Releasing Kubeflow Pipelines

WIP: this document is still incomplete.

Common Prerequisites

OS: Linux (MacOS not supported yet due to different behavior of sed)
Permissions needed
- Can create a branch in github.com/kubeflow/pipelines
Tools that should be in your $PATH
- jq 1.6 https://stedolan.github.io/jq/download/
- yq https://github.com/mikefarah/yq/releases/tag/3.3.0
- jdk 8
Preparations
1. Clone github.com/kubeflow/pipelines repo into $KFP_REPO
2. cd $KFP_REPO

Choose a good commit on master branch with commit hash as $COMMIT_SHA
Choose the next release branch's $MINOR_VERSION in format x.y, e.g. 1.0, 1.1...
Make a release branch of format release-$MINOR_VERSION, e.g. release-1.0, release-1.1. Branch from the commit and push to kubeflow pipelines upstream repo.
```
git checkout $COMMIT_SHA
git checkout -b release-$MINOR_VERSION
git push upstream HEAD
```

Choose the release's complete $VERSION following semantic versioning, e.g.
- 1.0.0-rc.1
- 1.0.0-rc.2
- 1.0.0
- 1.0.1
- ...
Update all version refs in release branch by
```
./hack/release.sh $VERSION release-$MINOR_VERSION
```
It will prompt you whether to push it to release branch. Press y and hit Enter.

Note, the script will clone kubeflow/pipelines repo into a temporary location on your computer, make those changes and attempt to push to upstream, so that it won't interfere with your current git repo.

TODO: this script should also regenerate
- changelog
- python api client
Wait and make sure the cloudbuild job that builds all images in gcr.io/ml-pipeline-test succeeded for above commit. Then submit the second cloudbuild job that copies these images to gcr.io/ml-pipeline.

TODO: we should have an automation KFP cluster, and the waiting and submiting second cloudbuild task should be automated.
Release kfp-server-api and kfp python packages on pypi.
Create a github release using $VERSION git tag, fill in the description.