Poetry¶

Poetry is a virtual environment and dependency manager, builder and publisher for python. We are using it as a tool to update our python dependencies.

The configuration file, pyproject.toml in the project root, contains the list of dependencies. From there, we can export a requirements.txt file for use with pip, which we use in our Dockerfile.

How to add poetry to the running container¶

Adding poetry to the running container lets us manage project dependencies without installing it on the host machine. Not adding it to the docker image keeps the image smaller.

Make sure the container is running
```
docker-compose up -d # (1)!
```
1. This runs the container in background (daemon) mode
Install poetry inside the running container
```
docker-compose exec mkdocs pip install poetry # (1)!
```
1. Runs pip install poetry in the mkdocs service container
Now we can call poetry that's in the image
Use poetry for dependency management or other purposes.
Take down the container when done
```
docker-compose down # (1)!
```
1. This stops and deletes the container. The poetry install is gone as well. If we don't want to delete the container, use docker-compose stop.

How to add poetry to the image¶

Adding poetry to the image lets us manage the project dependencies without installing it on the host machine.

Add poetry to Dockerfile

Dockerfile

...
# install dependencies
COPY requirements.txt .
RUN pip install --no-cache-dir -r requirements.txt
RUN pip install --no-cache-dir poetry==1.5.1
...

Build the image
```
docker-compose build
```
Now we can call poetry that's in the image

How to add a package¶

We use pyproject-fmt as an example. pyproject-fmt is an auto-formatter for the pyproject.toml configuration file.

Get a shell inside the container

docker-compose run mkdocs sh

Install and run the package

pip install pyproject-fmt
pyproject-fmt pyproject.toml

How to add package to a group¶

Organizing packages into groups allows better organization of dependencies. For example, dev dependencies and docs dependencies as opposed to the ones for the main application.

poetry add pytest --group dev