aslib_data_check.yml

# Author: Haniye Kashgarani
# Affiliation: University of Wyoming
# Date: 2023-08-17T15:39:37+00:00
# Github: haniyeka

name: aslib_data_check

on:
  push:
    branches:
      - main
  pull_request:
    branches:
      - main
  workflow_dispatch:

jobs:
  ASLIB_data_check_tool:
    runs-on: ubuntu-latest
    #container:
      # Docker image that includes python2.7, and libraries like liac-arff, pyyaml, and typing
      # This docker also includes files from https://github.com/coseal/aslib-spec/tree/master/data_check_tool_python in /usr/src/app/data_check_tool_python
      # image: haniyeka/aslib_data_check:latest
    steps:
    - name: Checkout code
      uses: actions/checkout@v3
      with:
        fetch-depth: 2
    - name: Run on push
      id: get-changes-push
      if: github.event_name == 'push'
      run: |
        echo "This is a push to the main branch."
        echo "Previous commit SHA:"
        echo ${{ github.event.before }}
        echo "Current commit SHA:"
        echo ${{ github.sha }}
        echo "Get changed directories:"
        dirs=$(git diff --name-only ${{ github.event.before }} ${{ github.sha }} | \
            grep -v '^\.github/' | \
            grep '/.*$' | \
            sed -E 's@(.*)/[^/]*$@\1@' | \
            sort -u | \
            tr '\n' ' ') 
        echo "Changed directories: $dirs"
        echo "DIRS=$dirs" >> $GITHUB_ENV

    - name: Run on pull request
      id: get-changes-pull-request
      if: github.event_name == 'pull_request'
      run: |
        echo "This is a pull request to the main branch."
        echo "Source branch:"
        echo ${{ github.event.pull_request.head.ref }}
        echo "Target branch:"
        echo ${{ github.event.pull_request.base.ref }}
        echo "Get changed directories:"
        git fetch origin ${{ github.event.pull_request.base.ref }} # Fetch the target branch
        dirs=$(git diff --name-only ${{ github.event.pull_request.base.sha }} ${{ github.event.pull_request.head.sha }} | \
             grep -v '^\.github/' | \
             grep '/.*$' | \
             sed -E 's@(.*)/[^/]*$@\1@' | \
             sort -u | \
             tr '\n' ' ')
        echo "Changed directories: $dirs"
        echo "DIRS=$dirs" >> $GITHUB_ENV

    - name: Run on workflow dispatch
      id: all-dirs-workflow-dispatch
      if: github.event_name == 'workflow_dispatch'
      run: |
        echo "This workflow was manually triggered."
        echo "Get all directories except .git and .github:"
        dirs=$(find . -maxdepth 1 -type d \( ! -name . \) \( ! -name .git \) \( ! -name .github \) | sed 's@^\./@@' | tr '\n' ' ')
        echo "Directories to check: $dirs"
        echo "DIRS=$dirs" >> $GITHUB_ENV

    - name: Run data check tool script on directories
      run: |
        echo "Current directory:"
        echo $(pwd)
        echo "Scenarios:"
        echo $DIRS
        if [ -z "$DIRS" ]; then
          echo "No scenario is changed or added."
          exit 0
        fi
        docker run -v $(pwd):/workspace -e DIRS="$DIRS" haniyeka/aslib_data_check:latest \
          bash -c "
            set -e
            ls /usr/src/app/data_check_tool_python/src/
            echo \$DIRS
            for dir in \$DIRS; do
              if [ -d \"/workspace/\$dir\" ]; then
                hidden_files=\$(find \"/workspace/\$dir/\" -name '.*' -type f)
                if [ ! -z \"\$hidden_files\" ]; then
                  echo 'Error: Hidden files found in' \$dir
                  echo 'Please remove them and try again!'
                  exit 1
                fi
                python /usr/src/app/data_check_tool_python/src/main.py --dir \"/workspace/\$dir\"
              fi
            done
          "