# Introduction

This repository contains the source code of Energinet's machine learning SDK and it's associated command-line interface (CLI).

The SDK enables rapid development and deployment of machine learning models, both locally and in the cloud. It uses AzureML as backend when training models in the cloud.

# Installing the SDK

The following steps details how to install the SDK via PIP.

## Requirements

- Python 9 installed (later versions may work, but are untested)
- Azure CLI installed

## Installation steps

Start by installing the Azure CLI on your computer. Follow the instructions given at [this page](https://docs.microsoft.com/en-us/cli/azure/install-azure-cli).

Upgrade your version of PIP:

    $ python -m pip install --upgrade pip

Install the SDK:

    $ pip install energinet-ml-sdk


## Installation using conda

    $ conda create -n energinetml python=3.8
    $ conda activate energinetml
    $ conda install pip
    $ pip install azure-cli
    $ pip install energinet-ml-sdk

# Verify installation

After installation, the **energinetml** command should be available in your terminal (console).

Verify installation succeeded:

    $ energinetml --help

If the command **energinetml** was not found, try access it via one of the following commands:

    $ python -m energinetml --help
    $ python3 -m energinetml --help

# Troubleshooting

- Package version not found?


# Releasing PIP package

To release a new version of energinet-ml-sdk as a PIP package, update the version number of **energinetml/meta/PACKAGE_VERSION**, then commit it to (or merge into) master. This triggers a DevOps pipeline which builds and releases the package.
