Pack a Python package¶

By the end of this tutorial you will be able to run pyfiglet via docker:

$ docker run --rm -it pyfiglet:1.0.4 exec pyfiglet hello
 _ _ _
| |__ ___| | | ___
| '_ \ / _ \ | |/ _ \
| | | | __/ | | (_) |
|_| |_|\___|_|_|\___/

It should take around 10 minutes to complete.

Setup your environment¶

We recommend starting from a clean Ubuntu installation. If we don’t have one available, we can create one using Multipass:

Is Multipass already installed and active? Check by running

snapservicesmultipass

If we see the multipass service but it isn’t "active", then we’ll need to run sudo snap start multipass. On the other hand, if we get an error saying snap "multipass" not found, then we must install Multipass:

sudosnapinstallmultipass

Then we can create the VM with the following command:

multipass launch --memory 2G --disk 10G --name rock-dev 26.04

Finally, once the VM is up, open a shell into it:

multipassshellrock-dev

Unless stated otherwise, we will work entirely within the VM from now on.

LXD will be required for building the rock. Make sure it is installed and initialised:

sudosnapinstalllxd
lxdinit--auto

There is a known connectivity issue with LXD and Docker. To avoid this issue, enable IPv4 forwarding before installing Docker:

echo"net.ipv4.conf.all.forwarding=1"|sudotee-a/etc/sysctl.d/99-forwarding.conf
sudosystemctlrestartsystemd-sysctl

Warning

In a production environment, make sure to use a hardened firewall configuration.

In order to create the rock, we’ll install Rockcraft with classic confinement, which grants it access to the whole file system:

sudosnapinstallrockcraft--classic

We’ll use Docker to run the rock. We can install it as a snap:

sudosnapinstalldocker

By default, Docker is only accessible with root privileges (sudo). We want to be able to use Docker commands as a regular user:

sudoaddgroup--systemdocker
sudoadduser$USERdocker
newgrpdocker

Restart Docker:

sudosnapdisabledocker
sudosnapenabledocker

Note that we’ll also need a text editor. We can either install one of our choice or simply use one of the already existing editors in the Ubuntu environment (like nano).

Project setup¶

To create a new Rockcraft project, create a new directory and change into it:

mkdirpyfiglet-rock&&cdpyfiglet-rock

Next, create a file called rockcraft.yaml with the following contents:

rockcraft.yaml¶

name:pyfiglet
base:ubuntu@26.04
version:"1.0.4"# Note: should match `pyfiglet` below
summary:A rock for pyfiglet
description:A rock for pyfiglet
license:Apache-2.0
platforms:
amd64:
parts:
pyfiglet:
plugin:python
source:.
python-packages:
-pyfiglet==1.0.4# Note: should match `version` above
stage-packages:
-python3-venv

Note

This example uses python plugin along python-packages, so pip is invoked implicitly to install such packages. PEP 668 changes the behavior of Python package managers like pip by introducing a special marker (EXTERNALLY-MANAGED) that prevents pip from installing packages outside a virtual environment. Thus, python3-venv is required as a stage package.

Pack the rock with Rockcraft¶

To build the rock, run:

rockcraftpack

Run the rock in Docker¶

First, import the recently created rock into Docker:

sudorockcraft.skopeo--insecure-policycopyoci-archive:pyfiglet_1.0.4_amd64.rockdocker-daemon:pyfiglet:1.0.4

Now run the pyfiglet command from the rock:

dockerrun--rmpyfiglet:1.0.4execpyfigletitworks!

Which should print:

 _ _ _ _ 
(_) |_ __ _____ _ __| | _____| |
| | __| \ \ /\ / / _ \| '__| |/ / __| |
| | |_ \ V V / (_) | | | <\__ \_|
|_|\__| \_/\_/ \___/|_| |_|\_\___(_)

Explore the running container¶

Since the rock uses an ubuntu base, you can poke around in a running container using bash, via:

$ docker run --rm -it pyfiglet:1.0.4 exec bash
root@14d1812a2681:/# pyfiglet hi
_ _
| |__ (_)
| '_ \| |
| | | | |
|_| |_|_|