Publish Python packages with Azure Pipelines

Azure DevOps Services | Azure DevOps Server 2022 - Azure DevOps Server 2019

Using Azure Pipelines, developers can publish Python packages to Azure Artifacts feeds, public registries, or store them as pipeline artifacts. This article will guide you through how to:

• Install the prerequisites
• Connect to an Azure Artifacts feed
• Publish Python packages to an Azure Artifacts feed

Prerequisites

• An Azure DevOps organization and a project. Create an organization or a project if you haven't already.

• An Azure Artifacts feed. Create a feed if you don't have one already.

Authenticate with Azure Artifacts

To use twine for publishing your Python packages, you must first authenticate with your Azure Artifacts feed. The TwineAuthenticate task provides twine credentials to a PYPIRC_PATH environment variable. This variable is then used bytwine to facilitate the publishing of your packages directly from your pipeline.

YAML

- task: TwineAuthenticate@1
  inputs:
    artifactFeed: <PROJECT_NAME/FEED_NAME>     ## For an organization-scoped feed, artifactFeed: <FEED_NAME>

Classic

1. Sign in to your Azure DevOps organization, and then navigate to your project.

2. Select Pipelines, select your pipeline definition, and then select Edit.

3. Select the + sign to add a new task. Search for the Python twine upload authenticate task, and then select Add to add it to your pipeline.

4. Select your feed from the My feed name dropdown menu, or select a Feed from external organizations if you want to authenticate with a feed outside your organization.

   

Important

The credentials stored in the PYPIRC_PATH environment variable supersede those in your .ini and .conf files.

If you add multiple TwineAuthenticate tasks at different stages in your pipeline, each additional task execution will extend (not override) the existing PYPIRC_PATH environment variable.

Publish Python packages to an Azure Artifacts feed

YAML

- script: |
    pip install build
    pip install twine
  displayName: 'Install build and twine'

- script: |
    python -m build -w
  displayName: 'Python build'

- task: TwineAuthenticate@1
  inputs:
    artifactFeed: <PROJECT_NAME/FEED_NAME>
  displayName: 'Twine Authenticate'

- script: |
    python -m twine upload -r <FEED_NAME> --config-file $(PYPIRC_PATH) dist/*.whl
  displayName: 'Upload to feed'

Classic

1. Sign in to your Azure DevOps organization, and then navigate to your project.

2. Select Pipelines, select your pipeline definition, and then select Edit.

3. Select the + sign to add a new task. Search for the Use Python version task, and then select Add to add it to your pipeline. In this example we're setting the Version spec to >= 3.7.

4. Select the + sign to add a new task. Search for the Command line task, and then select Add to add it to your pipeline. Paste the following commands in the Script textbox to install build and twine on your agent:

   pip install build
   pip install twine
   

5. Select the + sign to add a new task. Search for the Command line task, and then select Add to add it to your pipeline. Paste the following command in the Script textbox to build your project:

   python -m build -w
   

6. Select the + sign to add a new task. Search for the Twine Authenticate task, and then select Add to add it to your pipeline. Select your feed from the My feed name dropdown menu.

7. Select the + sign to add a new task. Search for the Command line task, and then select Add to add it to your pipeline. Paste the following command in the Script textbox to publish your package to your feed, replacing the placeholder with your feed name:

   python -m twine upload -r <FEED_NAME> --config-file $(PYPIRC_PATH) dist/*.whl
   

   

Note

To publish your packages to a feed using Azure Pipelines, both the Project Collection Build Service and your project's Build Service identities must have the Feed Publisher (Contributor) role assigned in your feed settings. See Manage permissions for details.

Related articles

• Publish and download Python packages CLI
• Search for packages in upstream sources
• Configure permissions