Release Process#
This is a guide to the process of creating a new release, and is meant for the maintainers of Raster Vision.
Minor or Major Version Release#
It’s a good idea to update any major dependencies before the release.
Update
THIRD_PARTY_LICENSES.txt
:Install third_party_license_file_generator:
pip install third_party_license_file_generator
Update
third_party_license_file_generator_overrides.yml
, if necessary.Collect all
requirements.txt
files into a single file,requirements-all.txt
:cat `find **/requirements.txt` requirements.txt requirements-dev.txt > requirements-all.txt;
Generate new
THIRD_PARTY_LICENSES.txt
file:python -m third_party_license_file_generator \ -r requirements-all.txt \ -p `which python` \ -o THIRD_PARTY_LICENSES.txt \ -l third_party_license_file_generator_overrides.yml;
Test examples:
Checkout the
master
branch, re-build the docker image (docker/build
), and push it to ECR (docker/ecr_publish
).Follow the instructions in this README to do the following:
Run all Examples and check that evaluation metrics are close to the scores from the last release. (For each example, there should be a link to a JSON file with the evaluation metrics from the last release.) This stage often uncovers bugs, and is the most time consuming part of the release process.
Collect all model bundles, and check that they work with the
predict
command and sanity check output in QGIS.Update the Model Zoo by uploading model bundles and sample images to the right place on S3. If you use the
collect
command (see), you should be able to sync thecollect_dir
tos3://azavea-research-public-data/raster-vision/examples/model-zoo-<version>
.Screenshot the outputs of the
compare
command (for each example) and include them in the PR described below.
Test notebooks:
Update the tutorial notebooks that use models from the model zoo so that they use the latest version.
Execute all tutorial notebooks and make sure they work correctly. Do not commit output changes unless code behavior has changed.
Test/update docs:
Update the docs if needed. See the docs README for instructions.
Update tiny_spacenet.py if needed and ensure the line numbers in every
literalinclude
of that file are correct. Tip: you can find all instances by searching the repo using the regex:\.\. literalinclude:: .+tiny_spacenet\.py$
.Test Installation and Quickstart instructions and make sure they work.
Update the the changelog, and point out API changes.
Fix any broken badges on the GitHub repo readme.
Update the version number. This occurs in several places, so it’s best to do this with a find-and-replace over the entire repo.
Make a PR to the
master
branch with the preceding updates. In the PR, there should be a link to preview the docs. Check that they are building and look correct.Make a git branch with the version as the name, and push to GitHub.
Ensure that the docs are building correctly for the new version branch on readthedocs. You will need to have admin access on your RTD account. Once the branch is building successfully, Under Versions -> Activate a Version, you can activate the version to add it to the sidebar of the docs for the latest version. (This might require manually triggering a rebuild of the docs.) Then, under Admin -> Advanced Settings, change the default version to the new version.
GitHub Actions is supposed to publish an image whenever there is a push to a branch with a version number as the name. If this doesn’t work or you want to publish it immediately, then you can manually make a Docker image for the new version and push to Quay. For this you will need an account on Quay.io under the Azavea organization.
./docker/build docker login quay.io docker tag raster-vision-pytorch:latest quay.io/azavea/raster-vision:pytorch-<version> docker push quay.io/azavea/raster-vision:pytorch-<version>
Make a GitHub tag and release using the previous release as a template.
Remove artifacts from previous builds. From the repo root:
rm -rf build/ dist/ *.egg-info rm -rf rastervision_*/build rastervision_*/dist rastervision_*/*.egg-info
Publish all packages to PyPI. This step requires twine which you can install with
pip install twine
To store settings for PyPI you can set up a
~/.pypirc
file containing:[pypi] username=__token__ password=<password> [testpypi] username=__token__ password=<password>
Once packages are published they cannot be changed, so be careful. (It’s possible to practice using TestPyPI.) Navigate to the repo’s root directory on your local filesystem. With the version branch checked out, run the following scripts to build packages and publish to PyPI.
Build:
scripts/pypi_build
Publish to TestPyPI:
scripts/pypi_publish --test
You can then test it with
pip
like so:pip install --index-url https://test.pypi.org/simple/ rastervision
Finally, if everything looks okay, publish to Pypi:
scripts/pypi_publish
Announce the new release in our forum, and with a blog post if it’s a big release.
Make a PR to the master branch that updates the version number to the next development version,
X.Y.Z-dev
. For example, if the last release was0.20.1
, update the version to0.20.2-dev
.
Patch Release#
This describes how to create a new patch release (AKA a bug-fix release), using an increment from 0.8.0 to 0.8.1 as an example. This assumes that there is already a branch for a minor release called 0.8
.
Backport changes to the
0.8
branch. To create a patch release (version 0.8.1), we need to backport all the commits on themaster
branch that have been added since the last patch release onto the0.8
branch. To do this:Create a new branch from the
0.8
branch. Let’s call itbackport
.Cherry-pick each commit that we want to include from the
master
branch onto thebackport
branch.Make a PR against the
0.8
branch from thebackport
branch. The title of the PR should start with[BACKPORT]
.
Update changelog and version on the
0.8
branch. Make and merge a PR against0.8
(but notmaster
) that adds a changelog for the new release and increments the version to0.8.1
throughout the repo. Wait for the0.8
branch to be built by GitHub Actions and the0.8
Docker images to be published to Quay. If that is successful, we can proceed to the next steps of actually publishing a release.Publish the new version to PyPI. Follow the same instructions for PyPI as listed above for minor/major version releases.
Using the GitHub UI, make a new release. Use
v0.8.1
as the tag, and the0.8
branch as the target.Update changelog and version on the
master
branch. Make and merge a PR againstmaster
thatincludes the cherry-picked commit that updates the changelog for
0.8.1
andincrements the version to
0.8.2-dev
throughout the repo.