Skip to content

Replace git checkout with datalad checkout in action workflows wherever appropriate #24

Replace git checkout with datalad checkout in action workflows wherever appropriate

Replace git checkout with datalad checkout in action workflows wherever appropriate #24

Workflow file for this run

name: Add a new model
# Trigger the workflow on pull request
on:
issues:
types: [ assigned, labeled ]
jobs:
create_new_branch:
if: ${{ startsWith(github.event.issue.title, 'New Model:') && github.event.label.name != 'failed' }}
runs-on: ubuntu-latest
outputs:
BRANCHNAME: ${{ steps.branch.outputs.branchName }}
# All the issue form data
MODELPATH: ${{ steps.get_path.outputs.path }}
WEIGHTS: ${{ steps.get_weights.outputs.weights }}
DOCKER: ${{ steps.get_docker.outputs.docker }}
PYTHON: ${{ steps.python_scripts.outputs.pythons }}
MODELINFO: ${{ env.model_info }}
SAMPLEDATA: ${{ steps.get_data.outputs.sample_data }}
PYTHONS: ${{ env.pythons }}
CONFIG: ${{ steps.get_model_config.outputs.model_config }}
TESTCOMMAND: ${{ steps.get_test_command.outputs.test_command }}
steps:
# This will automatically create a new branch from this issue, using custom config at /.github/issue-branch.yml 🟢
- name: Create Issue Branch
id: branch
uses: robvanderleek/create-issue-branch@main
env:
GITHUB_TOKEN: ${{ secrets.GH_TOKEN }}
# Set up Python 🟢
- name: Set up Python
uses: actions/setup-python@v4
with:
python-version: 3.11
# Checkout the repository to the GitHub Actions runner 🟢
- uses: actions/checkout@v4
with:
ref: ${{ steps.branch.outputs.branchName }}
fetch-depth: 0
# Scrape the info from the issue 🟢
- uses: stefanbuck/github-issue-parser@v3
id: issue-parser
with:
template-path: .github/ISSUE_TEMPLATE/addModel.yml # optional but recommended
# Scrape info from the form 🟢
- name: Get the Path
id: get_path
run: echo "path=${{ steps.issue-parser.outputs.issueparser_path }}" >> $GITHUB_OUTPUT
- name: Get the Weights
id: get_weights
run: echo "weights=${{ steps.issue-parser.outputs.issueparser_weights }}" >> $GITHUB_OUTPUT
- name: Get docker folder
id: get_docker
run: echo "docker=${{ steps.issue-parser.outputs.issueparser_docker }}" >> $GITHUB_OUTPUT
- name: Get Python Scripts
id: python_scripts
run: |
echo "pythons<<EOF" >> $GITHUB_ENV
echo "${{ steps.issue-parser.outputs.issueparser_python-scripts }}" >> $GITHUB_ENV
echo "EOF" >> $GITHUB_ENV
- name: Get Model Info
id: get_model_info
run: |
echo "model_info<<EOF" >> $GITHUB_ENV
echo "${{ steps.issue-parser.outputs.issueparser_model_info }}" >> $GITHUB_ENV
echo "EOF" >> $GITHUB_ENV
- name: Get Test Command
id: get_test_command
run: |
echo "test_command=${{ steps.issue-parser.outputs.issueparser_testCommand }}" >> $GITHUB_OUTPUT
- name: Get Sample Data
id: get_data
run: echo "sample_data=${{ steps.issue-parser.outputs.issueparser_sample-data }}" >> $GITHUB_OUTPUT
- name: Get Config
id: get_model_config
run: |
echo "model_config=${{ steps.issue-parser.outputs.issueparser_modelConfig }}" >> $GITHUB_OUTPUT
push-model:
needs: create_new_branch
runs-on: ubuntu-latest
steps:
# Checkout the repository to the GitHub Actions runner to the new branch created for the issue 🟢
- uses: actions/checkout@v4
if: needs.create_new_branch.outputs.BRANCHNAME == ''
with:
ref: issue-${{ github.event.issue.number }}
fetch-depth: 0
# If branchName is empty use issue number, else, use the branchName 🟢
- uses: actions/checkout@v4
if: needs.create_new_branch.outputs.BRANCHNAME != ''
