Docker Compose Python Eve Mongodb
Aug 18, 2017
4 minute read


As a Python developer you are probably familiar with PyCharm, an amazing IDE from JetBrains. If you don’t know anything about it, have a look, you’ll probably not regret it :) Some of the steps in this article will make mention of PyCharm, however, it’s not required and you can still follow if you’re using another IDE. If you’re not using any IDE, well, god save your soul.

What we will do here

We will go through Eve’s quickstart guide adding the docker context.


Check you have the required software on your machine.

$ docker -v
Docker version 17.06.0-ce, build 02c1d87
$ docker-compose -v
docker-compose version 1.14.0, build c7bdf9e

If any of these commands are not found, check the docker documentation for install procedure.

Let’s get started!

Eve’s requirements

Eve’s quickstart guide requires MongoDB installed and running.

If we translate that into docker, we will have:

  • a Python container, running our Eve app
  • a MongoDB container, linked to the first one

Ok, that’s the role of docker compose when you can instruct docker about multiple containers and how they interact.

Python app

The Python app is our Eve app, so let’s get it started:

Create a new file, with

from eve import Eve
app = Eve()

if __name__ == '__main__':'')

and another one, with

MONGO_HOST = "mongo"
DOMAIN = {'people': {}}

We need to specify here the host on which mongo runs, otherwise Eve assumes it’s local.

Note the host is just called “mongo”, which is the name we will give it in the docker-compose.yml file below.

This is our Eve app! Let’s instruct docker about it now.

add a new file, called Dockerfile

FROM python:3


WORKDIR /usr/src/app

COPY requirements.txt ./
RUN pip install --no-cache-dir -r requirements.txt

COPY . .

CMD [ "python", "./" ]

This will use a base python image and run our app in it, by:

  • Exposing port 5000 so that it can be accessed outside the docker container
  • Installing the requirements.txt with pip
  • Embedding our code into /usr/src/app
  • Running the command python

Ok, where are the requirements.txt? Let’s add them

eve == 0.7.4

We just need the eve Python package here.


For MongoDB, we will just use their official container.

Let’s wrap it all together

Add a new file, docker-compose.yml

version: '2'
    build: .
      - "5000:5000"
      - .:/usr/src/app
      - mongo
    image: "mongo"

This defines our stack, according to version 3 of the docker-compose reference:

  • One container called “web”, where our Eve app is
    • which port 5000 is bound to host’s port 5000
    • where the project directory being shared with /usr/src/app in the container (this allows us to make changes to the code without having to rebuild the container each time)
    • which links to another container called “mongo”
  • One container called “mongo”

Let’s run it now!

$ docker-compose up

After putting all the layers together, this command should end with:

web_1    |  * Running on (Press CTRL+C to quit)

In a terminal, just do as Eve’s quickstart suggests:

$ curl -i

HTTP/1.0 200 OK
Content-Type: application/json
Content-Length: 62
Server: Eve/0.7.4 Werkzeug/0.11.15 Python/3.6.2
Date: Fri, 18 Aug 2017 15:31:52 GMT

{"_links": {"child": [{"href": "people", "title": "people"}]}}%   

And then try:

$ curl

    "_items": [],
    "_links": {
        "parent": {
            "title": "home",
            "href": "/"
        "self": {
            "title": "people",
            "href": "people"
    "_meta": {
        "page": 1,
        "max_results": 25,
        "total": 0

This command is useful to test the connection with mongo. If it’s not properly configured, it will fail.

Also try the DELETE command and verify it’s not authorised as suggested by the Quickstart from Eve.

Awww, Stop it!

Ok we can now stop the containers (CTRL+C) and configure PyCharm now.

PyCharm setup

Wouldn’t it be cool to be able to debug your app while it runs in the container?

That’s what we will do now.

PyCharm doesn’t currently support docker-compose reference version 3.

Configure docker

Configure docker-compose

Under project interpreter, click “Add remote”, then docker-compose and make sure you select “web”.

Add path mapping:

You can now run your debugger which will automatically launch the docker compose command.

Source code

Source code is available at

Further reading

This article is inspired by a blog post from JetBrains.