Skip to content

Completely refactor the fulltext operations (#1093) #1

Completely refactor the fulltext operations (#1093)

Completely refactor the fulltext operations (#1093) #1

name: Docker build and publish
on:
push:
branches: [ master ]
pull_request:
branches: [ master ]
concurrency:
# When this is not a pull request, then we want all the docker containers to be pushed, we therefore
# directly fall back to the commit hash which will be distinct for each push to master.
group: '${{ github.workflow }} @ ${{ github.event.pull_request.head.label || github.sha}}'
cancel-in-progress: true
jobs:
docker:
runs-on: ubuntu-20.04
steps:
- name: Checkout
uses: actions/checkout@v4
with:
submodules: 'recursive'
- name: Set up QEMU
uses: docker/setup-qemu-action@v3
- name: Set up Docker Buildx
uses: docker/setup-buildx-action@v3
- name: Login to Docker Hub
if: github.event_name != 'pull_request'
uses: docker/login-action@v3
with:
username: ${{ secrets.DOCKERHUB_USERNAME }}
password: ${{ secrets.DOCKERHUB_TOKEN }}
- name: Get short sha
id: sha
run: echo "sha_short=$(git rev-parse --short HEAD)" >> $GITHUB_OUTPUT
- name: Get PR number
id: pr
run: echo "pr_num=$(git log --format=%s -n 1 | sed -nr 's/.*\(\#([0-9]+)\)/\1/p')" >> $GITHUB_OUTPUT
- name: Build X-86-64
uses: docker/build-push-action@v5
with:
context: .
# Docker multiplatform images require an entry in their manifests. If this
# entry is missing, then the images for each platform are considered seperate.
# If this action runs on a single runner (i.e. the attribute `platforms`
# contains a list of all platforms) then the manifest is configured automatically.
# If the build and push is split over multiple runners, then the action requires an additional `merge`
# job, which merges the images for each platform into a single multiplatform image.
# References:
# https://docs.docker.com/build/ci/github-actions/multi-platform/#distribute-build-across-multiple-runners
# https://www.docker.com/blog/multi-arch-build-and-images-the-simple-way/
platforms: linux/amd64
# Only export to the local docker for testing.
load: true
tags: adfreiburg/qlever:test
- name: E2E in Docker
run: |
sudo mkdir ${{github.workspace}}/e2e_data
sudo chmod a+rwx ${{github.workspace}}/e2e_data
sudo docker run -i --rm -v "${{github.workspace}}/e2e_data:/app/e2e_data/" --entrypoint e2e/e2e.sh adfreiburg/qlever:test
- name: Build ARM-64
# The cross-compilation to ARM64 takes a long time, so we typically
# only perform it on the pushes to master when we actually need that
# container. Should the building of the container however fail in the
# future, then we can set this to `true`. Then the ARM64 image will also
# be built on pull requests which allows for debugging without changing
# the master branch.
if: false
uses: docker/build-push-action@v5
with:
context: .
platforms: linux/arm64
# Only build, don't run.
load: false
push: false
- name: Build and push
if: github.event_name != 'pull_request'
uses: docker/build-push-action@v5
with:
context: .
platforms: linux/amd64,linux/arm64
# Push to dockerhub, reuse the cached steps from the previous build.
push: true
# If this is a push on master, publish with short commit sha
# else use the ref name, which has to be the tag in this case.
# We have to explicitly add the "qlever:latest" tag for it to work correctly,
# see e.g. https://stackoverflow.com/questions/27643017/do-i-need-to-manually-tag-latest-when-pushing-to-docker-public-repository
tags: >
adfreiburg/qlever:latest,
adfreiburg/qlever:${{ github.ref_name == 'master' && format('pr-{0}', steps.pr.outputs.pr_num) || github.ref_name }},
adfreiburg/qlever:commit-${{ steps.sha.outputs.sha_short }},