name: Query Agent (SQL Regression)

# On-demand SQL regression suite for the OpenObserve query engine.
# Runs 672+ sqllogictest-style queries against a live OO instance and reports
# pass/fail on the PR. Triggered by PR comment or manual dispatch.
#
# Two entry points, both PR-bound:
#   - a PR comment, EXACTLY:
#       /query-agent  → FULL run: triage → build → test → report
#   - Manual dispatch with pr_number (required; bare number, "#123", or full URL)
#
# ACCESS CONTROL:
#   - TRIGGER: only the engineering-team ALLOWLIST can start a run.
#   - TEST EXECUTION: only runs when the PR's AUTHOR is also on the allowlist.
#
# IMPORTANT: merged to main before triggers take effect.

on:
  workflow_dispatch:
    inputs:
      pr_number:
        description: "PR number or URL (required)"
        required: true
      dry_run:
        description: "Dry-run: triage only, don't run tests"
        type: boolean
        required: true
        default: true
  issue_comment:
    types: [created]

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

concurrency:
  group: query-agent-${{ github.event.issue.number || github.event.inputs.pr_number || github.ref }}
  cancel-in-progress: true

env:
  # Server config — matches api-testing.yml
  ZO_ROOT_USER_EMAIL: root@example.com
  ZO_ROOT_USER_PASSWORD: Complexpass#123
  ZO_BASE_URL: http://localhost:5080/
  ZO_BASE_URL_SC: http://localhost:5080/
  ZO_USAGE_REPORTING_ENABLED: true
  ZO_USAGE_PUBLISH_INTERVAL: 5
  ZO_ALERT_SCHEDULE_INTERVAL: 3
  ZO_INGEST_ALLOWED_UPTO: 48
  ZO_CREATE_ORG_THROUGH_INGESTION: true

jobs:
  # ---------------------------------------------------------------------------
  # JOB 1 — Triage. Resolve PR, gate allowlist, check if diff touches
  # query-engine paths, post decision comment.
  # ---------------------------------------------------------------------------
  triage:
    if: >
      contains(fromJSON('["bjp232004","hengfeiyang","prabhatsharma","uddhavdave","oasisk","haohuaijin","Subhra264","Loaki07","007harshmahajan","chaitanya-sistla","ktx-kirtan","omkarK06","ktx-vaidehi","ktx-abhay","neha00290","ktx-akshay","ktx-sadhna","nikhilsaikethe","YashodhanJoshi1","ktx-riya","mmosarafO2","Shrinath-O2","ktx-mihir"]'), github.actor) &&
      (github.event_name == 'workflow_dispatch' ||
       (github.event_name == 'issue_comment' &&
        github.event.issue.pull_request &&
        github.event.comment.body == '/query-agent'))
    runs-on: ubuntu-latest
    timeout-minutes: 10
    outputs:
      skip: ${{ steps.ctx.outputs.skip }}
      pr_number: ${{ steps.resolve.outputs.pr_number }}
      author_allowed: ${{ steps.prinfo.outputs.author_allowed }}
      author: ${{ steps.prinfo.outputs.author }}
      dry_run: ${{ steps.mode.outputs.dry_run }}
    steps:
      - name: Resolve PR number
        id: resolve
        uses: actions/github-script@v9
        with:
          script: |
            const raw = context.eventName === 'issue_comment'
              ? String(context.payload.issue.number)
              : String(context.payload.inputs.pr_number || '');
            const urlMatch = raw.match(/\/pull\/(\d+)/);
            const numMatch = raw.match(/\d+/);
            const pr = urlMatch ? urlMatch[1] : (numMatch ? numMatch[0] : '');
            if (!pr) core.setFailed(`Could not parse a PR number from input: "${raw}"`);
            core.setOutput('pr_number', pr);

