Tutorial (T7): Data Versioning Demo

In this tutorial, we will cover data versioning techniques using the cheese app dataset. Everything will be run inside containers using Docker.

Prerequisites

• Have the latest Docker installed

Make sure we do not have any running containers and clear up an unused images

• Run docker container ls
• Stop any container that is running
• Run docker system prune
• Run docker image ls

Clone the data-versioning-ac215 branch from the data-versioning repository to your local machine

In order to work on this task, you need to have a repository with write permission (the class repo only gives you read permissions). This can be done by either forking the repository or copying its code.

To fork the repository, follow these steps:

• Clone the branch data-versioning-ac215 of the repository using the following command:

git clone -b data-versioning-ac215 [email protected]:dlops-io/data-versioning.git This example uses SSH, so you'll need to have your public SSH key added to your GitHub account. Alternatively, you can clone using HTTPS (not explained here).

• Remove the .git folder from the cloned repository. This will delete the Git metadata, unlinking the code from the original repository. rm -rf data-versioning/.git

• Initialize a new git repository. cd data-versioning git init

• (Optional) Create a LICENSE to define usage permissions. (.gitignore is already included you may update it to exclude files you don't want to track)

• Create a new private repository on GitHub without any files ie. no README.md, no LICENSE, no .gitignore etc. New repo url: https://github.com/YOUR_GITHUB_USERNAME/data-versioning

• Add the remote and push your changes

git init
git remote add origin [email protected]:YOUR_GITHUB_USERNAME/data-versioning.git
git add .
git commit -m "Initial commit"
git push -u origin main

Adding the secrets folder

Your folder structure should look like this:

   |-data-versioning
   |-secrets

Setup GCP Service Account

• Here are the step to create a service account:
• To setup a service account you will need to go to GCP Console, search for "Service accounts" from the top search box. or go to: "IAM & Admins" > "Service accounts" from the top-left menu and create a new service account called "data-service-account". For "Service account permissions" select "Cloud Storage" > "Storage Admin" (Type "cloud storage" in filter and scroll down till you find). Then click continue and done.
• This will create a service account
• On the right "Actions" column click the vertical ... and select "Manage keys". A prompt for Create private key for "data-service-account" will appear select "JSON" and click create. This will download a Private key json file to your computer. Copy this json file into the secrets folder. Rename the json file to data-service-account.json

Create a Data Store folder in GCS Bucket

• Go to https://console.cloud.google.com/storage/browser
• Go to the bucket cheese-app-data-versioning (REPLACE WITH YOUR BUCKET NAME)
• Create a folder dvc_store inside the bucket
• Create a folder images inside the bucket (This is where we will store the images that need to be versioned)

Run DVC Container

We will be using DVC as our data versioning tool. DVC (Data Version Control) is an open-source, Git-based data science tool. It applies version control to machine learning development, make your repo the backbone of your project.

Setup DVC Container Parameters

In order for the DVC container to connect to our GCS Bucket open the file docker-shell.sh and edit some of the values to match your setup

export GCS_BUCKET_NAME="cheese-app-data-versioning" [REPLACE WITH YOUR BUCKET NAME]
export GCP_PROJECT="ac215-project" [REPLACE WITH YOUR GCP PROJECT]
export GCP_ZONE="us-central1-a"  [REPLACE WITH YOUR GCP ZONE]

Note: Addition of docker-entrypoint.sh

Note that we have added a new file called docker-entrypoint.sh to our development flow. A docker-entrypoint.sh is used to simplify some task when running containers such as:

• Helps with Initialization and Setup:
  • The entrypoint file is used to perform necessary setup tasks when the container starts.
  • It is a way to ensure that certain operations occur every time the container runs, regardless of the command used to start it.
• Helps with Dynamic Configuration:
  • It allows for dynamic configuration of the container environment based on runtime variables or mounted volumes.
  • This is more flexible than hardcoding everything into the Dockerfile.

For this container we need to:

• Mount a GCS bucket to a volume mount in the container
• We then mount the "images" folder in the bucket mount to the "/app/cheese_dataset" folder

Run docker-shell.sh

• Make sure you are inside the data-versioning folder and open a terminal at this location
• Run sh docker-shell.sh

Version Data using DVC

In this step we will start tracking the dataset using DVC

Initialize Data Registry

In this step we create a data registry using DVC dvc init

Add Remote Registry to GCS Bucket (For Data)

dvc remote add -d cheese_dataset gs://cheese-app-data-versioning/dvc_store [REPLACE WITH YOUR BUCKET NAME]

Add the dataset to registry

dvc add cheese_dataset

Push to Remote Registry

dvc push

You can go to your GCS Bucket folder dvc_store to view the tracking files

Update Git to track DVC

Run this outside the container.

• First run git status git status
• Add changes git add .
• Commit changes git commit -m 'dataset updates...'
• Add a dataset tag git tag -a 'dataset_v20' -m 'tag dataset'
• Push changes git push --atomic origin main dataset_v20

Download Data to view version

In this Step we will use Colab to view various version of the dataset

• Open Colab Notebook
• Follow instruction in the Colab Notebook

Make changes to data

Upload images

• Upload a few more images to the images folder in your bucket (We are simulating some change in data)

Add the dataset (changes) to registry

dvc add cheese_dataset

Push to Remote Registry

dvc push

Update Git to track DVC changes (again remember this should be done outside the container)

• First run git status git status
• Add changes git add .
• Commit changes git commit -m 'dataset updates...'
• Add a dataset tag git tag -a 'dataset_v21' -m 'tag dataset'
• Add a remote origin git remote add origin [email protected]:YOUR_GITHUB_USERNAME/data-versioning.git
• Push changes git push --atomic origin main dataset_v21

Download Data to view version

In this Step we will use Colab to view the new version of the dataset

• Open Colab Notebook
• Follow instruction in the Colab Notebook to view dataset_v21

🎉 Congratulations we just setup and tested data versioning using DVC

Docker Cleanup

To make sure we do not have any running containers and clear up an unused images

• Run docker container ls
• Stop any container that is running
• Run docker system prune
• Run docker image ls