auto-fix-issues.yml

name: Auto Fix Issues
on:
  schedule:
    - cron: '*/10 * * * *'
  workflow_dispatch:
    inputs:
      issue_number:
        description: 'Issue number (empty = pick oldest)'
        required: false

permissions:
  contents: write
  pull-requests: write
  issues: write

jobs:
  check:
    runs-on: ubuntu-latest
    outputs:
      matrix: ${{ steps.pick.outputs.matrix }}
    steps:
      - id: pick
        env:
          GH_TOKEN: ${{ secrets.GH_PAT }}
          ISSUE_NUMBER: ${{ inputs.issue_number || '' }}
        run: |
          REPO="${{ github.repository }}"

          # Collect issues
          if [ -n "$ISSUE_NUMBER" ]; then
            ISSUES=$(gh issue view "$ISSUE_NUMBER" --repo "$REPO" --json number,title,body -q '[.]')
          else
            ISSUES=$(gh issue list --repo "$REPO" --label auto-fix --state open --limit 10 --json number,title,body)
          fi

          COUNT=$(echo "$ISSUES" | jq 'length')
          if [ "$COUNT" -eq 0 ]; then
            echo "No auto-fix issues"
            echo "matrix=" >> "$GITHUB_OUTPUT"
            exit 0
          fi

          # Collect open auto-fix PRs and their changed file paths
          OPEN_PRS=$(gh pr list --repo "$REPO" --label auto-fix --state open \
            --json number,headRefName,files --jq '[.[] | {number, branch: .headRefName, paths: [.files[].path]}]' 2>/dev/null || echo '[]')

          # For each issue, check if an existing PR touches the same doc category
          # Category = first two path segments under docs/en/ (e.g. docs/en/sql-reference, docs/en/guides)
          MATRIX="[]"
          for row in $(echo "$ISSUES" | jq -c '.[]'); do
            NUM=$(echo "$row" | jq -r '.number')
            TITLE=$(echo "$row" | jq -r '.title')
            BODY=$(echo "$row" | jq -r '.body')

            # Guess the doc category from issue title/body keywords
            CATEGORY=""
            TEXT=$(echo "${TITLE} ${BODY}" | tr '[:upper:]' '[:lower:]')
            if echo "$TEXT" | grep -qE 'function|func[ _-]'; then
              CATEGORY="docs/en/sql-reference/20-sql-functions"
            elif echo "$TEXT" | grep -qE 'command|statement|ddl|dml|dql'; then
              CATEGORY="docs/en/sql-reference/10-sql-commands"
            elif echo "$TEXT" | grep -qE 'sql.reference|sql-reference'; then
              CATEGORY="docs/en/sql-reference"
            elif echo "$TEXT" | grep -qE 'guide|load|unload|stage|warehouse|security'; then
              CATEGORY="docs/en/guides"
            fi

            # Find existing PR with overlapping category
            EXISTING_BRANCH=""
            EXISTING_PR=""
            if [ -n "$CATEGORY" ]; then
              MATCH=$(echo "$OPEN_PRS" | jq -c --arg cat "$CATEGORY" \
                '[.[] | select(.paths[] | startswith($cat))] | first // empty')
              if [ -n "$MATCH" ] && [ "$MATCH" != "null" ]; then
                EXISTING_BRANCH=$(echo "$MATCH" | jq -r '.branch')
                EXISTING_PR=$(echo "$MATCH" | jq -r '.number')
                echo "Issue #${NUM} matches existing PR #${EXISTING_PR} (branch: ${EXISTING_BRANCH})"
              fi
            fi

            # Mark in-progress
            gh issue edit "$NUM" --repo "$REPO" --add-label auto-fix-in-progress --remove-label auto-fix

            ENTRY=$(echo "$row" | jq -c --arg eb "$EXISTING_BRANCH" --arg ep "$EXISTING_PR" \
              '. + {existing_branch: $eb, existing_pr: $ep}')
            MATRIX=$(echo "$MATRIX" | jq -c --argjson e "$ENTRY" '. + [$e]')
          done

          if [ "$(echo "$MATRIX" | jq 'length')" -eq 0 ]; then
            echo "matrix=" >> "$GITHUB_OUTPUT"
          else
            echo "matrix=$(echo "$MATRIX" | jq -c '{include: .}')" >> "$GITHUB_OUTPUT"
          fi

  fix:
    needs: check
    if: needs.check.outputs.matrix != ''
    runs-on: ubuntu-latest
    strategy:
      matrix: ${{ fromJson(needs.check.outputs.matrix) }}
      max-parallel: 1
    steps:
      - uses: actions/checkout@v4

      - name: Checkout databend source
        uses: actions/checkout@v4
        with:
          repository: databendlabs/databend
          path: _databend
          sparse-checkout: src
          fetch-depth: 1

      - name: Fetch databend release tags
        env:
          GH_TOKEN: ${{ secrets.GH_PAT }}
        run: |
          gh release list --repo databendlabs/databend --limit 50 \
            --json tagName,publishedAt --jq '.[] | "\(.tagName) \(.publishedAt)"' \
            > _databend/RELEASES.txt

      - name: Install evot
        env:
          GH_TOKEN: ${{ secrets.GH_PAT }}
        run: |
          for TAG in $(gh release list --repo evotai/evot --limit 10 --json tagName --jq '.[].tagName'); do
            if gh release download "$TAG" --repo evotai/evot --pattern "evot-*-x86_64-unknown-linux-gnu.tar.gz" 2>/dev/null; then
              echo "Downloaded evot from $TAG"
              break
            fi
          done
          tar -xzf evot-*.tar.gz && chmod +x bin/evot
          mkdir -p ~/.evotai/lib
          cp lib/*.node ~/.evotai/lib/
          echo "$PWD/bin" >> "$GITHUB_PATH"

      - name: Write evot.env
        run: |
          mkdir -p ~/.evotai
          cat > ~/.evotai/evot.env << ENVEOF
          EVOT_LLM_PROVIDER=anthropic
          EVOT_LLM_ANTHROPIC_BASE_URL=${{ secrets.EVOT_LLM_ANTHROPIC_BASE_URL }}
          EVOT_LLM_ANTHROPIC_API_KEY=${{ secrets.EVOT_LLM_ANTHROPIC_API_KEY }}
          EVOT_LLM_ANTHROPIC_MODEL=${{ secrets.EVOT_LLM_ANTHROPIC_MODEL }}
          EVOT_LLM_OPENAI_BASE_URL=${{ secrets.EVOT_LLM_OPENAI_BASE_URL }}
          EVOT_LLM_OPENAI_API_KEY=${{ secrets.EVOT_LLM_OPENAI_API_KEY }}
          EVOT_LLM_OPENAI_MODEL=${{ secrets.EVOT_LLM_OPENAI_MODEL }}
          ENVEOF
          sed -i 's/^[[:space:]]*//' ~/.evotai/evot.env

      - name: Run
        run: |
          git config user.name "github-actions[bot]"
          git config user.email "41898282+github-actions[bot]@users.noreply.github.com"
          bash .github/scripts/auto-fix.sh
        env:
          REPO: ${{ github.repository }}
          ISSUE_JSON: ${{ toJson(matrix) }}
          EXISTING_BRANCH: ${{ matrix.existing_branch }}
          EXISTING_PR: ${{ matrix.existing_pr }}
          GENERATOR_MODEL: ''
          REVIEWER_MODEL: gpt-5.4
          EVOT_ID: auto-fix-ci
          GH_TOKEN: ${{ secrets.GH_PAT }}