Closed geoHeil closed 4 months ago
This, and its daughter dagster-libs-feedstock
, are already very complex feedstocks (even by conda-forge standards), and at present all quite manual, as the upstream dependency data is all in setup.py
files which are... complex to reason about (usually manual diff). However, it's a diff against one repo at two tags.
dagster-cloud
(and its novel daster-
dependencies) would bring in two additional repos, and has [extras]
that are probably important to users of those packages. I would recommend, if this is important, to start a new feedstock (potentially with multiple outputs) that ship this independent chain of packages.
As for automation, starting is easy:
conda install grayskull
git clone https://github.com/conda-forge/staged-recipes
cd staged-recipes/recipes
grayskull pypi dagster-cloud
grayskull pypi dagster-cloud-cli
Make a PR, handle requests, off to the races.
Once a feedstock, it will get bot updates, and these can be further simplified with a conda-forge.yml
bot:
inspection: update-all
But when that breaks, maintaining it would be another, semi-weekly task for which I am not super interested in signing up at this time.
understood - thanks for sharing
starting to prepare something here https://github.com/conda-forge/staged-recipes/pull/26708
Comment:
how to move the dagster-cloud pip package over to conda?
is there any automation to ease the process to fill in https://github.com/conda-forge/dagster-feedstock/blob/main/recipe/meta.yaml?