Jubilant-backports, the joyful library for integration-testing charms in Juju 2.9 and above

Jubilant is a Python library that wraps the Juju CLI for use in charm integration tests. It provides methods that map 1:1 to Juju CLI commands, but with a type-annotated, Pythonic interface.

Jubilant-backports is a Python library that modifies Jubilant to provide support for running with Juju 2.9. It is a drop-in replacement for Jubilant, using the regular Jubilant package code for Juju 3.0 and above, as well as when there are no differences between 2.9 and higher, and custom code from a Jubilant branch otherwise. It supports the Jubilant 1.4 feature set, and future improvements to Jubilant are unlikely to be backported to Jubilant-backports (but security fixes and bug fixes, where required, will be backported).

You should consider using Jubilant-backports if you need to run the same charm integration test suite against Juju 2.9 as well as more modern Juju version.

Using Jubilant-backports

Jubilant-backports is published to PyPI, so you can install and use it with your favorite Python package manager:

$ pip install jubilant-backports
# or
$ uv add jubilant-backports

Because Jubilant-backports calls the Juju CLI, you'll also need to install Juju.

To use Jubilant-backports in Python code:

import jubilant_backports as jubilant

juju = jubilant.Juju()
juju.deploy('snappass-test')
juju.wait(jubilant.all_active)

# Or only wait for specific applications:
juju.wait(lambda status: jubilant.all_active(status, 'snappass-test', 'another-app'))

Note that we have done import jubilant_backports as jubilant and then used jubilant.Juju() rather than import jubilant_backports and jubilant_backports.Juju(). This will result in a simpler and cleaner change when you are ready to drop support for Juju 2.9, and want to move to the regular Jubilant package: all you need to do is change the dependency and the import statements, and the rest of your tests can be left unchanged.

Below is an example of a charm integration test. First we define a module-scoped pytest fixture named juju which creates a temporary model and runs the test with a Juju instance pointing at that model. Jubilant-backports's temp_model context manager creates the model during test setup and destroys it during teardown:

# conftest.py
@pytest.fixture(scope='module')
def juju():
    with jubilant.temp_model() as juju:
        yield juju


# test_deploy.py
def test_deploy(juju: jubilant.Juju):        # Use the "juju" fixture  # type: ignore
    juju.deploy('snappass-test')             # Deploy the charm
    status = juju.wait(jubilant.all_active)  # Wait till the app and unit are 'active'

    # Hit the Snappass HTTP endpoint to ensure it's up and running.
    address = status.apps['snappass-test'].units['snappass-test/0'].address
    response = requests.get(f'http://{address}:5000/', timeout=10)
    response.raise_for_status()
    assert 'snappass' in response.text.lower()

You don't have to use pytest with Jubilant-backports, but it's what we recommend. Pytest's assert-based approach is a straight-forward way to write tests, and its fixtures are helpful for structuring setup and teardown.

Contributing and developing

Anyone can contribute to Jubilant-backports. The code lives in a branch of the regular Jubilant. It's best to start by opening an issue with a clear description of the problem or feature request, but you can also open a pull request directly.

Jubilant-backports uses uv to manage Python dependencies and tools, so you'll need to install uv to work on the library. You'll also need make to run local development tasks (but you probably have make installed already).

After that, clone the Jubilant codebase, check out the jubilant-backports branch, and use make all to run various checks and the unit tests:

$ git clone https://github.com/canonical/jubilant
Cloning into 'jubilant-backports'...
...
$ cd jubilant-backports
$ git checkout jubilant-backports
$ make all
...
========== 93 passed in 0.81s ==========

To contribute a code change, write your fix or feature, add tests and docs, then run make all before you push and create a PR. Once you create a PR, GitHub will also run the integration tests, which takes several minutes.

Doing a release

To create a new release of Jubilant-backports:

1. Update the __version__ field in jubilant_backports/__init__.py to the new version you want to release.
2. Push up a PR with this change and get it reviewed and merged.
3. Create a new release on GitHub with good release notes. The tag should start with a backports-v, like backports-v1.4.3. Once you've created the release, the publish.yaml workflow will automatically publish it to PyPI.
4. Once the publish workflow has finished, check that the new version appears in the PyPI version history.