Lovely Pytest Docker
Create simple Pytest_ fixtures for writing integration tests based on Docker
containers. The framework provides a service class to start and stop containers
based Docker Compose. Each single container can be started individually.
Some parts of this package are taken from
https://github.com/AndreLouisCaron/pytest-docker
Usage with Pytest
The docker compose file should contain all desired containers and the ports
should be exposed. In the following example we want to start the app to test
and a SQL database (Crate). Let's assume there is a Dockerfile
for the app
in the same folder as the docker compose file.
version: "3"
services:
app:
build: .
ports:
- "8080"
depends_on:
- "crate"
crate:
image: crate:latest
ports:
- "4200"
In the conftest.py
file we can declare the docker fixtures for each service
we want to be able to start in the tests.
import pytest
@pytest.fixture(scope='session')
def docker_app(docker_services):
docker_services.start('app')
public_port = docker_services.wait_for_service("app", 8080)
url = "http://{docker_services.docker_ip}:{public_port}".format(**locals())
return url
@pytest.fixture(scope='session')
def docker_crate(docker_services):
docker_services.start('crate')
public_port = docker_services.wait_for_service("crate", 4200)
dsn = "{docker_services.docker_ip}:{public_port}".format(**locals())
return dsn
By default the fixture will look for the docker-compose.yml
file in the
tests
subfolder of the path where pytest.ini
resides (or the project's
root directory if no ini file is given - as in the tests example). In many
cases you will want to override the location for the docker compose files. Just
overwrite the docker_compose_files
fixture in your conftest.py
file.
@pytest.fixture(scope='session')
def docker_compose_files(pytestconfig):
"""Get the docker-compose.yml absolute path.
Override this fixture in your tests if you need a custom location.
"""
return [
project_path('docker', 'docker-compose.yml'),
]
In your test file declare the fixtures you want to use:
def test_something(docker_app, docker_crate):
...
...
A working configuration and test example can be found in the tests
folder
of this package.
Execution in Docker Container
It's possible to execute a command inside one of the Docker containers. Use
the exec
method of the docker_services
fixture::
def test_execute(docker_services):
res = docker_services.execute('crate', 'ls', '-a')
Stopping a Docker Container
It's possible to stop single Docker containers. Use
the stop
method of the docker_services
fixture::
def test_stop(docker_services):
# the first argument is the service name of the compose file,
# the following arguments build the command to run
res = docker_services.stop('crate')
Wait for Service
The wait_for_service
method of the service module checks whether the
docker service is really started. By default it makes a HTTP GET request to the
server's /
endpoint. The service will retry to check until a timeout of
30 seconds has passed.
Custom Service Checker
Some services may work differently and require a custom checker.
Create a custom service checker function which receives the IP address and the
port as parameters::
def custom_service_checker(ip_address, port):
return True
In the fixture provide the custom service checker function as check_service
parameter to the wait_for_service
method::
@pytest.fixture(scope='session')
def docker_custom_service(docker_services):
docker_services.start('custom_service')
public_port = docker_services.wait_for_service(
"app",
8080,
check_server=custom_service_checker
)
url = "http://{docker_services.docker_ip}:{public_port}".format(**locals())
return url
To use another request path with the default checker the url_checker
method
can be used to create a check_url
method for another path::
docker_services.wait_for_service(
"app",
8080,
check_server=url_checker('/probe_status'),
)
Run Tests
Tests are held in the tests
directory. Running tests is done via the
pytest package with::
./gradlew pytest
Publish a new Version
Use gradle to build and twine to upload like this:
./gradlew sdist
twine upload build/sdist/lovely-pytest-docker-<version>.tar.gz
Note: twine needs to be installed by othher means (e.g. pip install twine
)