testcontainers-python facilitates the use of Docker containers for functional and integration testing. The collection of packages currently supports the following features.
.. toctree:: core/README arangodb/README azurite/README clickhouse/README compose/README elasticsearch/README google/README kafka/README keycloak/README localstack/README minio/README mongodb/README mssql/README mysql/README neo4j/README nginx/README opensearch/README oracle/README postgres/README rabbitmq/README redis/README selenium/README
>>> from testcontainers.postgres import PostgresContainer >>> import sqlalchemy >>> with PostgresContainer("postgres:9.5") as postgres: ... engine = sqlalchemy.create_engine(postgres.get_connection_url()) ... result = engine.execute("select version()") ... version, = result.fetchone() >>> version 'PostgreSQL 9.5...'
The snippet above will spin up a postgres database in a container. The
get_connection_url() convenience method returns a
sqlalchemy compatible url we use to connect to the database and retrieve the database version.
The suite of testcontainers packages is available on PyPI, and individual packages can be installed using
pip. We recommend installing the package you need by running
pip install testcontainers-<feature>, e.g.,
pip install testcontainers-postgres.
For backwards compatibility, packages can also be installed by specifying extras, e.g.,
pip install testcontainers[postgres].
Docker in Docker (DinD)
When trying to launch a testcontainer from within a Docker container, e.g., in continuous integration testing, two things have to be provided:
- The container has to provide a docker client installation. Either use an image that has docker pre-installed (e.g. the official docker images) or install the client from within the Dockerfile specification.
- The container has to have access to the docker daemon which can be achieved by mounting /var/run/docker.sock or setting the DOCKER_HOST environment variable as part of your docker run command.
Development and Contributing
We recommend you use a virtual environment for development (
python>=3.7 is required). After setting up your virtual environment, you can install all dependencies and test the installation by running the following snippet.
pip install -r requirements/[your python version].txt pytest -s
Testcontainers is a collection of implicit namespace packages to decouple the development of different extensions, e.g.,
testcontainers-postgres for MySQL and PostgreSQL database containers, respectively. The folder structure is as follows.
# One folder per feature. [feature name] # Folder without __init__.py for implicit namespace packages. testcontainers # Implementation as namespace package with __init__.py. [feature name] __init__.py # Other files for this ... # Tests for the feature. tests test_[feature_name].py ... # README for this feature. README.rst # Setup script for this feature. setup.py
Contributing a New Feature
You want to contribute a new feature or container? Great! You can do that in six steps.
- Create a new feature directory and populate it with the [package structure]_ as described above. Copying one of the existing features is likely the best way to get started.
- Implement the new feature (typically in
__init__.py) and corresponding tests.
- Add a line
-e file:[feature name]to
make requirements. This command will find any new requirements and generate lock files to ensure reproducible builds (see the pip-tools documentation for details). Then run
pip install -r requirements/[your python version].txtto install the new requirements.
- Update the feature
README.rstand add it to the table of contents (
toctreedirective) in the top-level
- Add a line
[feature name]to the list of components in the GitHub Action workflow in
.github/workflows/main.ymlto run tests, build, and publish your package when pushed to the
- Rebase your development branch on
masterinto your development branch).