with:
ref: ${{ needs.create_new_branch.outputs.BRANCHNAME }}
fetch-depth: 0
# Install svn to clone subdir of github repos 🟢
- name: Install svn
run: sudo apt-get install subversion
# Get svn url for exporting docker 🟢
- name: Clone docker folder
run: |
url="${{ needs.create_new_branch.outputs.DOCKER }}"
svn_url=$(echo "$url" | sed -E 's|/tree/[^/]+|/trunk|; s|/blob/[^/]+|/trunk|')
svn export --force $svn_url ./${{ needs.create_new_branch.outputs.MODELPATH }}/docker
# Get svn url for exporting 🟢
- name: Generate Python SVN URLs
id: generate_urls
run: |
echo "pythons2<<EOF" >> $GITHUB_ENV
python ./.github/workflows/getPythonScripts.py >> $GITHUB_ENV
echo "EOF" >> $GITHUB_ENV
env:
pythons: ${{ needs.create_new_branch.outputs.PYTHONS }}
# Export the urls/clone the scripts 🟢
- name: Clone python scripts
run: |
cd ./${{ needs.create_new_branch.outputs.MODELPATH }}
svn_urls="${{ env.pythons2 }}"
for svn_url in $svn_urls; do
svn export --force $svn_url
done
# Get svn urls for exporting card and spec urls 🟢
- name: Generate Model Info SVN URLs
id: generate_model_info_urls
run: |
echo "model_info2<<EOF" >> $GITHUB_ENV
python ./.github/workflows/getModelInfo.py >> $GITHUB_ENV
echo "EOF" >> $GITHUB_ENV
env:
model_info: ${{ needs.create_new_branch.outputs.MODELINFO }}
# Export the urls/clone the model info 🟢
- name: Clone model info
run: |
model_info="${{ env.model_info2 }}"
for svn_url in $model_info; do
svn export --force $svn_url ./${{ needs.create_new_branch.outputs.MODELPATH }}
done
# Clone config files
- name: Config file clone
# Only run if needs.create_new_branch.outputs.CONFIG is not empty
if: ${{ needs.create_new_branch.outputs.CONFIG != '' }}
run: |
mkdir ./${{ needs.create_new_branch.outputs.MODELPATH }}/config/
url="${{ needs.create_new_branch.outputs.CONFIG }}"
svn_url=$(echo "$url" | sed -E 's|/tree/[^/]+|/trunk|; s|/blob/[^/]+|/trunk|')
svn export --force $svn_url ./${{ needs.create_new_branch.outputs.MODELPATH }}/config
# Commit the new created files and folders to the branch needs.create_new_branch.outputs.BRANCHNAME 🟢
- name: Commit and Push the new files
run: |
git config --global user.name "trained_models"
git config --global user.email "trained_models"
git add ${{ needs.create_new_branch.outputs.MODELPATH }}
git commit -m "Added model files"
git push origin ${{ needs.create_new_branch.outputs.BRANCHNAME }}
build:
needs: [create_new_branch, push-model]
runs-on: ubuntu-latest
steps:
# Checkout the repository to the GitHub Actions runner 🟢
- uses: actions/checkout@v4
if: needs.create_new_branch.outputs.BRANCHNAME == ''
with:
ref: issue-${{ github.event.issue.number }}
fetch-depth: 0
# If branchName is empty use issue number, else, use the branchName 🟢
- uses: actions/checkout@v4
if: needs.create_new_branch.outputs.BRANCHNAME != ''
with:
ref: ${{ needs.create_new_branch.outputs.BRANCHNAME }}
fetch-depth: 0
# Set up Python
- name: Set up Python
uses: actions/setup-python@v4
with:
python-version: 3.11
# Install yaml
- name: Install yaml and oyaml
run: |
pip install pyyaml
pip install oyaml
# Install linkml
- name: Install LinkML
run: pip install linkml
# Create model card and spec.yaml file
- name: Validate model card and spec files
run: |
echo "## Model Card and Spec Validation :white_check_mark:" >> $GITHUB_STEP_SUMMARY
echo "Model Card and spec yaml files are being validated here with LinkML Schema" >> $GITHUB_STEP_SUMMARY
python ./.github/workflows/create_model_card_and_spec.py