      - name: Inspect PR — author allowlist gate
        id: prinfo
        if: steps.resolve.outputs.pr_number != ''
        uses: actions/github-script@v9
        with:
          script: |
            const ALLOW = ["bjp232004","hengfeiyang","prabhatsharma","uddhavdave","oasisk","haohuaijin","Subhra264","Loaki07","007harshmahajan","chaitanya-sistla","ktx-kirtan","omkarK06","ktx-vaidehi","ktx-abhay","neha00290","ktx-akshay","ktx-sadhna","nikhilsaikethe","YashodhanJoshi1","ktx-riya","mmosarafO2","Shrinath-O2","ktx-mihir"].map(s => s.toLowerCase());
            const pr = Number('${{ steps.resolve.outputs.pr_number }}');
            const { data } = await github.rest.pulls.get({
              owner: context.repo.owner, repo: context.repo.repo, pull_number: pr,
            });
            const author = (data.user && data.user.login) ? data.user.login : '';
            const allowed = ALLOW.includes(author.toLowerCase());
            core.setOutput('author', author);
            core.setOutput('author_allowed', allowed ? 'true' : 'false');

      - name: Determine dry-run mode
        id: mode
        env:
          EVENT_NAME: ${{ github.event_name }}
          INPUT_DRY: ${{ github.event.inputs.dry_run }}
        run: |
          DRY=true
          if [ "$EVENT_NAME" = "workflow_dispatch" ] && [ "$INPUT_DRY" = "false" ]; then
            DRY=false
          elif [ "$EVENT_NAME" = "issue_comment" ]; then
            DRY=false  # /query-agent comment is always a full run
          fi
          echo "dry_run=$DRY" >> "$GITHUB_OUTPUT"

      - name: Acknowledge — react + post in-progress comment
        id: ack
        if: steps.resolve.outputs.pr_number != '' && steps.prinfo.outputs.author_allowed == 'true'
        uses: actions/github-script@v9
        with:
          script: |
            const pr = Number('${{ steps.resolve.outputs.pr_number }}');
            const runUrl = `${context.serverUrl}/${context.repo.owner}/${context.repo.repo}/actions/runs/${context.runId}`;
            if (context.eventName === 'issue_comment') {
              try {
                await github.rest.reactions.createForIssueComment({
                  owner: context.repo.owner, repo: context.repo.repo,
                  comment_id: context.payload.comment.id, content: 'eyes',
                });
              } catch (e) { core.info(`reaction failed: ${e.message}`); }
            }
            const { data } = await github.rest.issues.createComment({
              owner: context.repo.owner, repo: context.repo.repo, issue_number: pr,
              body: `### 🤖 Query Agent — triage in progress…\n\nAnalyzing the change; I'll update this comment with the decision shortly.\n\n[View run](${runUrl})`,
            });
            core.setOutput('comment_id', data.id);

      - uses: actions/checkout@v4
        if: steps.prinfo.outputs.author_allowed == 'true'
        with:
          fetch-depth: 1
          persist-credentials: false

      - name: Fetch PR diff
        if: steps.prinfo.outputs.author_allowed == 'true'
        env:
          GH_TOKEN: ${{ secrets.GITHUB_TOKEN }}
          PR_NUMBER: ${{ steps.resolve.outputs.pr_number }}
        run: |
          mkdir -p /tmp/query-agent-ci
          gh pr diff "$PR_NUMBER" --repo "$GITHUB_REPOSITORY" > /tmp/query-agent-ci/diff.patch || true
          if [ ! -s /tmp/query-agent-ci/diff.patch ]; then
            echo "::warning::Empty diff — nothing to triage"
          fi
          echo "diff bytes: $(wc -c < /tmp/query-agent-ci/diff.patch)"

      - name: Triage — check for query-engine changes
        id: ctx
        if: steps.prinfo.outputs.author_allowed == 'true'
        run: |
          # Paths that warrant running the SQL regression suite
          QUERY_PATHS='src/query/|src/sql/|src/service/search/|src/infra/table/|src/job/files/|tests/test-data/query-agent/|tests/api-testing/tests/query_agent/'
          if grep -qE "$QUERY_PATHS" /tmp/query-agent-ci/diff.patch 2>/dev/null; then
            echo "skip=false" >> "$GITHUB_OUTPUT"
            echo "skip_reason=Query-engine paths detected in diff" >> "$GITHUB_OUTPUT"
          else
            echo "skip=true" >> "$GITHUB_OUTPUT"
            echo "skip_reason=No query-engine paths in diff" >> "$GITHUB_OUTPUT"
          fi

