Spark context accessible as 'sc' (grasp = native[*], app id = local-1740643079929).
SparkSession accessible as 'spark'.
>>> 

With this REPL shell, you’ll be able to code and take a look at interactively.

pytest

For unit testing, you need to use pytest for AWS Glue Spark job scripts.

Run the next instructions for preparation:

$ WORKSPACE_LOCATION=/local_path_to_workspace
$ SCRIPT_FILE_NAME=pattern.py
$ UNIT_TEST_FILE_NAME=test_sample.py
$ mkdir -p ${WORKSPACE_LOCATION}/exams
$ vim ${WORKSPACE_LOCATION}/exams/${UNIT_TEST_FILE_NAME}

Now let’s invoke pytest utilizing docker run:

$ docker run -i --rm 
    -v ~/.aws:/residence/hadoop/.aws 
    -v $WORKSPACE_LOCATION:/residence/hadoop/workspace/ 
    --workdir /residence/hadoop/workspace 
    -e AWS_PROFILE=$PROFILE_NAME 
    --name glue5_pytest 
    public.ecr.aws/glue/aws-glue-libs:5 
    -c "python3 -m pytest --disable-warnings"

When pytest finishes executing unit exams, your output will look one thing like the next:

============================= take a look at session begins ==============================
platform linux -- Python 3.11.6, pytest-8.3.4, pluggy-1.5.0
rootdir: /residence/hadoop/workspace
plugins: integration-mark-0.2.0
collected 1 merchandise

exams/test_sample.py .                                                   [100%]

======================== 1 handed, 1 warning in 34.28s =========================

Visible Studio Code

To arrange the container with Visible Studio Code, full the next steps:

1. Set up Visible Studio Code.
2. Set up Python.
3. Set up Dev Containers.
4. Open the workspace folder in Visible Studio Code.
5. Press Ctrl+Shift+P (Home windows/Linux) or Cmd+Shift+P (Mac).
6. Enter Preferences: Open Workspace Settings (JSON).
7. Press Enter.
8. Enter following JSON and reserve it:

{
    "python.defaultInterpreterPath": "/usr/bin/python3.11",
    "python.evaluation.extraPaths": [
        "/usr/lib/spark/python/lib/py4j-0.10.9.7-src.zip:/usr/lib/spark/python/:/usr/lib/spark/python/lib/",
    ]
}

Now you’re able to arrange the container.

9. Run the Docker container:

$ docker run -it --rm 
    -v ~/.aws:/residence/hadoop/.aws 
    -v $WORKSPACE_LOCATION:/residence/hadoop/workspace/ 
    -e AWS_PROFILE=$PROFILE_NAME 
    --name glue5_pyspark 
    public.ecr.aws/glue/aws-glue-libs:5 
    pyspark

10. Begin Visible Studio Code.
11. Select Distant Explorer within the navigation pane.
12. Select the container ecr.aws/glue/aws-glue-libs:5 (right-click) and select Connect in Present Window.

13. If the next dialog seems, select Obtained it.

14. Open /residence/hadoop/workspace/.

15. Create an AWS Glue PySpark script and select Run.

You must see the profitable run on the AWS Glue PySpark script.

Modifications between the AWS Glue 4.0 and AWS Glue 5.0 Docker picture

The next are main modifications between the AWS Glue 4.0 and Glue 5.0 Docker picture:

• In AWS Glue 5.0, there’s a single container picture for each batch and streaming jobs. This differs from AWS Glue 4.0, the place there was one picture for batch and one other for streaming.
• In AWS Glue 5.0, the default person identify of the container is hadoop. In AWS Glue 4.0, the default person identify was glue_user.
• In AWS Glue 5.0, a number of extra libraries, together with JupyterLab and Livy, have been faraway from the picture. You may manually set up them.
• In AWS Glue 5.0, all of Iceberg, Hudi, and Delta libraries are pre-loaded by default, and the surroundings variable DATALAKE_FORMATS is now not wanted. Till AWS Glue 4.0, the surroundings variable DATALAKE_FORMATS was used to specify whether or not the particular desk format is loaded.

The previous record is restricted to the Docker picture. To be taught extra about AWS Glue 5.0 updates, see Introducing AWS Glue 5.0 for Apache Spark and Migrating AWS Glue for Spark jobs to AWS Glue model 5.0.

