When a sufficient number of changes have been accumulated, a new release can be made. There are three places in which a release must be published: github, pypi and bioconda.
Note: panfeed follow's the semantic versioning convention.
In the main branch,
edit the panfeed/__init__.py file to indicate the target release version (e.g. X.X.X).
Then do git add panfeed/__init__.py followed by git commit -m "Version bump".
This step requires the hatchling python package, and that the panfeed directory
does not contain any file that is not part of the git repository; doing this on a fresh
repository is the easiest way to avoid adding unwanted files.
python3 -m build
The above command should create two files in the dist directory:
panfeed-X.X.X-py3-none-any.whl and panfeed-X.X.X.tar.gz.
Apply a tag to identify the release in the git history by doing: git tag X.X.X
(where X.X.X is the target version), followed by git push --tags.
Go to panfeed's release page
, then click on "tags", click on the tag you just pushed, and finally click on
"Make new release from tag".
Fill the various fields following what has been done for previous releases
(e.g. this one),
and add the two files generated by the previous step, then publish the new release.
This step requires the twine python package, an account on pypi.org and
that the account has been added as a maintainer for panfeed.
python3 -m twine upload dist/*
NOTE: it might be that bioconda will automatically create a pull request when panfeed is updated
on pypi. So please wait up to a day before opening a pull request manually.
This step requires that the upload on pypi was successful, as it relies
on the new release to be present there. It also assumes you have a fork of the huge
bioconda-recipes repository and at least ~8GB
of RAM and a linux system. Follow the steps here
to set up your own fork.
Set up a bioconda build environment by creating a bioconda environment in your local conda installation:
mamba create -n bioconda -c conda-forge -c bioconda bioconda-utils grayskull
conda activate bioconda
Then cd into your local copy of bioconda-recipes and update it to the latest upstream commits:
git checkout master
git pull upstream master
git push origin master
Then create a new branch to update the recipe:
git checkout -b update_panfeed_XXX
Then use grayskull to update the recipe:
cd recipes
grayskull pypi panfeed
Then use git add -p to accept/reject changes. Please just add changes to the version and the
sha256 hash of the package file. The other changes are just cosmetic, unless a new subcommand
or dependency has been added. Create a commit with git commit, followed by git stash to get
rid of the unwanted changes.
Test the updated recipes with:
cd ..
bioconda-utils lint --packages panfeed
bioconda-utils build --packages panfeed
If everything looks good, do git push --set-upstream origin update_panfeed_XXX.
The remote server will likely provide a link to directly open a pull request based on these
changes, otherwise visit your fork of bioconda-recipes and there should be a button somewhere to
create a pull request. Follow the indications in the template pull request to provide
information, then wait for the remote tests to complete and add a command saying:
@bioconda-bot please add label. A bioconda maintainer will eventually merge and the new version will
then appear on bioconda.
On panfeed's main branch, edit the panfeed/__init__.py so that it's clear the next version
is a draft (e.g. if latest release is 1.5.0 you could do 1.5.1-dev), followed by
git add panfeed/__init__.py, git commit -m "Dev bump" and git push.