SetupLocalDevEnvironment.md

Setup Local Dev Environment

This can be your local machine or Datascience VM in Azure:

1. Fork the MLOps project: visit the MLOps github project and click Fork on top right.

   Now you can save your changes to your github.This step is optional but recommended.

2. Clone the project to your local directory: Click on the "Clone" button in github and copy the https url (like https://github.com/rsethur/MLOps.git)

3. If you have Git installed in your machine, go to command prompt to a project directory and execute git clone <URL>

   Optionally you click the "Clone or Download" button in github to download zip file & extract it.

4. In your local machine if you do not have Anaconda or Miniconda, please install. Alternatively if you do not want to use your local machine, please provision a Datascience VM - Ubuntu or Windows VM

5. Create conda environment: Via command prompt, navigate to project root folder and execute: conda env create -f "code/train/train_conda_env.yml"

6. Install the Azure CLI

7. Install the Azure ML CLI extension by executing:
   az extension add -n azure-cli-ml
   Incase you already have it, you can upgrade it:
   az extension update -n azure-cli-ml

8. Setup the environment variables to be used by python-dotenv

   1. Rename the .envtemplate to .env
   2. Fill in the values: BASE_NAME (if applicable), WORKSPACE_NAME, RESOURCE_GROUP, SUBSCRIPTION_ID, SP_APP_ID, SP_APP_SECRET, TENANT_ID

9. (Optional) Install postman tool. This will give an easy way to inspect the deployed ML service