Issues

Take into account that the next options aren’t supported when utilizing the AWS Glue container picture to develop job scripts domestically:

Conclusion

On this publish, we explored how the AWS Glue 5.0 Docker pictures present a versatile basis for growing and testing AWS Glue job scripts in your most popular surroundings. These pictures, available within the Amazon ECR Public Gallery, streamline the event course of by providing a constant, moveable surroundings for AWS Glue growth.

To be taught extra about tips on how to construct end-to-end growth pipeline, see Finish-to-end growth lifecycle for knowledge engineers to construct an information integration pipeline utilizing AWS Glue. We encourage you to discover these capabilities and share your experiences with the AWS group.

Appendix A: AWS Glue job pattern codes for testing

This appendix introduces three completely different scripts as AWS Glue job pattern codes for testing functions. You should use any of them within the tutorial.

The next pattern.py code makes use of the AWS Glue ETL library with an Amazon Easy Storage Service (Amazon S3) API name. The code requires Amazon S3 permissions in AWS Identification and Entry Administration (IAM). You should grant the IAM-managed coverage arn:aws:iam::aws:coverage/AmazonS3ReadOnlyAccess or IAM customized coverage that means that you can make ListBucket and GetObject API requires the S3 path.

import sys
from pyspark.context import SparkContext
from awsglue.context import GlueContext
from awsglue.job import Job
from awsglue.utils import getResolvedOptions


class GluePythonSampleTest:
    def __init__(self):
        params = []
        if '--JOB_NAME' in sys.argv:
            params.append('JOB_NAME')
        args = getResolvedOptions(sys.argv, params)

        self.context = GlueContext(SparkContext.getOrCreate())
        self.job = Job(self.context)

        if 'JOB_NAME' in args:
            jobname = args['JOB_NAME']
        else:
            jobname = "take a look at"
        self.job.init(jobname, args)

    def run(self):
        dyf = read_json(self.context, "s3://awsglue-datasets/examples/us-legislators/all/individuals.json")
        dyf.printSchema()

        self.job.commit()


def read_json(glue_context, path):
    dynamicframe = glue_context.create_dynamic_frame.from_options(
        connection_type="s3",
        connection_options={
            'paths': [path],
            'recurse': True
        },
        format="json"
    )
    return dynamicframe


if __name__ == '__main__':
    GluePythonSampleTest().run()

The next test_sample.py code is a pattern for a unit take a look at of pattern.py:

import pytest
from pyspark.context import SparkContext
from awsglue.context import GlueContext
from awsglue.job import Job
from awsglue.utils import getResolvedOptions
import sys
from src import pattern


@pytest.fixture(scope="module", autouse=True)
def glue_context():
    sys.argv.append('--JOB_NAME')
    sys.argv.append('test_count')

    args = getResolvedOptions(sys.argv, ['JOB_NAME'])
    context = GlueContext(SparkContext.getOrCreate())
    job = Job(context)
    job.init(args['JOB_NAME'], args)

Appendix B: Including JDBC drivers and Java libraries

So as to add a JDBC driver not at present accessible within the container, you’ll be able to create a brand new listing underneath your workspace with the JAR recordsdata you want and mount the listing to /decide/spark/jars/ within the docker run command. JAR recordsdata discovered underneath /decide/spark/jars/ inside the container are mechanically added to Spark Classpath and might be accessible to be used throughout the job run.

For instance, you need to use the next docker run command so as to add JDBC driver jars to a PySpark REPL shell:

$ docker run -it --rm 
    -v ~/.aws:/residence/hadoop/.aws 
    -v $WORKSPACE_LOCATION:/residence/hadoop/workspace/ 
    -v $WORKSPACE_LOCATION/jars/:/decide/spark/jars/ 
    --workdir /residence/hadoop/workspace 
    -e AWS_PROFILE=$PROFILE_NAME 
    --name glue5_jdbc 
    public.ecr.aws/glue/aws-glue-libs:5 
    pyspark

As highlighted earlier, the customJdbcDriverS3Path connection choice can’t be used to import a customized JDBC driver from Amazon S3 in AWS Glue container pictures.

