name: Generate code scanning query lists

# **What it does**: This workflow is currently run manually approximately every two weeks as part
#      of the release process for the CodeQL CLI. We hope to automate this in the future
#   When run, this workflow generates updated query lists with data from the codeql
#   repository, and creates a pull request if there are updates.
# **Why we have it**: So we can automate CodeQL query tables and show code scanning users the built in queries.
# **Who does it impact**: Anyone making CodeQL query suite changes in `github/codeql`, and wanting to get them published on the docs site.

on:
  workflow_dispatch:
    inputs:
      SOURCE_BRANCH:
        description: 'Branch to pull the source files from in the codeql repo (for example codeql-cli-2.x.x).'
        type: string
        required: true
        default: 'main'

  pull_request:
    paths:
      - .github/workflows/generate-code-scanning-query-lists.yml
      - src/codeql-queries/scripts/generate-code-scanning-query-list.ts
      - src/codeql-queries/scripts/generate-code-quality-query-list.ts
      - .github/actions/install-cocofix/action.yml

permissions:
  contents: write
  pull-requests: write

jobs:
  generate-security-query-lists:
    if: github.repository == 'github/docs-internal'
    runs-on: ubuntu-latest
    steps:
      - name: Checkout repository code
        uses: actions/checkout@8e8c483db84b4bee98b60c0593521ed34d9990e8 # v6.0.1

      - uses: ./.github/actions/node-npm-setup

      - name: Checkout codeql repo
        uses: actions/checkout@8e8c483db84b4bee98b60c0593521ed34d9990e8 # v6.0.1
        with:
          repository: github/codeql
          path: codeql
          ref: ${{ inputs.SOURCE_BRANCH || 'main' }}

      - name: Get the codeql SHA being synced
        id: codeql
        shell: bash
        run: |
          cd codeql
          OPENAPI_COMMIT_SHA=$(git rev-parse HEAD)
          echo "OPENAPI_COMMIT_SHA=$OPENAPI_COMMIT_SHA" >> $GITHUB_OUTPUT
          echo "Copied files from github/codeql repo. Commit SHA: $OPENAPI_COMMIT_SHA"

      - name: Download CodeQL CLI
        # Look under the `codeql` directory, as this is where we checked out the `github/codeql` repo
        uses: ./codeql/.github/actions/fetch-codeql

      - name: Test CodeQL CLI Download
        shell: bash
        run: codeql --version

      # "Server for running multiple commands while avoiding repeated JVM initialization."
      # Having started this should speed up the execution of the various
      # CLI calls of the executable.
      - name: Start CodeQL CLI server in the background
        shell: bash
        run: |
          codeql execute cli-server &
          sleep 3
          codeql --version

      - uses: ./.github/actions/install-cocofix
        with:
          token: ${{ secrets.DOCS_BOT_PAT_BASE }}

      - name: Build code scanning security query lists
        shell: bash
        run: |
          for lang in "actions" "cpp" "csharp" "go" "java" "javascript" "python" "ruby" "rust" "swift"; do
            echo "Generating code scanning query list for $lang"
            npm run generate-code-scanning-query-list -- \
              --verbose \
              --codeql-path codeql \
              --codeql-dir codeql \
              -o data/reusables/code-scanning/codeql-query-tables/$lang.md \
              $lang
          done

      - name: Upload security query lists
        uses: actions/upload-artifact@b7c566a772e6b6bfb58ed0dc250532a479d7789f # v6.0.0
        with:
          name: security-query-lists
          path: data/reusables/code-scanning/codeql-query-tables/

  generate-quality-query-lists:
    if: github.repository == 'github/docs-internal'
    runs-on: ubuntu-latest
    steps:
      - name: Checkout repository code
        uses: actions/checkout@8e8c483db84b4bee98b60c0593521ed34d9990e8 # v6.0.1

      - uses: ./.github/actions/node-npm-setup

      - name: Checkout codeql repo
        uses: actions/checkout@8e8c483db84b4bee98b60c0593521ed34d9990e8 # v6.0.1
        with:
          repository: github/codeql
          path: codeql
          ref: ${{ inputs.SOURCE_BRANCH || 'main' }}

      - name: Get the codeql SHA being synced
        id: codeql
        shell: bash
        run: |
          cd codeql
          OPENAPI_COMMIT_SHA=$(git rev-parse HEAD)
          echo "OPENAPI_COMMIT_SHA=$OPENAPI_COMMIT_SHA" >> $GITHUB_OUTPUT
          echo "Copied files from github/codeql repo. Commit SHA: $OPENAPI_COMMIT_SHA"

      - name: Download CodeQL CLI
        # Look under the `codeql` directory, as this is where we checked out the `github/codeql` repo
        uses: ./codeql/.github/actions/fetch-codeql

      - name: Test CodeQL CLI Download
        shell: bash
        run: codeql --version

      # "Server for running multiple commands while avoiding repeated JVM initialization."
      # Having started this should speed up the execution of the various
      # CLI calls of the executable.
      - name: Start CodeQL CLI server in the background
        shell: bash
        run: |
          codeql execute cli-server &
          sleep 3
          codeql --version