env:
model_path: ${{ needs.create_new_branch.outputs.MODELPATH }}
# Update the model's spec.yaml file
- name: Update yaml file
run: |
model_name=$(echo "${{ needs.create_new_branch.outputs.MODELPATH }}" | awk -F '/' '{print $(NF-1)}')
python ./.github/workflows/update_yaml_info.py ${{ needs.create_new_branch.outputs.MODELPATH }}/docker $model_name
# Commit the changes (spec.yaml file and model card) 🟢
- name: Commit changes
run: |
git config --global user.name "trained_models"
git config --global user.email "trained_models"
git add ${{ needs.create_new_branch.outputs.MODELPATH }}/spec.yaml
git commit -m "Updated spec.yaml"
git push origin ${{ needs.create_new_branch.outputs.BRANCHNAME }}
build-docker:
needs: [create_new_branch, push-model, build]
runs-on: ubuntu-latest
outputs:
IMAGENAME: ${{ steps.set_image_name.outputs.image_name }}
MODELNAME: ${{ steps.set_image_name.outputs.model_name }}
steps:
# Checkout the repository to the GitHub Actions runner 🟢
- uses: actions/checkout@v4
if: needs.create_new_branch.outputs.BRANCHNAME == ''
with:
ref: issue-${{ github.event.issue.number }}
fetch-depth: 0
# If branchName is empty use issue number, else, use the branchName 🟢
- uses: actions/checkout@v4
if: needs.create_new_branch.outputs.BRANCHNAME != ''
with:
ref: ${{ needs.create_new_branch.outputs.BRANCHNAME }}
fetch-depth: 0
# Get image name 🟢
- name: Set docker image name
id: set_image_name
run: |
full_path="${{ needs.create_new_branch.outputs.MODELPATH }}"
model_name=$(echo "$full_path" | awk -F '/' '{print $(NF-1)}')
echo "image_name=neuronets/$model_name" >> $GITHUB_OUTPUT
echo "model_name=$model_name" >> $GITHUB_OUTPUT
push-weights:
needs: [create_new_branch, push-model, build, build-docker]
runs-on: ubuntu-latest
steps:
# Checkout the repository to the GitHub Actions runner to the new branch created for the issue 🟢
- uses: actions/checkout@v4
if: needs.create_new_branch.outputs.BRANCHNAME == ''
with:
ref: issue-${{ github.event.issue.number }}
fetch-depth: 0
# If branchName is empty use issue number, else, use the branchName 🟢
- uses: actions/checkout@v4
if: needs.create_new_branch.outputs.BRANCHNAME != ''
with:
ref: ${{ needs.create_new_branch.outputs.BRANCHNAME }}
fetch-depth: 0
# Set up Python
- name: Set up Python
uses: actions/setup-python@v4
with:
python-version: 3.11
# Install datalad 🟢
- name: Install Datalad
id: testing
run: |
sudo apt-get install datalad
python3 -m pip install datalad-osf
pip install requests
git config --global user.name "trained_models"
git config --global user.email "trained_models"
# Get file extensions for the weights and sample data
# - name: Get extensions for weights
# id: get_extensions_weights
# run: |
# echo "weights_ext<<EOF" >> $GITHUB_OUTPUT
# python ./.github/workflows/getFileExtension.py "${{ needs.create_new_branch.outputs.WEIGHTS }}" >> $GITHUB_OUTPUT
# echo "EOF" >> $GITHUB_OUTPUT
# - name: Get extensions for sample data
# id: get_extensions_sample_data
# run: |
# echo "sample_ext<<EOF" >> $GITHUB_OUTPUT
# python ./.github/workflows/getFileExtension.py "${{ needs.create_new_branch.outputs.SAMPLEDATA }}" >> $GITHUB_OUTPUT
# echo "EOF" >> $GITHUB_OUTPUT
#Git annex addurl the weights
- name: Git-annex/Datalad add the weights
env:
OSF_TOKEN: ${{ secrets.OSF_TOKEN }}
run: |
# datalad siblings
mkdir ./${{ needs.create_new_branch.outputs.MODELPATH }}/weights
cd ./${{ needs.create_new_branch.outputs.MODELPATH }}/weights
