pr-issue-validator.yaml

name: Validate Pull Request

on:
  pull_request:
    types:
      - opened
      - synchronize
      - edited
      - reopened
    branches:    
      - 'main'
      - 'release-**'
    # paths-ignore:
    #   - 'docs/**'
    #   - '.github/'
    #   - 'CHANGELOG/'
    #   - 'charts/'
    #   - 'manifests/'
    #   - 'sample-docker-templates/'
  
jobs:
  validate-PR-issue:
    runs-on: ubuntu-latest
    permissions:
      issues: write
      contents: read
      pull-requests: write

    steps:
    - name: Checkout repository
      uses: actions/checkout@v2
      with:
        ref: ${{ github.event.pull_request.head.sha }}
        fetch-depth: 0


    - name: Validate Issue Reference
      env:
        GITHUB_TOKEN: ${{ github.token }}
        PR_BODY: ${{ github.event.pull_request.body }}
        url: ${{ github.event.pull_request.url }}
        PRNUM: ${{ github.event.pull_request.number }}
        TITLE: ${{ github.event.pull_request.title }}
      run: |
        set -x
        echo "base or target repo : ${{ github.event.pull_request.base.repo.full_name }}"
        echo "head or source repo : ${{ github.event.pull_request.head.repo.full_name }}"
        if [[ ${{ github.event.pull_request.head.repo.full_name }} == ${{ github.event.pull_request.base.repo.full_name }} ]]; then
          export forked=false
        else
          export forked=true
        fi
        echo "forked: $forked"
        if [[ "$TITLE" == *"doc:"* || "$TITLE" == *"docs:"* || "$TITLE" == *"misc:"* || "$TITLE" == *"release:"* || "$TITLE" == *"Release:"*  ]]; then
          echo "Skipping validation as this is a PR for documentation or misc."
          if [[ $forked == true ]]; then
            echo "PR:Ready-to-Review, exiting gracefully"
            exit 0
          fi
          gh pr edit $PRNUM --remove-label "PR:Issue-verification-failed"
          gh pr edit $PRNUM --add-label "PR:Ready-to-Review"
          exit 0
        fi
        
        ### For ex: Fixes #2123
        ### For ex: Fixes: #2123
        pattern1="((Fixes|Resolves):? #[0-9]+)"
        
        ### For ex: Resolves https://github.com/devtron-labs/devtron/issues/2123
        pattern2="((Fixes|Resolves):? https://github.com/devtron-labs/devtron/issues/[0-9]+)"

        
        ### For ex: Fixes devtron-labs/devtron#2123
        pattern3="((Fixes|Resolves):? devtron-labs/devtron#[0-9]+)"

        ### For ex: Fixes [#4839](https://github.com/devtron-labs/devtron/issues/4839)
        pattern4="(Fixes|Resolves):?\s+\[#([0-9]+)\]"
        
        # Get the pull request body
        PR_BODY=$(jq -r '.pull_request.body' $GITHUB_EVENT_PATH)
        echo "PR_BODY = $PR_BODY"

        ### Checks if PR_BODY matches pattern1 or pattern2 or pattern3 or none
        ### grep -i (case insensitive) -E (enables extended regular expression in grep) -q (this option suppresses normal output)
        if echo "$PR_BODY" | grep -iEq "$pattern1"; then
            ### Here we are taking only the numerical value ie. issue number
            ### head -n1 only prints the 1st line.
            ### grep -o -E "[0-9]+ basically outputs only the number between [0-9]+
            echo "$PR_BODY" | grep -iE "$pattern1" | head -n1 | grep -o -E "[0-9]+" | tr -d '\r\n' > issue_num
            issue_num=$(cat issue_num)
            echo "issue_num is : $issue_num"
        elif echo "$PR_BODY" | grep -iEq "$pattern2"; then
            echo "$PR_BODY" | grep -iE "$pattern2" | head -n1 | awk -F '/' '{print $NF}' | tr -d '\r\n' > issue_num
            issue_num=$(cat issue_num)
            echo "issue_num is : $issue_num"
        elif echo "$PR_BODY" | grep -iEq "$pattern3"; then
            echo "$PR_BODY" | grep -iE "$pattern3" | head -n1 | awk -F '#' '{print $NF}' | tr -d '\r\n' > issue_num
            issue_num=$(cat issue_num)
            echo "issue_num is : $issue_num"
        elif echo "$PR_BODY" | grep -iEq "$pattern4"; then
            echo "$PR_BODY" | grep -oP "$pattern4" | head -n1 | grep -oP '#\K[0-9]+' | tr -d '\r\n' > issue_num
            issue_num=$(cat issue_num)
            echo "issue_num is : $issue_num"
        else
            echo "No Issue number detected hence failing the PR Validation check."
            if [[ $forked == true ]]; then
              echo "PR:Issue-verification-failed, exiting forcefully!"
              exit 1
            fi
            gh pr edit $PRNUM --add-label "PR:Issue-verification-failed"
            gh pr edit $PRNUM --remove-label "PR:Ready-to-Review"
            exit 1
        fi

        ### Here we are setting the Internal Field Separator to "/"
        ### read -r -> reads input from variable $url
        ### -a url_parts -> tells read command to store input into an array named url_parts[]
        IFS="/" read -r -a url_parts <<< "$url"
        
        # Remove the last two elements (repos and the issue number)
        unset url_parts[-1]
        unset url_parts[-1]
        # Reattach the URL pieces
        url=$(IFS=/; echo "${url_parts[*]}")