      - name: Post or update triage comment
        if: always() && steps.resolve.outputs.pr_number != ''
        uses: actions/github-script@v9
        with:
          script: |
            const runUrl = `${context.serverUrl}/${context.repo.owner}/${context.repo.repo}/actions/runs/${context.runId}`;
            const authorAllowed = '${{ steps.prinfo.outputs.author_allowed }}' === 'true';
            const skip = '${{ steps.ctx.outputs.skip }}' || 'true';
            const skipReason = '${{ steps.ctx.outputs.skip_reason }}' || '';
            const dryRun = '${{ steps.mode.outputs.dry_run }}' === 'true';
            let body = '### 🤖 Query Agent\n\n';
            if (!authorAllowed) {
              const a = '${{ steps.prinfo.outputs.author }}';
              body += `Skipped — PR author${a ? ` @${a}` : ''} is not on the engineering-team allowlist. Query agent only runs for PRs authored by the team.`;
            } else if (skip === 'true') {
              body += `- **Skip:** yes — ${skipReason}\n`;
              body += `- No SQL regression tests needed for this change.\n`;
            } else {
              body += `- **Skip:** no — ${skipReason}\n`;
              if (dryRun) {
                body += `- ⚠️ **Dry-run:** triage only, no tests executed.\n`;
              } else {
                body += `- ▶️ Running the query agent regression suite…\n`;
              }
            }
            body += `\n[View run](${runUrl})`;
            const commentId = '${{ steps.ack.outputs.comment_id }}';
            if (commentId) {
              await github.rest.issues.updateComment({
                owner: context.repo.owner, repo: context.repo.repo,
                comment_id: Number(commentId), body,
              });
            } else {
              await github.rest.issues.createComment({
                owner: context.repo.owner, repo: context.repo.repo,
                issue_number: Number('${{ steps.resolve.outputs.pr_number }}'), body,
              });
            }

  # ---------------------------------------------------------------------------
  # JOB 2 — Run the query agent regression suite against a live OO instance.
  # Builds OO from source, runs all 672+ queries, posts pass/fail on the PR.
  # ---------------------------------------------------------------------------
  run_tests:
    needs: triage
    if: >
      needs.triage.outputs.dry_run == 'false' &&
      needs.triage.outputs.skip == 'false' &&
      needs.triage.outputs.author_allowed == 'true'
    runs-on: ubicloud-standard-4
    timeout-minutes: 60
    # Run the full SQL regression suite with the single-node physical
    # optimizer both enabled (default) and disabled. The non-optimized
    # RemoteScanRewriter path triggers known upstream DataFusion bugs on
    # a handful of queries — those are tagged skip_without_single_node_opt
    # and filtered by load_all_queries() when SINGLE_NODE_OPT_MODE=false.
    strategy:
      fail-fast: false
      matrix:
        config:
          - label: single_node_opt_false
            env: "ZO_FEATURE_SINGLE_NODE_OPTIMIZE_ENABLED=false"
            mode: "false"
          - label: single_node_opt_true
            env: "ZO_FEATURE_SINGLE_NODE_OPTIMIZE_ENABLED=true"
            mode: "true"
    permissions:
      contents: read
      issues: write
      pull-requests: write
    steps:
      - name: Remove unused tools
        run: |
          sudo rm -rf /usr/share/dotnet
          sudo rm -rf /opt/ghc
          sudo rm -rf "/usr/local/share/boost"
          sudo rm -rf "$AGENT_TOOLSDIRECTORY"

      - name: Checkout git repo
        uses: actions/checkout@v5
        with:
          fetch-depth: 0

      - name: Setup Rust Toolchain
        uses: dtolnay/rust-toolchain@master
        with:
          toolchain: nightly-2026-05-20

      - name: Setup Rust Cache
        uses: Swatinem/rust-cache@v2
        with:
          cache-on-failure: true
          shared-key: backend-debug-mimalloc
          save-if: false

      - name: Install Protoc
        run: bash ./.github/protoc.sh

      - name: Build OpenObserve debug binary
        run: cargo build --features mimalloc

      - name: Setup Python
        uses: actions/setup-python@v5
        with:
          python-version: "3.11.6"