Appendix C: Including Livy and JupyterLab

The AWS Glue 5.0 container picture doesn’t have Livy put in by default. You may create a brand new container picture extending the AWS Glue 5.0 container picture as the bottom. The next Dockerfile demonstrates how one can prolong the Docker picture to incorporate extra parts it’s essential to improve your growth and testing expertise.

To get began, create a listing in your workstation and place the Dockerfile.livy_jupyter file within the listing:

$ mkdir -p $WORKSPACE_LOCATION/jupyterlab/
$ cd $WORKSPACE_LOCATION/jupyterlab/
$ vim Dockerfile.livy_jupyter

The next code is Dockerfile.livy_jupyter:

FROM public.ecr.aws/glue/aws-glue-libs:5 AS glue-base

ENV LIVY_SERVER_JAVA_OPTS="--add-opens java.base/java.lang.invoke=ALL-UNNAMED --add-opens=java.base/java.nio=ALL-UNNAMED --add-opens=java.base/solar.nio.ch=ALL-UNNAMED --add-opens=java.base/solar.nio.cs=ALL-UNNAMED --add-opens=java.base/java.util.concurrent.atomic=ALL-UNNAMED"

# Obtain Livy
ADD --chown=hadoop:hadoop https://dlcdn.apache.org/incubator/livy/0.8.0-incubating/apache-livy-0.8.0-incubating_2.12-bin.zip ./

# Set up and configure Livy
RUN unzip apache-livy-0.8.0-incubating_2.12-bin.zip && 
rm apache-livy-0.8.0-incubating_2.12-bin.zip && 
mv apache-livy-0.8.0-incubating_2.12-bin livy && 
mkdir -p livy/logs && 
cat > livy/conf/livy.conf
livy.server.host = 0.0.0.0
livy.server.port = 8998
livy.spark.grasp = native
livy.repl.enable-hive-context = true
livy.spark.scala-version = 2.12
EOF && 
cat > livy/conf/log4j.properties
log4j.rootCategory=INFO,console
log4j.appender.console=org.apache.log4j.ConsoleAppender
log4j.appender.console.goal=System.err
log4j.appender.console.format=org.apache.log4j.PatternLayout
log4j.appender.console.format.ConversionPattern=%d{yy/MM/dd HH:mm:ss} %p %c{1}: %mpercentn
log4j.logger.org.eclipse.jetty=WARN
EOF

# Switching to root person quickly to put in dev dependency packages
USER root 
RUN dnf replace -y && dnf set up -y krb5-devel gcc python3.11-devel
USER hadoop

# Set up SparkMagic and JupyterLab
RUN export PATH=$HOME/.native/bin:$HOME/livy/bin/:$PATH && 
printf "numpy /tmp/constraint.txt && 
pip3.11 --no-cache-dir set up --constraint /tmp/constraint.txt --user pytest boto==2.49.0 jupyterlab==3.6.8 IPython==7.14.0 ipykernel==5.5.6 ipywidgets==7.7.2 sparkmagic==0.21.0 jupyterlab_widgets==1.1.11 && 
jupyter-kernelspec set up --user $(pip3.11 --no-cache-dir present sparkmagic | grep Location | reduce -d" " -f2)/sparkmagic/kernels/sparkkernel && 
jupyter-kernelspec set up --user $(pip3.11 --no-cache-dir present sparkmagic | grep Location | reduce -d" " -f2)/sparkmagic/kernels/pysparkkernel && 
jupyter server extension allow --user --py sparkmagic && 
cat > /residence/hadoop/.native/bin/entrypoint.sh
#!/usr/bin/env bash
mkdir -p /residence/hadoop/workspace/
livy-server begin
sleep 5
jupyter lab --no-browser --ip=0.0.0.0 --allow-root --ServerApp.root_dir=/residence/hadoop/workspace/ --ServerApp.token='' --ServerApp.password=''
EOF

# Setup Entrypoint script
RUN chmod +x /residence/hadoop/.native/bin/entrypoint.sh

# Add default SparkMagic Config
ADD --chown=hadoop:hadoop https://uncooked.githubusercontent.com/jupyter-incubator/sparkmagic/refs/heads/grasp/sparkmagic/example_config.json .sparkmagic/config.json

