> *There are only two ways to live your life. One is as though nothing is a miracle. The other is as though everything is a miracle. -Albert Einstein*

In this article, we will talk about what Docker is, how it works and how to deploy a Jupyter notebook to a Docker Container.

## What is Docker?

 ![docker-about.png](https://cdn.hackernoon.com/images/ckzto-7-ery-008-n-0-as-69-its-610-u.jpg)According to the __[Docker Website](https://www.docker.com/)__, Docker is an open platform for developing, shipping, and running applications. Docker enables you to separate your applications from your infrastructure so you can deliver software quickly.

In other words, Docker is a platform that provides a container for you to run host, and run your applications in without bothering about things like platform dependence, it provides infrastructure called a container where your platforms can be held and run.

## What makes up Docker (In a Nutshell)

Here we will provide an overview of what docker up, if you want a comprehensive overview of how docker works, check __[this article](https://devopscube.com/what-is-docker/)__ out.

The Docker Architecture is divided into three(3) sections:

* Docker Engine(dockerd)
* docker-containerd (contained)
* docker-runc (runc)

 ![docker-2.png](https://cdn.hackernoon.com/images/ckzto-7-erz-008-o-0-as-6-hrgo-7-x-4-q.jpg)

### Docker Engine(dockerd)

Docker engine comprises the docker daemon, an API interface, and Docker CLI. Docker daemon (dockerd) runs continuously as dockerd system service. It is responsible for building the docker images.

### Docker-containerd

*containerd* is another system daemon service that is responsible for downloading the docker images and running them as a container. It exposes its API to receive instructions from the dockerd service

### Docker-runc

*runc* is the container runtime responsible for creating the namespaces and *cgroups* required for a container. It then runs the container commands inside those namespaces. *runc* runtime is implemented as per the OCI specification.

## How to Deploy a Colab Jupyter Notebook to a Docker Container

 ![docker-0.png](https://cdn.hackernoon.com/images/ckzto-7-es-0008-p-0-as-67-cvead-94.jpg)

In this part, we are going to work build a simple classifier model using the __[Iris Dataset](https://scikit-learn.org/stable/auto_examples/datasets/plot_iris_dataset.html)__, after that we will import the code from colab and finally we install and run deploy the script containing the model into a docker container.

### Building Model

In this section, we will build the classifier model using the sklearn's inbuilt __[Iris Dataset](https://scikit-learn.org/stable/auto_examples/datasets/plot_iris_dataset.html)__.

**STEP 1:** Create a new notebook in __[google colab](https://colab.research.google.com/drive/16jcaJoc6bCFAQ96jDe2HwtXj7BMD_-m5)__.

 ![colab-opening.png](https://cdn.hackernoon.com/images/ckzto-7-es-0008-q-0-as-60-sdgeqpu.jpg)

**STEP 2** Import the dependencies.

\
```javascript
import matplotlib.pyplot as plt
from sklearn import metrics
from sklearn.datasets import load_iris
from sklearn.model_selection import train_test_split
from sklearn.neighbors import KNeighborsClassifier
```

**STEP 3** Here we are going to load the iris dataset, split the data into the training set and test set, and build our classification model.

\
```javascript
iris = load_iris()
X = iris.data
y = iris.target
```

Above we loaded the Iris dataset.

\
```javascript
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2 , random_state=4)
```

Above we used the *train_test_split* function in sklearn to split the iris dataset into a training set and test set.

\
```javascript
knn = KNeighborsClassifier(n_neighbors=10)
knn.fit(X,y)
```

Above we instantiated the KNeighbors Classifier model and tuned the n_neighbours hyperparameter to contain ten(10) neighbors.

### Installing and deploying to Docker

This is the final chapter, here we are going to install the docker desktop application and write the scripts which will deploy our script to a docker container.

**STEP 1** Firstly, you download the python script containing your trained model from colab.

 ![colab-import-code.png](https://cdn.hackernoon.com/images/ckzto-7-es-1008-r-0-as-63-ohsgbsn.jpg)

**STEP 2** Now we are going to install and set up docker. You can install docker using this__[link](https://docs.docker.com/get-started/#download-and-install-docker)__ **STEP 3** Now create a directory called **iris-classifier** where we are going to host our model and docker scripts.

Move the python file containing the iris classification model to the iris-classification folder just created.

In the same folder, create a text file called **requirements**, below are the contents it will contain.

\
```javascript
sklearn==0.0
matplotlib==3.2.2
```

**STEP 4** Here we will create the **Dockerfile**, go to your main directory and create a file called **Dockerfile** without any extension. A dockerfile is a script that is used to create a container image. Below are the items that will be contained in your Dockerfile.

\
```javascript
FROM python:3.8

ADD requirements.txt /

RUN pip install -r /requirements.txt

ADD iris-classifier.py /

ENV PYTHONUNBUFFERED=1

CMD [ "python", "./iris-classifier.py" ]
```

Above we simply told docker what to do each time the container is run.

**STEP 5** Here we are going to create our **Docker Compose file**, docker-compose files are simply configuration files that make it easy to maintain different Docker containers.

In your project directory, create a file called **docker-compose.yml**, below are the contents to be contained in the file.

\
```javascript
version: "3"
services:
  iris-classifier-uplink:
    # if failure  or server restarts, container will restart
    restart: always 
    container_name: iris-classifier-uplink
    image: iris-classifier-uplink
    build: 
      # build classifier image from the Dockerfile in the current directory
      context: .
```

Now in your directory, **iris-classifier** you should have three(3) files.

* docker-compose.yml
* Dockerfile
* __[iris-classifier.py](http://iris-classifier.py/)__
* requirements.txt

### Running Docker Container

This is the final step, here we will run our docker container using the commands stated below.

\
```javascript
docker compose build

docker compose up -d
```

This is the end, our Python model is now running in a docker container!

## Useful Links

* __[Docker Website](https://www.docker.com/)__
* __[Docker Blog](https://www.docker.com/blog/)__
* __[Docker Documentation](https://docs.docker.com/)__
* __[Deep Dive into how Docker works](https://devopscube.com/what-is-docker/)__

## EndNote

Jupyter Notebooks are really good places for building models and you can also use them as back ends for applications, unfortunately, they don't run forever.

Docker helps you fix that by re-running your jupyter notebook when it fails and this makes it a tool worth knowing.

\
*First Published [here](https://edemgold.blog/how-to-deploy-a-jupyter-notebook-to-docker)*

Google

Deploy Your First Jupyter Notebook to Docker

Too Long; Didn't Read

People Mentioned

Companies Mentioned

@edemgold

RELATED STORIES