      - name: Setup rye
        uses: eifinger/setup-rye@v4
        with:
          enable-cache: true
          working-directory: "tests/api-testing/"
          version: "0.27.0"

      - name: Pin cpython
        run: rye pin cpython@3.11.6
        working-directory: tests/api-testing/

      - name: Rye sync
        run: rye sync
        working-directory: tests/api-testing/

      - name: Regenerate query-agent expected values
        run: |
          cd ${{ github.workspace }}/tests/test-data/query-agent
          ${{ github.workspace }}/tests/api-testing/.venv/bin/python compute_counts.py

      - name: Run query agent tests
        id: qa_tests
        run: |
          ZO_MAX_FILE_RETENTION_TIME=1 ${{ matrix.config.env }} ${{ github.workspace }}/target/debug/openobserve > o2_query_agent_${{ matrix.config.label }}.log 2>&1 &
          SERVER_PID=$!
          echo "server_pid=$SERVER_PID" >> "$GITHUB_OUTPUT"
          for i in $(seq 1 60); do
            if curl -sf http://localhost:5080/healthz > /dev/null 2>&1; then
              echo "Server ready after ${i}s"; break
            fi
            sleep 1
          done
          # SINGLE_NODE_OPT_MODE controls which queries load_all_queries() includes
          SINGLE_NODE_OPT_MODE=${{ matrix.config.mode }} rye run pytest tests/query_agent/ --ignore=tests/query_agent/test_3_vortex.py -v \
            --junitxml=junit-query-agent-${{ matrix.config.label }}.xml 2>&1 | tee pytest-query-agent-${{ matrix.config.label }}.log
          echo "exit_code=${PIPESTATUS[0]}" >> "$GITHUB_OUTPUT"
        working-directory: tests/api-testing/

      - name: Kill server
        if: always()
        run: kill ${{ steps.qa_tests.outputs.server_pid }} || true

      - name: Upload server log on failure
        if: failure()
        uses: actions/upload-artifact@v4
        with:
          name: openobserve-query-agent-${{ matrix.config.label }}-log
          path: tests/api-testing/o2_query_agent_${{ matrix.config.label }}.log
          if-no-files-found: warn
          retention-days: 14

      - name: Upload pytest log on failure
        if: failure()
        uses: actions/upload-artifact@v4
        with:
          name: pytest-query-agent-${{ matrix.config.label }}-log
          path: tests/api-testing/pytest-query-agent-${{ matrix.config.label }}.log
          if-no-files-found: warn
          retention-days: 14

      - name: Upload JUnit report
        if: always()
        uses: actions/upload-artifact@v4
        with:
          name: junit-query-agent-${{ matrix.config.label }}
          path: tests/api-testing/junit-query-agent-${{ matrix.config.label }}.xml
          if-no-files-found: warn
          retention-days: 7

      - name: Post results comment
        if: always() && needs.triage.outputs.pr_number != ''
        uses: actions/github-script@v9
        with:
          script: |
            const fs = require('fs');
            const pr = Number('${{ needs.triage.outputs.pr_number }}');
            const runUrl = `${context.serverUrl}/${context.repo.owner}/${context.repo.repo}/actions/runs/${context.runId}`;
            const mode = '${{ matrix.config.label }}';
            const exitCode = '${{ steps.qa_tests.outputs.exit_code }}';
            const passed = exitCode === '0';
            let body = `### 🤖 Query Agent — Results (${mode})\n\n`;
            body += passed ? '✅ **All query agent tests passed.**\n\n' : '❌ **Query agent tests failed.**\n\n';
            // Parse pytest summary from log if available
            try {
              const log = fs.readFileSync(`tests/api-testing/pytest-query-agent-${mode}.log`, 'utf8');
              const m = log.match(/(\d+) passed.*(\d+) failed.*(\d+) error/);
              if (m) {
                body += `\`\`\`\n${m[0]}\n\`\`\`\n`;
              }
            } catch (e) { /* no log to parse */ }
            body += `\n[View run](${runUrl})`;
            // Find the ack comment and update it, or create new
            await github.rest.issues.createComment({
              owner: context.repo.owner, repo: context.repo.repo, issue_number: pr, body,
            });