      - name: Build code quality query lists
        shell: bash
        run: |
          for lang in "csharp" "go" "java" "javascript" "python" "ruby"; do
            echo "Generating code quality query list for $lang"
            npm run generate-code-quality-query-list -- \
              --verbose \
              --codeql-path codeql \
              --codeql-dir codeql \
              -o data/reusables/code-quality/codeql-query-tables/$lang.md \
              $lang
          done

      - name: Upload quality query lists
        uses: actions/upload-artifact@b7c566a772e6b6bfb58ed0dc250532a479d7789f # v6.0.0
        with:
          name: quality-query-lists
          path: data/reusables/code-quality/codeql-query-tables/

  create-pull-request:
    if: github.repository == 'github/docs-internal'
    runs-on: ubuntu-latest
    needs: [generate-security-query-lists, generate-quality-query-lists]
    steps:
      - name: Checkout repository code
        uses: actions/checkout@8e8c483db84b4bee98b60c0593521ed34d9990e8 # v6.0.1

      - name: Checkout codeql repo
        uses: actions/checkout@8e8c483db84b4bee98b60c0593521ed34d9990e8 # v6.0.1
        with:
          repository: github/codeql
          path: codeql
          ref: ${{ inputs.SOURCE_BRANCH || 'main' }}

      - name: Get the codeql SHA being synced
        id: codeql
        shell: bash
        run: |
          cd codeql
          OPENAPI_COMMIT_SHA=$(git rev-parse HEAD)
          echo "OPENAPI_COMMIT_SHA=$OPENAPI_COMMIT_SHA" >> $GITHUB_OUTPUT
          echo "Copied files from github/codeql repo. Commit SHA: $OPENAPI_COMMIT_SHA"

      - name: Download security query lists
        uses: actions/download-artifact@37930b1c2abaa49bbe596cd826c3c89aef350131 # v7.0.0
        with:
          name: security-query-lists
          path: data/reusables/code-scanning/codeql-query-tables/

      - name: Download quality query lists
        uses: actions/download-artifact@37930b1c2abaa49bbe596cd826c3c89aef350131 # v7.0.0
        with:
          name: quality-query-lists
          path: data/reusables/code-quality/codeql-query-tables/

      - name: Insight into diff
        shell: bash
        run: |
          git diff

      - name: Create pull request
        env:
          GITHUB_TOKEN: ${{ secrets.DOCS_BOT_PAT_BASE }}
          DRY_RUN: ${{ github.event_name == 'pull_request'}}
        shell: bash
        run: |

          # When we started, we downloaded the CodeQL CLI here in this workflow.
          # We have no intention of checking that in but we also don't want
          # `git status ...` to show it as an untracked file.
          rm -fr ./codeql

          # If nothing to commit, exit now. It's fine. No orphans.
          changes=$(git diff --name-only | wc -l)
          untracked=$(git status --untracked-files --short | wc -l)
          if [[ $changes -eq 0 ]] && [[ $untracked -eq 0 ]]; then
            echo "There are no changes to commit after running the generation and conversion scripts. Exiting..."
            exit 0
          fi

          git config --global user.name "docs-bot"
          git config --global user.email "77750099+docs-bot@users.noreply.github.com"

          branchname=codeql-query-tables-${{ steps.codeql.outputs.OPENAPI_COMMIT_SHA }}

          # Exit if the branch already exists. Since the actions/checkout fetch-depth is 1,
          # it doesn't "know" about branches locally, so we need to manually list them.
          branchExists=$(git ls-remote --heads origin refs/heads/$branchname | wc -l)

          # When run on a pull_request, we're just testing the tooling.
          # Exit before it actually pushes the possible changes.
          if [ "$DRY_RUN" = "true" ]; then
            echo "Dry-run mode when run in a pull request"
            echo "See the 'Insight into diff' step for the changes it would create PR about."
            exit 0
          fi

          if [ $branchExists -ne 0 ]; then
            echo "Branch $branchname already exists in the remote repository."
            exit 0
          else
            git checkout -b $branchname
          fi

          git add data/reusables/code-scanning/codeql-query-tables
          git add data/reusables/code-quality/codeql-query-tables
          git commit -m "Update CodeQL query tables"
          git push -u origin $branchname

          echo "Creating pull request..."
          gh pr create \
            --title "Update CodeQL query tables" \
            --repo github/docs-internal \
           --label "codeql-query-tables,skip FR board,ready-for-doc-review,workflow-generated" \
            --body '👋 humans. This PR updates the **CodeQL query table reusables** with the latest changes in preparation for the next **CodeQL CLI** release. (Synced from codeql@${{ steps.codeql.outputs.OPENAPI_COMMIT_SHA }})


            No action is required from the first responder for the Docs content team. This PR is automatically added to the Docs content review board. Any writer can review this by checking that the PR looks sensible. If CI does not pass or other problems arise, contact #docs-engineering on slack.


            When the DRI for the CodeQL CLI release is ready to publish, they will ask us to merge this PR in #docs-content.'