Closed kaczmarj closed 1 year ago
here are the things i did to get started.
First, I forked https://github.com/conda-forge/staged-recipes. Then I cloned that fork onto my computer and ran the following code to begin adding wsinfer.
git clone git@github.com:kaczmarj/staged-recipes.git
cd staged-recipes
git switch --create=add/wsinfer
cd recipes/
mkdir wsinfer
cd wsinfer/
cp ../example/meta.yaml .
at this point, I need to edit meta.yaml
.
after we push a new version of wsinfer to pypi, we can use grayskull to generate the meta.yaml file. https://github.com/conda/grayskull
related to https://github.com/conda-forge/staged-recipes/pull/23479 -- this proposes adding wsinfer-zoo to conda-forge. once that is completed, we can add wsinfer.
fixed by https://github.com/conda-forge/staged-recipes/pull/23487
once that is merged, wsinfer can be installed via conda-forge
wsinfer can be installed via conda now!
conda install -c conda-forge wsinfer
Making wsinfer available via conda can make installation easier for some users. See https://conda-forge.org/docs/maintainer/adding_pkgs.html for how to contribute a package to conda-forge.
My concern at this point is that we depend on opencv-python-headless (set in pyproject.toml). This package is available in conda-forge by a different name --
opencv
. We will have to make sure we specify this properly.