name: DeepWiki Documentation Crawler on: release: types: [published] workflow_dispatch: concurrency: group: deepwiki-${{ github.ref }} cancel-in-progress: true permissions: contents: read jobs: crawl: name: Crawl DeepWiki Documentation runs-on: ubuntu-latest timeout-minutes: 10 env: CRAWLER_DIR: ${{ github.workspace }}/.deepwiki-crawler REPOSITORY: ${{ github.repository }} # The pinned crawler schedules extra workers before updating its active # count. Use its sequential path to avoid an unbounded scheduling loop. MAX_CONCURRENCY: '1' REQUEST_TIMEOUT: '30000' RETRY_ATTEMPTS: '3' USER_AGENT: DeepWiki-Crawler/1.0 steps: - name: Check out workflow helpers uses: actions/checkout@v4 with: persist-credentials: false - name: Check out pinned crawler source uses: actions/checkout@v4 with: repository: erayack/deepwiki-action ref: b9fc8378235f298d736d12c5f4651d6ea0820c61 path: .deepwiki-crawler persist-credentials: false - uses: actions/setup-node@v4 with: node-version: '24' - name: Install crawler runtime dependencies # Install beside the scripts: Node resolves require() from their directory, # not the shell's cwd. Pin CommonJS-compatible versions; skip lifecycle hooks. shell: bash run: | set -euo pipefail npm install --prefix "$CRAWLER_DIR" --ignore-scripts --no-audit --no-fund --package-lock=false --save-exact node-fetch@2.7.0 cheerio@1.0.0 - name: Check repository and crawl DeepWiki shell: bash run: | set -euo pipefail [[ "$REPOSITORY" =~ ^[a-zA-Z0-9._-]+/[a-zA-Z0-9._-]+$ ]] node "$GITHUB_WORKSPACE/scripts/run-deepwiki.cjs" "$CRAWLER_DIR" "$REPOSITORY"