# Replace PATH var
ENV PATH=/residence/hadoop/.native/bin:/residence/hadoop/livy/bin/:$PATH

ENTRYPOINT ["/home/hadoop/.local/bin/entrypoint.sh"]

Run the docker construct command to construct the picture:

docker construct 
    -t glue_v5_livy 
    --file $WORKSPACE_LOCATION/jupyterlab/Dockerfile.livy_jupyter 
    $WORKSPACE_LOCATION/jupyterlab/

When the picture construct is full, you need to use the next docker run command to begin the newly constructed picture:

docker run -it --rm 
    -v ~/.aws:/residence/hadoop/.aws 
    -v $WORKSPACE_LOCATION:/residence/hadoop/workspace/ 
    -p 8998:8998 
    -p 8888:8888 
    -e AWS_PROFILE=$PROFILE_NAME 
    --name glue5_jupyter  
    glue_v5_livy

Appendix D: Including further Python libraries

On this part, we focus on including further Python libraries and putting in Python packages utilizing

Native Python libraries

So as to add native Python libraries, place them underneath a listing and assign the trail to $EXTRA_PYTHON_PACKAGE_LOCATION:

$ docker run -it --rm 
    -v ~/.aws:/residence/hadoop/.aws 
    -v $WORKSPACE_LOCATION:/residence/hadoop/workspace/ 
    -v $EXTRA_PYTHON_PACKAGE_LOCATION:/residence/hadoop/workspace/extra_python_path/ 
    --workdir /residence/hadoop/workspace 
    -e AWS_PROFILE=$PROFILE_NAME 
    --name glue5_pylib 
    public.ecr.aws/glue/aws-glue-libs:5 
    -c 'export PYTHONPATH=/residence/hadoop/workspace/extra_python_path/:$PYTHONPATH; pyspark'

To validate that the trail has been added to PYTHONPATH, you’ll be able to test for its existence in sys.path:

Python 3.11.6 (primary, Jan  9 2025, 00:00:00) [GCC 11.4.1 20230605 (Red Hat 11.4.1-2)] on linux
Sort "assist", "copyright", "credit" or "license" for extra info.
Setting default log stage to "WARN".
To regulate logging stage use sc.setLogLevel(newLevel). For SparkR, use setLogLevel(newLevel).
Welcome to
      ____              __
     / __/__  ___ _____/ /__
    _ / _ / _ `/ __/  '_/
   /__ / .__/_,_/_/ /_/_   model 3.5.2-amzn-1
      /_/

Utilizing Python model 3.11.6 (primary, Jan  9 2025 00:00:00)
Spark context Net UI accessible at None
Spark context accessible as 'sc' (grasp = native[*], app id = local-1740719582296).
SparkSession accessible as 'spark'.
>>> import sys
>>> "/residence/hadoop/workspace/extra_python_path" in sys.path
True

Putting in Python packages utilizing pip

To put in packages from PyPI (or every other artifact repository) utilizing pip, you need to use the next method:

docker run -it --rm 
    -v ~/.aws:/residence/hadoop/.aws 
    -v $WORKSPACE_LOCATION:/residence/hadoop/workspace/ 
    --workdir /residence/hadoop/workspace 
    -e AWS_PROFILE=$PROFILE_NAME 
    -e SCRIPT_FILE_NAME=$SCRIPT_FILE_NAME 
    --name glue5_pylib 
    public.ecr.aws/glue/aws-glue-libs:5 
    -c 'pip3 set up snowflake==1.0.5; spark-submit /residence/hadoop/workspace/src/$SCRIPT_FILE_NAME'

Concerning the Authors

Subramanya Vajiraya is a Sr. Cloud Engineer (ETL) at AWS Sydney specialised in AWS Glue. He’s enthusiastic about serving to clients remedy points associated to their ETL workload and implementing scalable knowledge processing and analytics pipelines on AWS. Exterior of labor, he enjoys occurring bike rides and taking lengthy walks together with his canine Ollie.

Noritaka Sekiyama is a Principal Massive Knowledge Architect on the AWS Glue group. He works primarily based in Tokyo, Japan. He’s liable for constructing software program artifacts to assist clients. In his spare time, he enjoys biking together with his street bike.