# datalad download-url --overwrite -m "Added Weights" -O ./${{ needs.create_new_branch.outputs.MODELPATH }}/weights/best_model.${{ steps.get_extensions_weights.outputs.weights_ext }} "${{ needs.create_new_branch.outputs.WEIGHTS }}"
datalad download-url --overwrite -m "Added Sample Dataset" "${{ needs.create_new_branch.outputs.WEIGHTS }}"
datalad save . -m "Added model weights"
datalad push --to osf-storage
datalad push --to origin
- name: Git-annex/Datalad add the sample datasets
env:
OSF_TOKEN: ${{ secrets.OSF_TOKEN }}
run: |
mkdir ./${{ needs.create_new_branch.outputs.MODELPATH }}/example-data
cd ./${{ needs.create_new_branch.outputs.MODELPATH }}/example-data
# datalad download-url --overwrite -m "Added Sample Dataset" -O ./${{needs.create_new_branch.outputs.MODELPATH}}/example-data/sample.${{ steps.get_extensions_sample_data.outputs.sample_ext }} "${{needs.create_new_branch.outputs.SAMPLEDATA}}"
datalad download-url --overwrite -m "Added Sample Dataset" "${{ needs.create_new_branch.outputs.SAMPLEDATA }}"
datalad save . -m "Added sample data"
datalad push --to osf-storage
datalad push --to origin
start-runner:
needs: [create_new_branch, push-model, build, build-docker, push-weights]
runs-on: ubuntu-latest
outputs:
label: ${{ steps.start-ec2-runner.outputs.label }}
ec2-instance-id: ${{ steps.start-ec2-runner.outputs.ec2-instance-id }}
steps:
- name: Configure AWS credentials
uses: aws-actions/configure-aws-credentials@v1
with:
aws-access-key-id: ${{ secrets.AWS_KEY_ID }}
aws-secret-access-key: ${{ secrets.AWS_KEY_SECRET }}
aws-region: ${{ vars.AWS_REGION }}
- name: Start EC2 runner
id: start-ec2-runner
uses: machulav/ec2-github-runner@v2
with:
mode: start
github-token: ${{ secrets.GH_TOKEN }}
ec2-image-id: ${{ vars.AWS_IMAGE_ID }}
ec2-instance-type: ${{ vars.AWS_INSTANCE_TYPE }}
subnet-id: ${{ vars.AWS_SUBNET }}
security-group-id: ${{ vars.AWS_SECURITY_GROUP }}
test-model:
name: Do the job on the runner
needs: [create_new_branch, push-model, build, build-docker, push-weights, start-runner] # required to start the main job when the runner is ready
runs-on: ${{ needs.start-runner.outputs.label }} # run the job on the newly created runner
steps:
# Setups singularity to the job to make it accessible to other steps
# Cleanup steps to free up disk space
- name: Cleanup disk space for large docker images
run: |
sudo rm -rf /usr/share/dotnet
sudo rm -rf /opt/ghc
sudo rm -rf "/usr/local/share/boost"
sudo rm -rf /opt/hostedtoolcache
sudo rm -rf "$AGENT_TOOLSDIRECTORY"
- name: Free Disk Space (Ubuntu)
uses: jlumbroso/free-disk-space@main
with:
# this might remove tools that are actually needed,
# if set to "true" but frees about 6 GB
tool-cache: false
# all of these default to true, but feel free to set to
# "false" if necessary for your workflow
android: true
dotnet: true
haskell: true
large-packages: true
docker-images: false
- uses: actions/checkout@v4
if: needs.create_new_branch.outputs.BRANCHNAME == ''
with:
ref: issue-${{ github.event.issue.number }}
fetch-depth: 0
# If branchName is empty use issue number, else, use the branchName 🟢
- uses: actions/checkout@v4
if: needs.create_new_branch.outputs.BRANCHNAME != ''
with:
ref: ${{ needs.create_new_branch.outputs.BRANCHNAME }}
fetch-depth: 0
- name: Install singularity
run: |
wget https://github.com/sylabs/singularity/releases/download/v4.0.0/singularity-ce_4.0.0-focal_amd64.deb
sudo apt install ./singularity-ce_4.0.0-focal_amd64.deb -y
