From d5c9bfa320a0ac5a0161ed76f7d26397d92b066b Mon Sep 17 00:00:00 2001 From: Your Name Date: Thu, 22 Jan 2026 23:51:18 +0800 Subject: [PATCH] feat(ci): add model/headless verification, use --no-ff merge strategy - Add verify-image job in release.yml with: - Headless environment verification (DISPLAY, Xvfb, LibreOffice, Inkscape, Calibre) - Model download verification (HuggingFace cache, ONNX models, offline mode) - Basic functionality verification (container startup, tool availability) - Add Test 3 and Test 4 in upstream-sync.yml for headless/model verification - Update auto-upstream-sync.yml to use git merge --no-ff for better history tracking --- .github/workflows/auto-upstream-sync.yml | 4 +- .github/workflows/release.yml | 242 ++++++++++++++++++++++- .github/workflows/upstream-sync.yml | 91 ++++++++- 3 files changed, 329 insertions(+), 8 deletions(-) diff --git a/.github/workflows/auto-upstream-sync.yml b/.github/workflows/auto-upstream-sync.yml index 1716491..a78aa1d 100644 --- a/.github/workflows/auto-upstream-sync.yml +++ b/.github/workflows/auto-upstream-sync.yml @@ -152,8 +152,8 @@ jobs: run: | echo "🔄 Merging upstream/${{ env.UPSTREAM_BRANCH }} into ${{ env.TARGET_BRANCH }}..." - # 嘗試自動 merge - if git merge upstream/${{ env.UPSTREAM_BRANCH }} --no-edit -m "🔄 Auto-merge upstream changes ($(date +'%Y-%m-%d'))"; then + # 嘗試自動 merge(使用 --no-ff 確保創建 merge commit,保留完整歷史) + if git merge upstream/${{ env.UPSTREAM_BRANCH }} --no-ff --no-edit -m "🔄 Merge upstream/${{ env.UPSTREAM_BRANCH }} ($(date +'%Y-%m-%d'))"; then echo "✅ Merge successful!" echo "merge_success=true" >> $GITHUB_OUTPUT else diff --git a/.github/workflows/release.yml b/.github/workflows/release.yml index 63e8660..a29ea29 100644 --- a/.github/workflows/release.yml +++ b/.github/workflows/release.yml @@ -185,11 +185,249 @@ jobs: echo "**Platforms:** linux/amd64, linux/arm64" >> $GITHUB_STEP_SUMMARY # ============================================================================== - # 📦 Create GitHub Release + # � Verify Docker Image(模型驗證 + Headless 驗證) + # ============================================================================== + verify-image: + name: Verify Docker Image + runs-on: ubuntu-latest + needs: build-and-push + timeout-minutes: 15 + + steps: + - name: Get tag name + id: tag + run: | + if [ "${{ github.event_name }}" == "workflow_dispatch" ]; then + echo "version=${{ github.event.inputs.tag }}" >> $GITHUB_OUTPUT + else + echo "version=${GITHUB_REF#refs/tags/}" >> $GITHUB_OUTPUT + fi + + - name: Pull Docker image + run: | + echo "📥 Pulling Docker image..." + docker pull ${{ env.DOCKER_IMAGE }}:${{ steps.tag.outputs.version }} + + # ======================================== + # 驗證 1: Headless 環境驗證 + # ======================================== + - name: "🖥️ Headless Environment Verification" + run: | + echo "========================================" + echo "🖥️ Headless 環境驗證" + echo "========================================" + + IMAGE="${{ env.DOCKER_IMAGE }}:${{ steps.tag.outputs.version }}" + FAILED=0 + + # 檢查 DISPLAY 環境變數(應為空或不存在) + echo "" + echo "📋 檢查 DISPLAY 環境變數..." + DISPLAY_VAR=$(docker run --rm "${IMAGE}" printenv DISPLAY 2>/dev/null || echo "") + if [ -z "$DISPLAY_VAR" ]; then + echo "✅ DISPLAY 未設定(headless 模式正確)" + else + echo "⚠️ DISPLAY 已設定: ${DISPLAY_VAR}(可能影響 headless 模式)" + fi + + # 檢查 Xvfb(應不存在或不需要) + echo "" + echo "📋 檢查 Xvfb..." + if docker run --rm "${IMAGE}" which Xvfb 2>/dev/null; then + echo "⚠️ Xvfb 存在(但應不需要)" + else + echo "✅ Xvfb 不存在(headless 模式正確)" + fi + + # 驗證 LibreOffice headless 模式 + echo "" + echo "📋 驗證 LibreOffice headless..." + if docker run --rm "${IMAGE}" libreoffice --headless --version 2>/dev/null | head -1; then + echo "✅ LibreOffice headless 模式可用" + else + echo "❌ LibreOffice headless 模式失敗" + FAILED=1 + fi + + # 驗證 Inkscape headless 模式 + echo "" + echo "📋 驗證 Inkscape headless..." + if docker run --rm "${IMAGE}" inkscape --version 2>/dev/null; then + echo "✅ Inkscape 可用" + else + echo "❌ Inkscape 不可用" + FAILED=1 + fi + + # 驗證 Calibre headless 模式(使用 ebook-convert) + echo "" + echo "📋 驗證 Calibre headless..." + if docker run --rm -e QTWEBENGINE_CHROMIUM_FLAGS="--no-sandbox" "${IMAGE}" ebook-convert --version 2>/dev/null | head -1; then + echo "✅ Calibre headless 模式可用" + else + echo "❌ Calibre headless 模式失敗" + FAILED=1 + fi + + echo "" + echo "========================================" + if [ $FAILED -eq 0 ]; then + echo "✅ Headless 環境驗證通過!" + else + echo "❌ Headless 環境驗證失敗!" + exit 1 + fi + echo "========================================" + + # ======================================== + # 驗證 2: 模型下載驗證 + # ======================================== + - name: "🤖 Model Download Verification" + run: | + echo "========================================" + echo "🤖 模型下載驗證" + echo "========================================" + + IMAGE="${{ env.DOCKER_IMAGE }}:${{ steps.tag.outputs.version }}" + FAILED=0 + + # 檢查 HuggingFace cache 目錄 + echo "" + echo "📋 檢查 HuggingFace cache..." + HF_CACHE=$(docker run --rm "${IMAGE}" ls -la /root/.cache/huggingface/ 2>/dev/null || echo "NOT_FOUND") + if [ "$HF_CACHE" != "NOT_FOUND" ]; then + echo "✅ HuggingFace cache 存在:" + echo "$HF_CACHE" | head -10 + else + echo "⚠️ HuggingFace cache 不存在(可能使用其他路徑)" + fi + + # 檢查 PDFMathTranslate 模型(DocLayout-YOLO) + echo "" + echo "📋 檢查 PDFMathTranslate 模型..." + ONNX_MODEL=$(docker run --rm "${IMAGE}" find /root -name "*.onnx" -type f 2>/dev/null | head -5 || echo "") + if [ -n "$ONNX_MODEL" ]; then + echo "✅ ONNX 模型已預下載:" + echo "$ONNX_MODEL" + else + echo "⚠️ 未找到 ONNX 模型(可能在其他位置或動態下載)" + fi + + # 檢查 MinerU 模型目錄 + echo "" + echo "📋 檢查 MinerU 模型..." + MINERU_MODELS=$(docker run --rm "${IMAGE}" ls -la /root/.cache/modelscope/ 2>/dev/null || \ + docker run --rm "${IMAGE}" ls -la /app/models/ 2>/dev/null || echo "NOT_FOUND") + if [ "$MINERU_MODELS" != "NOT_FOUND" ]; then + echo "✅ MinerU 模型目錄存在:" + echo "$MINERU_MODELS" | head -10 + else + echo "⚠️ MinerU 模型目錄不存在(可能使用其他路徑)" + fi + + # 檢查 BabelDOC cache + echo "" + echo "📋 檢查 BabelDOC cache..." + BABEL_CACHE=$(docker run --rm "${IMAGE}" ls -la /root/.cache/babeldoc/ 2>/dev/null || echo "NOT_FOUND") + if [ "$BABEL_CACHE" != "NOT_FOUND" ]; then + echo "✅ BabelDOC cache 存在:" + echo "$BABEL_CACHE" | head -10 + else + echo "⚠️ BabelDOC cache 不存在(將在首次使用時下載)" + fi + + # 檢查離線模式環境變數 + echo "" + echo "📋 檢查離線模式環境變數..." + HF_OFFLINE=$(docker run --rm "${IMAGE}" printenv HF_HUB_OFFLINE 2>/dev/null || echo "") + if [ "$HF_OFFLINE" = "1" ]; then + echo "✅ HF_HUB_OFFLINE=1(離線模式已啟用)" + else + echo "⚠️ HF_HUB_OFFLINE 未設定(模型可能在 runtime 下載)" + fi + + echo "" + echo "========================================" + echo "📊 模型驗證摘要" + echo "========================================" + echo "此驗證檢查模型是否已預下載至 Docker image。" + echo "若模型未預下載,轉檔功能仍可運作(首次使用時下載)。" + echo "========================================" + + # ======================================== + # 驗證 3: 基本功能驗證 + # ======================================== + - name: "⚙️ Basic Functionality Verification" + run: | + echo "========================================" + echo "⚙️ 基本功能驗證" + echo "========================================" + + IMAGE="${{ env.DOCKER_IMAGE }}:${{ steps.tag.outputs.version }}" + + # 啟動容器 + echo "🚀 啟動容器..." + docker run -d --name verify-test -p 3000:3000 "${IMAGE}" + + # 等待啟動 + echo "⏳ 等待容器啟動..." + for i in {1..60}; do + if docker exec verify-test curl -sf http://localhost:3000/healthcheck > /dev/null 2>&1; then + echo "✅ 容器在 ${i} 秒內啟動成功" + break + fi + if [ $i -eq 60 ]; then + echo "❌ 容器啟動超時" + docker logs verify-test + docker rm -f verify-test + exit 1 + fi + sleep 1 + done + + # 檢查關鍵工具 + echo "" + echo "📋 檢查關鍵轉檔工具..." + TOOLS=("libreoffice --version" "pandoc --version" "ffmpeg -version" "tesseract --version" "pdf2zh --help" "mineru --help" "babeldoc --help") + + for tool_cmd in "${TOOLS[@]}"; do + TOOL_NAME=$(echo "$tool_cmd" | cut -d' ' -f1) + echo -n " ${TOOL_NAME}: " + if docker exec verify-test $tool_cmd >/dev/null 2>&1; then + echo "✅" + else + echo "⚠️ 不可用或未安裝" + fi + done + + # 清理 + echo "" + echo "🧹 清理測試容器..." + docker rm -f verify-test + + echo "" + echo "========================================" + echo "✅ 基本功能驗證完成!" + echo "========================================" + + - name: Verification Summary + run: | + echo "## 🔍 Image Verification Results" >> $GITHUB_STEP_SUMMARY + echo "" >> $GITHUB_STEP_SUMMARY + echo "**Image:** \`${{ env.DOCKER_IMAGE }}:${{ steps.tag.outputs.version }}\`" >> $GITHUB_STEP_SUMMARY + echo "" >> $GITHUB_STEP_SUMMARY + echo "### ✅ Verification Passed" >> $GITHUB_STEP_SUMMARY + echo "" >> $GITHUB_STEP_SUMMARY + echo "- 🖥️ Headless Environment: Verified" >> $GITHUB_STEP_SUMMARY + echo "- 🤖 Model Download: Checked" >> $GITHUB_STEP_SUMMARY + echo "- ⚙️ Basic Functionality: Verified" >> $GITHUB_STEP_SUMMARY + + # ============================================================================== + # �📦 Create GitHub Release # ============================================================================== create-release: runs-on: ubuntu-latest - needs: build-and-push + needs: [build-and-push, verify-image] permissions: contents: write diff --git a/.github/workflows/upstream-sync.yml b/.github/workflows/upstream-sync.yml index 350e4b2..c3dccd2 100644 --- a/.github/workflows/upstream-sync.yml +++ b/.github/workflows/upstream-sync.yml @@ -233,9 +233,88 @@ jobs: echo "✅ All critical tools available" # ======================================== - # Test 3: 最小轉換測試 (txt → pdf) + # Test 3: Headless 環境驗證 # ======================================== - - name: "Test 3: Minimal conversion test" + - name: "Test 3: Headless environment verification" + run: | + echo "🖥️ Verifying headless environment..." + + # 檢查 DISPLAY 環境變數 + echo "" + echo "📋 檢查 DISPLAY 環境變數..." + DISPLAY_VAR=$(docker exec convertx-test printenv DISPLAY 2>/dev/null || echo "") + if [ -z "$DISPLAY_VAR" ]; then + echo "✅ DISPLAY 未設定(headless 模式正確)" + else + echo "⚠️ DISPLAY 已設定: ${DISPLAY_VAR}" + fi + + # 檢查 Xvfb + echo "" + echo "📋 檢查 Xvfb..." + if docker exec convertx-test which Xvfb 2>/dev/null; then + echo "⚠️ Xvfb 存在(但應不需要)" + else + echo "✅ Xvfb 不存在(headless 模式正確)" + fi + + # 驗證 LibreOffice headless + echo "" + echo "📋 驗證 LibreOffice headless..." + if docker exec convertx-test libreoffice --headless --version 2>/dev/null | head -1; then + echo "✅ LibreOffice headless 模式可用" + else + echo "❌ LibreOffice headless 模式失敗" + exit 1 + fi + + echo "" + echo "✅ Headless environment verification passed" + + # ======================================== + # Test 4: 模型下載驗證 + # ======================================== + - name: "Test 4: Model download verification" + run: | + echo "🤖 Verifying model downloads..." + + # 檢查 HuggingFace cache + echo "" + echo "📋 檢查 HuggingFace cache..." + HF_CACHE=$(docker exec convertx-test ls -la /root/.cache/huggingface/ 2>/dev/null || echo "NOT_FOUND") + if [ "$HF_CACHE" != "NOT_FOUND" ]; then + echo "✅ HuggingFace cache 存在" + else + echo "⚠️ HuggingFace cache 不存在(模型可能在其他位置)" + fi + + # 檢查 ONNX 模型 + echo "" + echo "📋 檢查 ONNX 模型..." + ONNX_COUNT=$(docker exec convertx-test find /root -name "*.onnx" -type f 2>/dev/null | wc -l || echo "0") + if [ "$ONNX_COUNT" -gt 0 ]; then + echo "✅ 找到 ${ONNX_COUNT} 個 ONNX 模型" + else + echo "⚠️ 未找到 ONNX 模型(可能動態下載)" + fi + + # 檢查離線模式環境變數 + echo "" + echo "📋 檢查離線模式環境變數..." + HF_OFFLINE=$(docker exec convertx-test printenv HF_HUB_OFFLINE 2>/dev/null || echo "") + if [ "$HF_OFFLINE" = "1" ]; then + echo "✅ HF_HUB_OFFLINE=1(離線模式已啟用)" + else + echo "⚠️ HF_HUB_OFFLINE 未設定" + fi + + echo "" + echo "✅ Model verification completed" + + # ======================================== + # Test 5: 最小轉換測試 (txt → pdf) + # ======================================== + - name: "Test 5: Minimal conversion test" id: test_conversion run: | echo "📄 Testing minimal conversion (txt → pdf)..." @@ -271,9 +350,9 @@ jobs: } # ======================================== - # Test 4: API 端點檢查 + # Test 6: API 端點檢查 # ======================================== - - name: "Test 4: API endpoints check" + - name: "Test 6: API endpoints check" run: | echo "🌐 Checking API endpoints..." @@ -321,6 +400,8 @@ jobs: echo "|----------|------|" >> $GITHUB_STEP_SUMMARY echo "| Container 啟動 | ✅ 通過 (${startup_time:-N/A}s) |" >> $GITHUB_STEP_SUMMARY echo "| 工具檢查 (libreoffice, pandoc, ffmpeg, tesseract) | ✅ 通過 |" >> $GITHUB_STEP_SUMMARY + echo "| Headless 環境驗證 | ✅ 通過 |" >> $GITHUB_STEP_SUMMARY + echo "| 模型下載驗證 | ✅ 通過 |" >> $GITHUB_STEP_SUMMARY echo "| 最小轉換測試 | ✅ 通過 |" >> $GITHUB_STEP_SUMMARY echo "| API 端點檢查 | ✅ 通過 |" >> $GITHUB_STEP_SUMMARY @@ -403,6 +484,8 @@ jobs: echo "|----------|------|" >> $GITHUB_STEP_SUMMARY echo "| Container 啟動 | ✅ |" >> $GITHUB_STEP_SUMMARY echo "| 工具存在性檢查 | ✅ |" >> $GITHUB_STEP_SUMMARY + echo "| Headless 環境驗證 | ✅ |" >> $GITHUB_STEP_SUMMARY + echo "| 模型下載驗證 | ✅ |" >> $GITHUB_STEP_SUMMARY echo "| 最小轉換測試 | ✅ |" >> $GITHUB_STEP_SUMMARY echo "| API 端點檢查 | ✅ |" >> $GITHUB_STEP_SUMMARY echo "" >> $GITHUB_STEP_SUMMARY