testcontainers-python

testcontainers-python facilitates the use of Docker containers for functional and integration testing. The collection of packages currently supports the following features.

• Testcontainers Core
• Community Modules

Getting Started

>>> from testcontainers.postgres import PostgresContainer
>>> import sqlalchemy

>>> with PostgresContainer("postgres:16") as postgres:
...     psql_url = postgres.get_connection_url()
...     engine = sqlalchemy.create_engine(psql_url)
...     with engine.begin() as connection:
...         version, = connection.execute(sqlalchemy.text("SELECT version()")).fetchone()
>>> version
'PostgreSQL 16...'

The snippet above will spin up the current latest version of a postgres database in a container. The get_connection_url() convenience method returns a sqlalchemy compatible url (using the psycopg2 driver per default) to connect to the database and retrieve the database version.

>>> from testcontainers.postgres import PostgresContainer
>>> import psycopg

>>> with PostgresContainer("postgres:16", driver=None) as postgres:
...     psql_url = postgres.get_connection_url()
...     with psycopg.connect(psql_url) as connection:
...         with connection.cursor() as cursor:
...             version, = cursor.execute("SELECT version()").fetchone()
>>> version
'PostgreSQL 16...'

This snippet does the same, however using a specific version and the driver is set to None, to influence the get_connection_url() convenience method to not include a driver in the URL (e.g. for compatibility with psycopg v3).

Note, that the sqlalchemy and psycopg packages are no longer a dependency of testcontainers[postgres] and not needed to launch the Postgres container. Your project therefore needs to declare a dependency on the used driver and db access methods you use in your code.

By default, Testcontainers will search for the container via the gateway IP. You can manually specify your own IP with the environment variable TESTCONTAINERS_HOST_OVERRIDE.

Installation

The suite of testcontainers packages is available on PyPI, the package can be installed using pip.

Version 4.0.0 onwards we do not support the testcontainers-* packages as it is unsustainable to maintain ownership.

Instead packages can be installed by specifying extras, e.g., pip install testcontainers[postgres].

Please note, that community modules are supported on a best-effort basis and breaking changes DO NOT create major versions in the package. Therefore, only the package core is strictly following SemVer. If your workflow is broken by a minor update, please look at the changelogs for guidance.

Custom Containers

Crafting containers that are based on custom images is supported by the core module. Please check the core documentation for more information.

This allows you to create containers from images that are not part of the modules provided by testcontainers-python.

For common use cases, you can also use the generic containers provided by the testcontainers-generic module. Please check the generic documentation for more information. (example: ServerContainer for running a FastAPI server)

Docker in Docker (DinD)

When trying to launch Testcontainers from within a Docker container, e.g., in continuous integration testing, two things have to be provided:

1. 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.

2. 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.

Private Docker registry

Using a private docker registry requires the DOCKER_AUTH_CONFIG environment variable to be set. official documentation

The value of this variable should be a JSON string containing the authentication information for the registry.

Example:

DOCKER_AUTH_CONFIG='{"auths": {"https://myregistry.com": {"auth": "dXNlcm5hbWU6cGFzc3dvcmQ="}}}'

In order to generate the JSON string, you can use the following command:

echo -n '{"auths": {"<url>": {"auth": "'$(echo -n "<username>:<password>" | base64 -w 0)'"}}}'

Fetching passwords from cloud providers:

ECR_PASSWORD = $(aws ecr get-login-password --region eu-west-1)
GCP_PASSWORD = $(gcloud auth print-access-token)
AZURE_PASSWORD = $(az acr login --name <registry-name> --expose-token --output tsv)

Configuration

Env Variable	Example	Description
TESTCONTAINERS_DOCKER_SOCKET_OVERRIDE	/var/run/docker.sock	Path to Docker’s socket used by ryuk
TESTCONTAINERS_RYUK_PRIVILEGED	false	Run ryuk as a privileged container
TESTCONTAINERS_RYUK_DISABLED	false	Disable ryuk
RYUK_CONTAINER_IMAGE	testcontainers/ryuk:0.8.1	Custom image for ryuk
DOCKER_AUTH_CONFIG	{"auths": {"<url>": {"auth": "<encoded>"}}}	Custom registry auth config

Development and Contributing

We recommend you use a Poetry for development. After having installed poetry, you can run the following snippet to set up your local dev environment.

make install

Package Structure

Testcontainers is a collection of implicit namespace packages to decouple the development of different extensions, e.g., testcontainers[mysql] and testcontainers[postgres] for MySQL and PostgreSQL database containers, respectively.

The folder structure is as follows:

modules
# 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_[some_aspect_for_the_feature].py
        ...
    # README for this feature.
    README.rst

Contributing a New Feature

You want to contribute a new feature or container? Great! - We recommend you first open an issue - Then follow the suggestions from the team - We also have a Pull Request template for new containers!