Thank you for considering contributing to dlt! We appreciate your help in making dlt better. This document will guide you through the process of contributing to the project.
- Getting Started
- Submitting Changes
- Linting
- Testing
- Local Development
- Publishing (Maintainers Only)
- Resources
To get started, follow these steps:
- Fork the
dltrepository and clone it to your local machine. - Install
poetrywithmake install-poetry(or follow the official instructions). - Run
make devto install all dependencies including dev ones. - Start working in the
poetryshell by executingpoetry shell.
When you're ready to contribute, follow these steps:
- Create an issue describing the feature, bug fix, or improvement you'd like to make.
- Create a new branch in your forked repository for your changes.
- Write your code and tests.
- Lint your code by running
make lintand test common modules withmake test-common. - If you're working on destination code, contact us to get access to test destinations.
- Create a pull request targeting the devel branch of the main repository.
Note: for some special cases, you'd need to contact us to create a branch in this repository (not fork). See below.
We use devel (which is our default Github branch) to prepare a next release of dlt. We accept all regular contributions there (including most of the bugfixes).
We use master branch for hot fixes (including documentation) that needs to be released out of normal schedule.
On the release day, devel branch is merged into master. All releases of dlt happen only from the master.
We'll fix critical bugs and release dlt our of the schedule. Follow the regular procedure, but make your PR against master branch. Please ping us on Slack if you do it.
We enable our CI to run tests for contributions from forks. All the tests are run, but not all destinations are available due to credentials. Currently
only the duckdb and postgres are available to forks.
In case you submit a new destination or make changes to a destination that require credentials (so Bigquery, Snowflake, buckets etc.) you should contact us so we can add you as contributor. Then you should make a PR directly to the dlt repo.
dlt uses mypy and flake8 with several plugins for linting.
dlt uses pytest for testing.
To test common components (which don't require external resources), run make test-common.
To test local destinations (duckdb and postgres), run make test-load-local.
To test external destinations use make test. You will need following external resources
BigQueryprojectRedshiftclusterPostgresinstance. You can find a docker compose for postgres instance here. When run the instance is configured to work with the tests.
cd tests/load/postgres/
docker-compose up --build -dSee tests/.example.env for the expected environment variables and command line example to run the tests. Then create tests/.env from it. You configure the tests as you would configure the dlt pipeline.
We'll provide you with access to the resources above if you wish to test locally.
Use Python 3.8 for development, as it's the lowest supported version for dlt. You'll need distutils and venv. You may also use pyenv, as suggested by poetry.
This section is intended for project maintainers who have the necessary permissions to manage the project's versioning and publish new releases. If you're a contributor, you can skip this section.
Please read how we version the library first.
The source of truth of the current version is is pyproject.toml, and we use poetry to manage it.
Before publishing a new release, make sure to bump the project's version accordingly:
- Check out the devel branch.
- Use
poetry version patchto increase the patch version - Run
make build-libraryto apply the changes to the project. - Create a new branch, and submit the PR to devel. Go through standard process to merge it.
- Create a merge PR from
develtomasterand merge it with a merge commit.
- Check out the master branch
- Use
poetry version patchto increase the patch version - Run
make build-libraryto apply the changes to the project. - Create a new branch, and submit the PR to master and merge it.
Occasionally we may release an alpha version directly from the branch.
- Check out the devel branch
- Use
poetry version prereleaseto increase the alpha version - Run
make build-libraryto apply the changes to the project. - Create a new branch, and submit the PR to devel and merge it.
Once the version has been bumped, follow these steps to publish the new release to PyPI:
- Ensure that you are on the master branch and have the latest code that has passed all tests on CI.
- Verify the current version with
poetry version. - Obtain a PyPI access token and configure it with
poetry config pypi-token.pypi your-api-token. - Run
make publish-libraryto publish the new version. - Create a release on GitHub, using the version and git tag as the release name.
If you have any questions or need help, don't hesitate to reach out to us. We're here to help you succeed in contributing to dlt. Happy coding!