A easy strategy to run VS Code in a container, within the cloud
Earlier this 12 months, I revealed 2 guides (right here, and right here) to internet hosting code-server on SageMaker. It exhibits how one can run VS Code on scalable cloud compute, and code from wherever with out worrying about native setup. All you want is an web connection.
On this submit, we’ll go a step additional and present the way to do container-based growth within the cloud, to hurry up your ML venture supply.
We are going to pull an instance python 3.10 container picture from Dockerhub, and develop in it with VS Code on SageMaker. You have to entry to a SageMaker Pocket book Occasion to undergo the instance. A small one must be sufficient.
With container-based growth, you write, execute, and check your ML code inside containers. It means that you would be able to work with the identical libraries and dependencies as your group members, making collaboration simpler with constant and extremely reproducible growth environments.
One other benefit is that you simply develop and check your code in an atmosphere that may behave precisely like your manufacturing atmosphere, serving to transfer quicker from 0 to 1 in your venture.
And internet hosting code-server on SageMaker offers minimal modifications to an area growth expertise in VS Code, permitting you to code from wherever, on scalable cloud compute.
“However I assumed VS Code was Open Supply”
Usually with VS Code, you’ll use the Dev Containers extension to do container-based growth. Nonetheless, you can not use it with code-server, nor every other open supply VS Code tasks.
Whereas the core of VS Code is open supply, {the marketplace} and plenty of extensions revealed by Microsoft aren’t. And Microsoft prohibits the usage of any non-Microsoft VS Code from accessing their market. See Variations in comparison with VS Code for extra particulars.
So how will we do container-based growth with VS Code on SageMaker then?
So we can not deliver the container to the IDE attributable to Microsoft extension license points. However what if we deliver the IDE to the container? In reality, code-server is web-based and may really run wherever. The setup could be very simple and you are able to do it in 3 steps:
Step 1: Add code-server to your current Dockerfile
You’ll be able to add code-server to your container with a single line of code. Right here is an instance for including it to an a python 3.10 picture from Dockerhub
Add one line and that is it!
Step 2: Construct and launch the container picture on the Pocket book Occasion
Launch a terminal through the Jupyter/Jupyterlab of your Pocket book Occasion:
Now you can navigate to your Dockerfile folder, and construct the container picture domestically with the next command:
# BUILD THE CONTAINER IMAGE LOCALLY
docker construct -t ide-in-container:newest .
As soon as the picture is constructed, you possibly can run the container with the command to launch code-server:
# RUN CONTAINER WITH VS CODE IN IT
docker run --rm -it -p 8080:8080 ide-in-container:newest code-server --auth none --disable-telemetry --bind-addr 0.0.0.0:8080
We expose code-server to the occasion by the default port 8080.
Step 3: Entry VS Code by the Pocket book Occasion Jupyter Proxy
Now, all you want is to repeat your Pocket book Occasion URL, change it a bit, and previous it into a brand new tab:
/proxy/8080/ and pasted it in a brand new browser tab.The tab ought to now open VS Code:
Notes
To additional streamline your work, you possibly can mount the native Pocket book Occasion quantity, and level VS Code to your current configuration (extensions, key bindings, settings). Within the instance Dockerfile, I level VS Code to the /house/SageMaker/vscode-config config folder, I already had on the occasion. That is executed utilizing the XDG_DATA_HOME atmosphere variable.
And right here is how one can mount the Pocket book Occasion quantity to the container:
# RUN THIS COMMAND IF YOU WANT TO MOUNT THE NOTEBOOK INSTANCE HOME/SAGEMAKER FOLDER
docker run --rm -it -p 8080:8080 -v /house/ec2-user/SageMaker/:/house/SageMaker ide-in-container:newest code-server --auth none --disable-telemetry --bind-addr 0.0.0.0:8080
To enhance your collaboration, you possibly can add this “atmosphere Dockerfile” in your venture group GitHub repo. on this method it will likely be versioned and everybody can work with the very same dependencies, at any time.
Growing inside containers with cloud-based IDEs might help streamline your ML venture and group collaboration by making it simple to run and check code in a constant atmosphere, and by permitting you to entry that growth atmosphere from wherever.
On this submit, I’ve shared with you a easy strategy for working VS Code in a container, hosted on Amazon SageMaker.
To go additional, you possibly can go to 5 Easy Steps to MLOps with GitHub Actions, MLflow, and SageMaker Pipelines and find out how your group can simply run an end-to-end MLOps venture utilizing GitHub Actions and SageMaker.
