Manage Docker with Python¶

Video: to follow...

There are two main libraries for this:

Docs: Python on Whales
Docs: Docker Py

For Python on Whales, a very useful script is:

from python_on_whales import DockerClient

docker = DockerClient(compose_files=["../docker-compose.yml"])

docker.compose.build()
docker.compose.up(detach=True)

# docker.compose.down()

By using a docker-compose.yaml file, we can use PyYaml to read, write and edit these files, enabling programatic customisation and creation of containers.

Another example is :

docker run --name some-postgres -e POSTGRES_PASSWORD=mysecretpassword -d postgres

becomes

from python_on_whales import docker

docker.run(
    "postgres:9.6",
    name="some-postgres",
    envs={"POSTGRES_PASSWORD": "mysecretpassword"},
    detach=True,
)
print(docker.ps())
# [python_on_whales.Container(id='f5fb939c409d', name='some-postgres')]

which can also be managed programatically.

For Docker-Py we have:

import docker
client = docker.from_env()
image = client.images.pull("redis")
output = client.containers.run("ubuntu", "echo hello world")
output = output.decode("utf-8")
print(str(output))

containers = client.containers.list()
for container in containers:
    print(container.name)
PORT = random.randint(8000, 9000)
container = client.containers.run("nginx:latest", detach=True, ports={"80/tcp": PORT})

Docker enables the use of ARG and --build-arg to add arguments prior to FROM, FROM usually being the first allowed command

ARG BaseImage
FROM $BaseImage

and this means we can use this and other shell variables to script a lot of processes and create our own testing matrix.

For the Docker Py library we have:

import docker
client = docker.from_env()
image = client.images.pull("redis")
output = client.containers.run("ubuntu", "echo hello world")
output = output.decode("utf-8")
print(str(output))

containers = client.containers.list()
for container in containers:
    print(container.name)
PORT = random.randint(8000, 9000)
container = client.containers.run("nginx:latest", detach=True, ports={"80/tcp": PORT})

Sources of YT video on Python and YAML

MathByte - shorter and more concise video with its repo installed in this project in yaml folder.
DevOpsMadeEasy which has a YT video (1hr 45m) here.

The repo and video explore managing Docker programatically with Python, mostly through the use of editing docker compose files and then running them programatically.