AWS EMR install python libraries - python

I am running a Map Reduce Code in Amazon EMR using Python which uses the native boto library. I need to know which packages are pre-installed in the cluster nodes ? Also how do I automatically install some modules while bootstrapping ?

Related

Installing a Python library wheel in an Azure Machine Learning Compute Cluster

I am working on an Azure Machine Learning Studio pipeline via the Designer. I need to install a Python library wheel (a third-party tool) in the same compute, so that I can import it into the designer. I usually install packages to compute instances via the terminal, but the Azure Machine Learning Studio designer uses a compute cluster, not a compute instance.
Is there any way to access the terminal so that I can install the wheel in the compute cluster and have access to the library via the designer? Thanks!
There isn't an easy path for this. Your options are either, switch to a code-first pipeline definition approach, or try your darndest to extend the Designer UI to meet your needs.
Define pipelines with v2 CLI or Python SDK
It looks like you're already outside of I get the impression that you know Python quite well, you should really check out the v2 CLI or the Python SDK for Pipelines. I'd recommend maybe starting with the v2 CLI as it will be the way to define AML jobs in the future.
Both require some initial learning, but will give you all the flexibility that isn't currently available in the UI.
custom Docker image
The "Execute Python Script" module allows use a custom python Docker image. I think this works? I just tried it but not with a custom .whl file, and it looked like it worked

Accessing delta lake through Pyspark on EMR notebooks

I have a query with respect to using external libraries like delta-core over AWS EMR notebooks. Currently there isn’t any mechanism of installing the delta-core libraries through pypi packages. The available options include.
Launching out pyspark kernel with --packages option
The other option is to change the packages option in the python script through os configuration, but I don’t see that it is able to download the packages and I still get import error on import delta.tables library.
Third option is to download the JARs manually but it appears that there isn’t any option on EMR notebooks.
Has anyone tried this out before?
You can download the jars while creating EMR using bootstrap scripts.
You can place the jars in s3 and pass it to pyspark with --jars option

How to Connect to RDS Instance from AWS Glue Python Shell?

I am trying to access RDS Instance from AWS Glue, I have a few python scripts running in EC2 instances and I currently use PYODBC to connect, but while trying to schedule jobs for glue, I cannot import PYODBC as it is not natively supported by AWS Glue, not sure how drivers will work in glue shell as well.
From: Introducing Python Shell Jobs in AWS Glue announcement:
Python shell jobs in AWS Glue support scripts that are compatible with Python 2.7 and come pre-loaded with libraries such as the Boto3, NumPy, SciPy, pandas, and others.
The module list doesn't include pyodbc module, and it cannot be provided as custom .egg file because it depends on libodbc.so.2 and pyodbc.so libraries.
I think you have 2 options:
Create a jdbc connection to your DB from Glue's console, and use Glue's internal methods to query it. This will require code changes of course.
Use Lambda function instead. You'll need to pack pyodbc and the required libs along with your code in a zip file. Someone has already compiled those libs for AWS Lambda, see here.
Hope it helps
For AWS Glue use either Dataframe/DynamicFrame and specify the SQL Server JDBC driver. AWS Glue already contain JDBC Driver for SQL Server in its environment so you don't need to add any additional driver jar with glue job.
df1=spark.read.format("jdbc").option("driver", "com.microsoft.sqlserver.jdbc.SQLServerDriver").option("url", url_src).option("dbtable", dbtable_src).option("user", userID_src).option("password", password_src).load()
if you are using a SQL instead of table:
df1=spark.read.format("jdbc").option("driver", "com.microsoft.sqlserver.jdbc.SQLServerDriver").option("url", url_src).option("dbtable", ("your select statement here") A).option("user", userID_src).option("password", password_src).load()
As an alternate solution you can also use jtds driver for SQL server in your python script running in AWS Glue
If anyone needs a postgres connection with sqlalchemy using python shell, it is possible by referencing the sqlalchemy, scramp, pg8000 wheel files, it's important to reconstruct the wheel from pg8000 by eliminating the scramp dependency on the setup.py.
I needed to so something similar and ended up creating another Glue job in Scala while using Python for everything else. I know it may not work for everyone but wanted to mention How to run DDL SQL statement using AWS Glue
I was able to use the python library psycopg2 even though it is not written in pure python and it does not come preloaded with aws glue python shell environment. This runs contrary to aws glue documentation. So you might be able to use odbc related python libraries in a similar way. I created .egg files for psycopg2 library and used it successfully within glue python shell environment. Following are the logs from glue python shell if you have import psycopg2 in your script and the glue job refers to the related psycopg2 .egg files.
Creating /glue/lib/installation/site.py
Processing psycopg2-2.8.3-py2.7.egg
Copying psycopg2-2.8.3-py2.7.egg to /glue/lib/installation
Adding psycopg2 2.8.3 to easy-install.pth file
Installed /glue/lib/installation/psycopg2-2.8.3-py2.7.egg
Processing dependencies for psycopg2==2.8.3
Searching for psycopg2==2.8.3
Reading https://pypi.org/simple/psycopg2/
Downloading https://files.pythonhosted.org/packages/5c/1c/6997288da181277a0c29bc39a5f9143ff20b8c99f2a7d059cfb55163e165/psycopg2-2.8.3.tar.gz#sha256=897a6e838319b4bf648a574afb6cabcb17d0488f8c7195100d48d872419f4457
Best match: psycopg2 2.8.3
Processing psycopg2-2.8.3.tar.gz
Writing /tmp/easy_install-dml23ld7/psycopg2-2.8.3/setup.cfg
Running psycopg2-2.8.3/setup.py -q bdist_egg --dist-dir /tmp/easy_install-dml23ld7/psycopg2-2.8.3/egg-dist-tmp-9qwen3l_
creating /glue/lib/installation/psycopg2-2.8.3-py3.6-linux-x86_64.egg
Extracting psycopg2-2.8.3-py3.6-linux-x86_64.egg to /glue/lib/installation
Removing psycopg2 2.8.3 from easy-install.pth file
Adding psycopg2 2.8.3 to easy-install.pth file
Installed /glue/lib/installation/psycopg2-2.8.3-py3.6-linux-x86_64.egg
Finished processing dependencies for psycopg2==2.8.3
These are the steps that I used to connect to an RDS from glue python shell job:
Package up your dependency package into an egg file (these package must be pure python if I remember correctly). Put it in S3.
Set your job to reference that egg file under the job configuration > Python library path
Verify that your job can import the package/module
Create a glue connection to your RDS (it's in Database > Tables, Connections), test the connection make sure it can hit your RDS
Now in your job, you must set it to reference/use this connection. It's in the require connection as you configure your job or edit your job.
Once those steps are done and verify, you should be able to connect. In my sample I used pymysql.