singularity version
# Build the Docker image 🟢
- name: Build Docker image
run: |
docker build -t ${{ needs.build-docker.outputs.MODELNAME }} ./${{ needs.create_new_branch.outputs.MODELPATH }}/docker
docker save --output ./${{ needs.create_new_branch.outputs.MODELPATH }}/docker/${{ needs.build-docker.outputs.MODELNAME }}.tar ${{ needs.build-docker.outputs.MODELNAME }}
# Convert the docker image to a singularity image
- name: Convert docker image to singularity image
run: |
singularity build ./${{ needs.create_new_branch.outputs.MODELPATH }}/docker/${{ needs.build-docker.outputs.MODELNAME }}.sif docker-archive://./${{ needs.create_new_branch.outputs.MODELPATH }}/docker/${{ needs.build-docker.outputs.MODELNAME }}.tar
# Datalad get the sample dataset
- name: Get sample dataset
run: |
sudo apt install python3-pip -y
sudo apt-get install datalad -y
python3 -m pip install datalad-osf
sudo python3 -m pip install datalad-installer
datalad-installer --sudo ok git-annex -m datalad/git-annex:release
sudo git config --global filter.annex.process "git-annex filter-process"
git config --system user.name "trained_models"
git config --system user.email "trained_models"
datalad get ./${{ needs.create_new_branch.outputs.MODELPATH }}/weights/
datalad get ./${{ needs.create_new_branch.outputs.MODELPATH }}/example-data/
sudo apt install unzip
cd ./${{ needs.create_new_branch.outputs.MODELPATH }}/example-data/
find . -name "*.zip" -exec unzip {} \;
# Push the dataset to the repo
datalad save . -m "Added sample data"
datalad push --to osf-storage
datalad push --to origin
- name: Get the weights
run: |
cd ./${{ needs.create_new_branch.outputs.MODELPATH }}/weights/
find . -name "*.zip" -exec unzip {} \;
# Push the weights to the repo
datalad save . -m "Added model weights"
datalad push --to osf-storage
datalad push --to origin
- name: Collect Workflow Telemetry
uses: runforesight/workflow-telemetry-action@v1
with:
job_summary: true
proc_trace_sys_enable: true
proc_trace_table_show: true
# Run in singularity the test command saved under needs.create_new_branch.outputs.DEEPCSR
- name: Run test command in Singularity
run: |
# Parent directory as a bind path in a env variable
singularity exec --nv ./${{ needs.create_new_branch.outputs.MODELPATH }}/docker/${{ needs.build-docker.outputs.MODELNAME }}.sif ${{ needs.create_new_branch.outputs.TESTCOMMAND }}
# Load the Docker image 🟢
- name: Load Docker image
run: |
docker load --input "./${{ needs.create_new_branch.outputs.MODELPATH }}/docker/${{ needs.build-docker.outputs.MODELNAME }}.tar"
# Test model with docker image as well
- name: Run test command in Docker
run: |
docker run --gpus all -v /actions-runner/_work/${{ github.event.repository.name }}/${{ github.event.repository.name }}:/output ${{ needs.build-docker.outputs.MODELNAME }} "cd /output; ${{ needs.create_new_branch.outputs.TESTCOMMAND }}"
# Get model's version from model's path
- name: Get model's version
id: modelVersion
run: |
# Get the version
model_version=$(echo "${{ needs.create_new_branch.outputs.MODELPATH }}" | awk -F '/' '{print $(NF)}')
# Set the version as an output
echo "model_version=$model_version" >> $GITHUB_OUTPUT
# Push the Docker image to Docker Hub (only if the PR is merged) 🟢
- name: Push Docker image
run: |
docker login -u ${{ secrets.DOCKER_USERNAME }} -p ${{ secrets.DOCKER_PASSWORD }}
docker tag ${{ needs.build-docker.outputs.MODELNAME }} ${{ needs.build-docker.outputs.IMAGENAME }}:${{ steps.modelVersion.outputs.model_version }}