        # Add the issue number to the URL
        url="${url}/issues/${issue_num}"
        echo "$url"
        response_code=$(curl -s -o /dev/null -w "%{http_code}" "$url")
        if [[ "$response_code" -eq 200 ]]; then
          # Check if issue is open or closed
          text=$(curl -s "$url")
          echo "checking status of the issue"
          if [[ $(echo "$text" | jq -r '.state') == "open" ]]; then
            echo "Issue #$issue_num is open"
            echo "Issue reference found in the pull request body."
            if [[ $forked == true ]]; then
              echo "PR:Ready-to-Review, exiting gracefully"
              exit 0
            fi
            gh pr edit $PRNUM --remove-label "PR:Issue-verification-failed"
            gh pr edit $PRNUM --add-label "PR:Ready-to-Review"
            exit 0
          else
            echo "Issue #$issue_num is not open"
            if [[ $forked == true ]]; then
              echo "PR:Issue-verification-failed, exiting forcefully!"
              exit 1
            fi
            gh pr edit $PRNUM --add-label "PR:Issue-verification-failed"
            gh pr edit $PRNUM --remove-label "PR:Ready-to-Review"
            exit 1
          fi
        else
            echo "Invalid Response Code obtained - error code: $response_code"
            echo "No valid issue reference found in the pull request body."
            gh pr comment $PRNUM --body "PR is not linked to any issue, please make the corresponding changes in the body."
            if [[ $forked == true ]]; then
              echo "PR:Issue-verification-failed, exiting forcefully!"
              exit 1
            fi
            gh pr edit $PRNUM --add-label "PR:Issue-verification-failed"
            gh pr edit $PRNUM --remove-label "PR:Ready-to-Review"
            exit 1
        fi
    - name: Check SQL file format and duplicates
      shell: bash
      env: 
          pr_no: ${{ github.event.pull_request.number }}
          GH_TOKEN: ${{ github.token }}
      run: |
        
        # Fetch the latest changes from the main branch
        git fetch origin main
        
        # Get the list of changed files
        git diff origin/main...HEAD --name-only > diff
        
        # Specify the directory containing migration files
        MIGRATION_DIR="scripts/sql"
        
        # Print changed files
        echo "Changed files:"
        cat diff
        
        # Extract relevant .up.sql migration files from the changed files list
        changed_files=$(grep "^$MIGRATION_DIR/" diff | grep -E "\.up\.sql$")
        
        # Check if there are any .up.sql migration files in the changed files list
        if [ -z "$changed_files" ]; then
          echo "No .up.sql migration files found in the changes."
          exit 0
        fi
        
        
        # Extract unique migration numbers from the directory (considering only .up.sql files)
        existing_migrations=$(ls $MIGRATION_DIR | grep -E "\.up\.sql$" | grep -oE "[0-9]{3}[0-9]{3}[0-9]{2}" | sort | uniq)
        
        # Exclude migration numbers from changed files in existing_migrations
        while read -r file; do
          migration_number=$(basename "$file" | grep -oE "[0-9]{3}[0-9]{3}[0-9]{2}")
          existing_migrations=$(echo "$existing_migrations" | grep -v "$migration_number")
        done <<< "$changed_files"
        
        # Validate each changed .up.sql migration file
        is_valid=true
        processed_migrations=()
        while read -r file; do
          # Extract migration number from the filename
          migration_number=$(basename "$file" | grep -oE "[0-9]{3}[0-9]{3}[0-9]{2}")
        
          # Check if the filename has the full XXXPPPNN format
          if [[ ! $(basename "$file") =~ ^[0-9]{3}[0-9]{3}[0-9]{2}_ ]]; then
            echo "Error: Migration file $file does not have the complete XXXPPPNN format."
            is_valid=false
            continue
          fi
        
          if [ -z "$migration_number" ]; then
            echo "Warning: Could not extract migration number from $file."
            continue
          fi
        
          # Check if this migration number has already been processed
          if [[ " ${processed_migrations[@]} " =~ " $migration_number " ]]; then
            continue
          fi
          processed_migrations+=("$migration_number")
        
          # Check if the migration number is unique
          if echo "$existing_migrations" | grep -q "$migration_number"; then
            echo "Error: Migration number $migration_number already exists."
            is_valid=false
          fi
        
          # Check if the migration number is greater than previous ones
          last_migration=$(echo "$existing_migrations" | tail -n 1)
          if [ "$migration_number" -le "$last_migration" ]; then
            echo "Error: Migration number $migration_number is not greater than the latest ($last_migration)."
            is_valid=false
          fi
        
          # Check for sequential hotfix requirement (if NN > 01, check for NN-1)
          hotfix_number=$(echo "$migration_number" | grep -oE "[0-9]{2}$")
          if [ "$hotfix_number" -gt "01" ]; then
            previous_hotfix=$(printf "%02d" $((10#$hotfix_number - 1)))
            expected_previous_number="${migration_number:0:6}$previous_hotfix"
            if ! echo "$existing_migrations" | grep -q "$expected_previous_number"; then
              echo "Error: Previous hotfix migration $expected_previous_number not found for $migration_number."
              is_valid=false
            fi
          fi
        
        done <<< "$changed_files"
        
        if [ "$is_valid" = false ]; then
          echo "Validation failed. Please fix the errors before merging."
          gh pr comment $pr_no --body "The Migration files providede inside of the PR does not pass the criteria!!"
          exit 1
        fi
        
        echo "All .up.sql migration file validations passed."
        gh pr comment $pr_no --body "The migration files have successfully passed the criteria!!"