How does AWS know where my imports are?

I'm new to AWS Lambda and pretty new to Python.
I wanted to write a python lambda that uses the AWS API.
boto is the most popular python module to do this so I wanted to include it.
Looking at examples online I put import boto3 at the top of my Lambda and it just worked- I was able to use boto in my Lambda.
How does AWS know about boto? It's a community module. Are there a list of supported modules for Lambdas? Does AWS cache its own copy of community modules?
AWS Lambda's Python environment comes pre-installed with boto3. Any other libraries you want need to be part of the zip you upload. You can install them locally with pip install whatever -t mysrcfolder.
The documentation seems to suggest boto3 is provided by default on AWS Lambda:
AWS Lambda includes the AWS SDK for Python (Boto 3), so you don't need to include it in your deployment package. However, if you want to use a version of Boto3 other than the one included by default, you can include it in your deployment package.
As far as I know, you will need to manually install any other dependencies in your deployment package, as shown in the linked documentation, using:
pip install foobar -t <project path>
AWS Lambda includes the AWS SDK for Python (Boto 3), so you don't need to include it in your deployment package.
This link will give you a little more in-depth info on Lambda environment
https://aws.amazon.com/blogs/compute/container-reuse-in-lambda/
And this too
https://alestic.com/2014/12/aws-lambda-persistence/

How to create AWS Lambda deployment package that uses Couchbase Python client

I'm trying to use AWS Lambda to transfer data from my S3 bucket to Couchbase server, and I'm writing in Python. So I need to import couchbase module in my Python script. Usually if there are external modules used in the script, I need to pip install those modules locally and zip the modules and script together, then upload to Lambda. But this doesn't work this time. The reason is the Python client of couchbase works with the c client of couchbase: libcouchbase. So I'm not clear what I should do. When I simply add in the c client package (with that said, I have 6 package folders in my deployment package, the first 5 are the ones installed when I run "pip install couchbase": couchbase, acouchbase, gcouchbase, txcouchbase, couchbase-2.1.0.dist-info; and the last one is the c client of Couchbase I installed: libcouchbase), lambda doesn't work and said:
"Unable to import module 'lambda_function': libcouchbase.so.2: cannot open shared object file: No such file or directory"
Any idea on how I can get the this work? With a lot of thanks.
Following two things worked for me:
Manually copy /usr/lib64/libcouchbase.so.2 into ur project folder
and zip it with your code before uploading to AWS Lambda.
Use Python 2.7 as runtime on the AWS Lambda console to connect to couchbase.
Thanks !
Unfortunately AWS Lambda does not support executing C-based python modules, like the Couchbase SDK.
Your best bet would be to use a pure-python client. The easiest way to do this would be to use the unofficial memcached client https://github.com/couchbase/couchbase-cli/blob/master/cb_bin_client.py which uses server-side moxi to handle memcached clients on port 11211.

Categories