docker push ${{ needs.build-docker.outputs.IMAGENAME }}:${{ steps.modelVersion.outputs.model_version }}
stop-runner:
name: Stop self-hosted EC2 runner
needs: [create_new_branch, push-model, build, build-docker, push-weights, start-runner, test-model] # required to wait when the main job is done
runs-on: ubuntu-latest
if: ${{ always() }}
# required to stop the runner even if the error happened in the previous jobs
steps:
- name: Configure AWS credentials
uses: aws-actions/configure-aws-credentials@v1
with:
aws-access-key-id: ${{ secrets.AWS_KEY_ID }}
aws-secret-access-key: ${{ secrets.AWS_KEY_SECRET }}
aws-region: ${{ vars.AWS_REGION }}
- name: Stop EC2 runner
uses: machulav/ec2-github-runner@v2
with:
mode: stop
github-token: ${{ secrets.GH_TOKEN }}
label: ${{ needs.start-runner.outputs.label }}
ec2-instance-id: ${{ needs.start-runner.outputs.ec2-instance-id }}
failed:
runs-on: ubuntu-latest
needs: [create_new_branch, push-model, build, build-docker, push-weights, start-runner, test-model]
if: ${{ failure() }}
steps:
- name: Set labels
uses: actions-cool/issues-helper@v3
with:
actions: 'set-labels'
token: ${{ secrets.GITHUB_TOKEN }}
issue-number: ${{ github.event.issue.number }}
labels: 'failed'
- name: Create comment
uses: actions-cool/issues-helper@v3
with:
actions: 'create-comment'
token: ${{ secrets.GITHUB_TOKEN }}
issue-number: ${{ github.event.issue.number }}
body: |
[This is an automated message] Hello @${{ github.event.issue.user.login }},
🔴 The folders/scripts you provided did not pass our tests. Please review the Action, and modify this issue's urls accordingly.
When ready, simply append "Ready XX" in the issue title (where XX is a number incrementing with 01 each time a fix has been applied).
emoji: '+1,eyes'
success:
needs: [create_new_branch, push-model, build, build-docker, push-weights, start-runner, test-model, stop-runner]
runs-on: ubuntu-latest
steps:
- name: Set labels
uses: actions-cool/issues-helper@v3
with:
actions: 'set-labels'
token: ${{ secrets.GITHUB_TOKEN }}
issue-number: ${{ github.event.issue.number }}
labels: 'success'
- name: Create comment
uses: actions-cool/issues-helper@v3
with:
actions: 'create-comment'
token: ${{ secrets.GITHUB_TOKEN }}
issue-number: ${{ github.event.issue.number }}
body: |
[This is an automated message] Hello @${{ github.event.issue.user.login }}. The workflow linked to adding your model finished successfully! Please double check that this issue's tag is "success."
🟢 A Draft PR should be linked to this issue. Now that your model passed the checks, feel free to change the status of the PR to "ready for review."
⭐ Thank you for adding a model to Nobrainer-Zoo!
emoji: '+1,hooray,rocket'
##########################################
# This workflow will add a new model to the zoo
# The flow of the workflow is as follows:
# 1. Create a new branch from the issue
# 2. Scrape the info from the issue
# 3. Scrape info from the form
# 4. Push the new created files and folders to the branch
# 5. Build the model card and spec.yaml file
# 6. Update the model's spec.yaml file
# 7. Build the docker image
# 8. Push the weights and sample data to the datalad repo
# 9. Start the self-hosted runner
# 10. Test the model/Push docker image to dockerhub
# 11. Stop the self-hosted runner
# 12. If the model fails, add the failed label and comment
# 13. If the model succeeds, add the success label and comment