I noticed that our dependencies were quite out of date, so I updated them. I tested locally with docker compose and Airflow ran, loaded the DAGS and I watched a harvest run for a bit.
Since apache-airflow-providers-docker was listed as a dependency in our pyproject.toml it was pulling all of the latest airflow into our requirements.txt. We really only need airflow and apache-airflow-providers-docker in requirements-dev.txt for testing since they come with the Airflow Docker image we are using.
Specifying apache-airflow-providers-docker in the pyproject.toml added to quite a bit of install churn when we installed with:
$ uv pip install -r requirements.txt
and then did a:
$ uv run bin/get
which would go and fetch the latest airflow and overwrite the version we wanted because uv was looking to the pyproject.toml file.
I updated the README.md to reflect that we need to run the bin commands with uv run.
I noticed that our dependencies were quite out of date, so I updated them. I tested locally with docker compose and Airflow ran, loaded the DAGS and I watched a harvest run for a bit.
Since
apache-airflow-providers-docker
was listed as a dependency in ourpyproject.toml
it was pulling all of the latest airflow into ourrequirements.txt
. We really only needairflow
andapache-airflow-providers-docker
inrequirements-dev.txt
for testing since they come with the Airflow Docker image we are using.Specifying
apache-airflow-providers-docker
in thepyproject.toml
added to quite a bit of install churn when we installed with:and then did a:
which would go and fetch the latest airflow and overwrite the version we wanted because uv was looking to the
pyproject.toml
file.I updated the
README.md
to reflect that we need to run the bin commands withuv run
.