How to run Python Flask APIs using Docker

Python is a simple yet powerful coding language and it’s used widely in machine learning, artificial intelligence, and backend development. The reason why python is used widely is, its easy to learn syntax and debugging, since there are no compilation steps, it is easy to debug. Flask is a Python framework that makes it easier to develop backend or application code using python.

In this tutorial, we will see how to deploy Python Flask App using Docker, so you can easily transmit your working code in development, testing, staging, or production environment as well to a colleague’s or partner’s computer. Docker can save hours of time trying to deploy same code to another machine that is not compatible.

So Let’s get started

Prerequisite

Make sure, Docker and Docker Compose are installed, if not click on the link to get steps
Know Python version for your code

1. Go to your project

Clone your project in the required directory inside the server or local system.
Checkout required branch

2. Create Dockerfile

Create new Dockerfile
Paste below-mentioned code

FROM python:3.6
WORKDIR /app
COPY requirement* requirement.txt
RUN pip install -r requirement.txt
COPY . .

EXPOSE 5000
CMD ["flask","run"]

Here,

On line 1, Add Python version as per your requirement
2nd line defines working directory inside container, keep it as /app or define that custom
3rd line copies your requirements.txt file which has a list of packages that needs to be installed before running the code
- If you have a different file name for the packages file, change it accordingly
4th line will install all required packages as mentioned in the file copied in 3rd step
5th line will copy all your coding files from the current directory to /app working directory
6th line will expose your port number on which flask app is configured to run
7th line will tell your container at the time of startup to run flask run command

Here, you can define Environment variables in Dockerfile if required

3. Build Dockerfile

To run above Dockerfile, run this command

docker build -t python-flask:1 .

where, python-flask is your image name and 1 is image tag

You can change image tags when you create new revisions of your code

After successfully creating a Docker image out of your Dockerfile, Let’s run this Dockerfile and create a container

4. Deploy Docker image

docker run -d --name python-flask-app --restart always -p 5000:5000 python-flask:1

here -p contains 2 ports system port and container port,

If you are running some other code or service on 5000 port, you can change the system’s port and use any available port, that’s the flexibility Docker provides, where you don’t require to change the code to change the port number

For an e.g.

if 5000 port on the system is not available, you can change it via changing the first parameter -p 80:5000, it will start serving on the system’s 80 port.

Now, this command will run your flask app.

You can check console logs of your application using below command

docker logs -f python-flask-app

It will stream all live logs from your python container

5. Automation

You can also automate this use case, using simple shell-script as below

python-deployment.sh

#!/bin/bash

docker build -t python-flask:latest . 
docker rm -f python-flask-app || true
docker run -d --name python-flask-app --restart always -p 5000:5000 python-flask:latest

In above script,

You can also create variable for the docker image tag, which can be defined run-time
1st command will build your Docker image with new code
2nd command will destroy the currently running container if exists and if the container doesn’t exist it will return true and move to the next step
3rd command will run your newly build docker image

6. Testing

To test it,

curl localhost:5000
curl $Elastic-IP:5000

You can use the curl command to fetch the response from your flask API, if it doesn’t respond you can check docker logs and troubleshoot the error.

7. Sharing Docker image

If you want to share this Docker image with your teammates or colleagues, simply upload it to Docker Images repositories from where it can be shared with others.

There are various Docker image repositories available