Release guide

Release process - full release of `plotly` package

This is the release process for releasing plotly.py version X.Y.Z, including changelogs, Github release and forum announcement.

Finalize changelog

Review the contents of CHANGELOG.md. We try to follow the keepachangelog guidelines. Make sure the changelog includes the version being published at the top, along with the expected publication date.

Use the Added, Changed, Deprecated, Removed, Fixed, and Security labels for all changes to plotly.py. If the version of plotly.js has been updated, include this as the first Updated entry. Call out any notable changes as sub-bullets (new trace types in particular), and provide a link to the plotly.js CHANGELOG.

Finalize versions

Create a branch git checkout -b release-X.Y.Z from the tip of origin/master.

Manually update the versions to X.Y.Z in the files specified below.

CHANGELOG.md
- update the release date
Commit your changes on the branch:
- git commit -a -m "version changes for vX.Y.Z"

Triggering (and Retriggering) the build

Commit and add this specific tag which versioneer will pick up, and push to Github so that CI will build the release artifacts. This is an atomic push so that CI will read the tag on the commit:
- git tag vX.Y.Z
- git push --atomic origin release-X.Y.Z vX.Y.Z
Create a Github pull request from release-X.Y.Z to master and wait for CI to be green
If something goes wrong below, you'll need to trigger the build process again after a fix. You'll need to commit your changes in the release branch, move the tag and atomically force push:
- git commit ....
- git tag -f vX.Y.Z
- git push --force --atomic origin release-X.Y.Z vX.Y.Z

Download and QA CI Artifacts

The full_build job in the release_build workflow in CircleCI produces a tarball of artifacts output.tgz which you should download and decompress, which will give you a directory called output. The filenames contained within will contain version numbers.

Note: if any of the version numbers are not simply X.Y.Z but include some kind of git hash, then this is a dirty build and you'll need to clean up whatever is dirtying the tree and follow the instructions above to trigger the build again. (That said, you can do QA on dirty builds, you just can't publish them.)

To locally install the PyPI dist, make sure you have an environment with JupyterLab installed (maybe one created with conda create -n condatest python=3.10 jupyter anywidget pandas):

tar xzf output.tgz
pip uninstall plotly
conda uninstall plotly (just in case!)
pip install path/to/output/dist/plotly-X.Y.X-py3-none-any.whl

To locally install the Conda dist (generally do this in a different, clean environment from the one above!):

conda uninstall plotly
pip uninstall plotly (just in case!)
conda install path/to/output/plotly-X.Y.Z.tar.bz2

You'll want to check, in both Lab and Notebook, in a brand new notebook in each so that there is no caching of previous results, that go.Figure() and go.FigureWidget() work without error.

If something is broken, you'll need to fix it and trigger the build again (see above section).

Publishing

Once you're satisfied that things render in Lab and Notebook in Widget and regular mode, you can publish the artifacts. You will need special credentials from Plotly leadership to do this..

Publishing to PyPI:

(plotly_dev) $ cd path/to/output/dist
(plotly_dev) $ twine upload plotly-X.Y.Z*

Publishing to plotly conda channel (make sure you have run conda install anaconda-client to get the anaconda command):

(plotly_dev) $ cd path/to/output
(plotly_dev) $ anaconda upload plotly-X.Y.Z.tar.bz2

Merge the PR and make a Release

Merge the pull request you created above into master
Go to https://github.com/plotly/plotly.py/releases and "Draft a new release"
Enter the vX.Y.Z tag you created already above and make "Release title" the same string as the tag.
Copy the changelog section for this version as the "Describe this release"

Update documentation site

Search for the previous version string in the docs and replace it with the new version string, including but not necessarily limited to the following files:
- doc/python/getting-started.md
- doc/apidoc/conf.py
- doc/requirements.txt
- binder/requirements.txt
doc-prod should already have been merged on a regular basis into master, but start by doing it first if not. Then merge master into doc-prod to deploy the doc related to features in the release.
in a clone of the graphing-library-docs repo:
1. bump the version of Plotly.py in _data/pyversion.json
2. bump the version of Plotly.js with cd _data && python get_plotschema.py <PLOTLY.JS VERSION> fixing any errors that come up.
- If Plotly.js contains any new traces or trace or layout attributes, you'll get a warning “missing key in attributes: <attribute-name>. To resolve, add the attribute to the relevant section in /_data/orderings.json in the position you want it to appear in the reference docs.
1. rebuild the Algolia schema index with ALGOLIA_API_KEY=<key> make update_ref_search
2. Rebuild the Algolia python index with ALGOLIA_API_KEY=<key> make update_python_search
3. Commit and push the changes to master in that repo

Notify Stakeholders

Post an announcement to the Plotly Python forum, with links to the README installation instructions and to the CHANGELOG.
Update the previous announcement to point to this one
Update the Github Release entry and CHANGELOG entry to have the nice title and a link to the announcement
Follow up on issues resolved in this release or forum posts with better answers as of this release

Release process - Release Candidate of `plotly` package

(rough notes for a rough/ad hoc process!)

It's the same process as above except that the X.Y.Z version has a suffix and there are special instructions below for publishing an RC: note that the npm suffix is -rc.1 and the PyPI suffix is rc1. We also don't update the docs with RC information and we inform a limited number of stakeholders.

PyPI RC (no special flags, just the rc1 suffix):

(plotly_dev) $ twine upload dist/plotly-X.Y.Zrc1*

The --tag next part ensures that users won't install this version unless they explicitly ask for the version or for the version with the next tag.

Conda RC:

$ anaconda upload --label test plotly-*.tar.bz2

The --label test part ensures that users won't install this version unless they explicitly ask for the version or for the version with the next tag.

Uh oh!