Merge branch 'develop' into fix/tech-debt-security-round2

- unified_render_service: 采纳PR方案FFMPEG_BIN常量，保留str(e)[:200]截断
- generation.py: 合并两边docstring（基础SSRF校验+重定向每跳校验

apps/api/app/middleware/versioning.py

@@ -61,7 +61,7 @@ class APIVersionMiddleware(BaseHTTPMiddleware):
 class VersionNotFoundMiddleware(BaseHTTPMiddleware):
     """处理已下线的 API 版本"""
 
-    SUNSET_VERSIONS: list[str] = []  # 已下线的版本列表
+    SUNSET_VERSIONS = []  # 已下线的版本列表
 
     async def dispatch(self, request: Request, call_next):
         version = self._extract_version(request.url.path)

apps/worker/video_processing/trim_engine.py

@@ -186,8 +186,8 @@ class TrimEngine:
             FFmpeg filter 字符串，如 "[0:v]trim=start=10:duration=5,setpts=PTS-STARTPTS[v0_trimmed]"
         """
         if trim.is_noop:
-            # 不裁剪，直接直通（仅重置时间戳）
-            return f"{input_label}setpts=PTS-STARTPTS{output_label}"
+            # 不裁剪，直接直通
+            return f"{input_label}copy{output_label}" if False else f"{input_label}setpts=PTS-STARTPTS{output_label}"
 
         parts: list[str] = []
 

apps/worker/worker_app/tasks/generation.py

@@ -460,6 +460,8 @@ def _verify_url_accessible(
     import urllib.request
     from urllib.parse import urljoin
 
+    from shared.url_security import UrlSecurityError, validate_url_safety
+
     from video_processing.url_security import UrlSecurityError, validate_url_safety
 
     # P0-1 SSRF 防护：请求前先校验 URL 安全性

infra/docker/worker.Dockerfile

@@ -1,104 +1,48 @@
 # ============================================================
-# Worker Dockerfile - 优化版（多阶段构建 + 镜像瘦身）
-# 优化项：
-#   1. 多阶段构建：builder 阶段安装编译依赖，runtime 阶段只保留运行时
-#   2. ffmpeg 静态编译替换：从 apt 安装(457MB) 改为静态二进制(~80MB)
-#   3. Python 依赖瘦身：strip .so 调试符号 + 清理测试文件 + 清理缓存
+# Worker Dockerfile - 专门用于 Celery Worker
+# 优化：依赖分层缓存，基础大包和业务依赖分开
 # ============================================================
 
-# ==================== Builder 阶段 ====================
-FROM git.xiaoxiajianji.com/xiaoxia/base/python:3.12-slim AS builder
+# 基础镜像：Python 3.12 + ffmpeg
+FROM git.xiaoxiajianji.com/xiaoxia/base/python:3.12-slim
 
-# 使用阿里云镜像加速
-RUN sed -i 's|deb.debian.org|mirrors.aliyun.com|g' /etc/apt/sources.list.d/debian.sources 2>/dev/null || \
-    sed -i 's|deb.debian.org|mirrors.aliyun.com|g' /etc/apt/sources.list 2>/dev/null || true
-
-# 安装编译工具（仅 builder 需要）
-RUN apt-get update && apt-get install -y --no-install-recommends \
-    gcc \
-    g++ \
-    python3-dev \
-    binutils \
-    wget \
-    xz-utils \
-    && rm -rf /var/lib/apt/lists/*
-
-# ---- 下载静态编译 ffmpeg ----
-# 使用 johnvansickle.com 的静态编译版本（业界标准）
-RUN cd /tmp \
-    && wget -q https://johnvansickle.com/ffmpeg/releases/ffmpeg-release-amd64-static.tar.xz \
-    && tar xf ffmpeg-release-amd64-static.tar.xz \
-    && cp ffmpeg-*-amd64-static/ffmpeg /usr/local/bin/ffmpeg \
-    && cp ffmpeg-*-amd64-static/ffprobe /usr/local/bin/ffprobe \
-    && chmod +x /usr/local/bin/ffmpeg /usr/local/bin/ffprobe \
-    && rm -rf ffmpeg-*
-
-# ---- 安装 Python 依赖 ----
-WORKDIR /tmp
-
-# 创建 venv
-RUN python -m venv /opt/venv
-ENV PATH="/opt/venv/bin:$PATH"
-
-# 基础依赖
-COPY requirements-base.txt /tmp/requirements-base.txt
-RUN pip install --no-cache-dir -i https://mirrors.aliyun.com/pypi/simple/ --trusted-host mirrors.aliyun.com \
-    -r /tmp/requirements-base.txt \
-    && rm /tmp/requirements-base.txt
-
-# Worker 专属大包
-COPY requirements-worker.txt /tmp/requirements-worker.txt
-RUN pip install --no-cache-dir -i https://mirrors.aliyun.com/pypi/simple/ --trusted-host mirrors.aliyun.com \
-    -r /tmp/requirements-worker.txt \
-    && rm /tmp/requirements-worker.txt
-
-# 业务依赖
-COPY requirements.txt /tmp/requirements.txt
-RUN pip install --no-cache-dir -i https://mirrors.aliyun.com/pypi/simple/ --trusted-host mirrors.aliyun.com \
-    -r /tmp/requirements.txt \
-    && rm /tmp/requirements.txt
-
-# ---- Python 依赖瘦身 ----
-# 1. strip .so 文件的调试符号（节省约 80-100MB）
-RUN find /opt/venv -name "*.so" -type f -exec strip --strip-all {} \; 2>/dev/null || true
-
-# 2. 清理测试文件（节省约 20MB）
-RUN find /opt/venv -type d -name "tests" -exec rm -rf {} + 2>/dev/null; \
-    find /opt/venv -type d -name "test" -exec rm -rf {} + 2>/dev/null; \
-    find /opt/venv -name "test_*.py" -delete 2>/dev/null || true
-
-# 3. 清理 .pyc 缓存和 __pycache__（节省约 10MB，运行时按需生成）
-RUN find /opt/venv -type d -name "__pycache__" -exec rm -rf {} + 2>/dev/null; \
-    find /opt/venv -name "*.pyc" -delete 2>/dev/null || true
-
-# 4. 清理 dist-info 中的文档
-RUN find /opt/venv -name "*.dist-info" -type d -exec sh -c 'rm -f "$1"/DESCRIPTION.rst "$1"/INSTALLER "$1"/LICENSE* "$1"/WHEEL "$1"/entry_points.txt' _ {} \; 2>/dev/null || true
-
-# ==================== Runtime 阶段 ====================
-FROM git.xiaoxiajianji.com/xiaoxia/base/python:3.12-slim AS runtime
-
-# 构建参数：版本号
+# 构建参数：版本号（CI 传入 commit hash）
 ARG APP_VERSION=dev
 
 # 使用阿里云镜像加速
 RUN sed -i 's|deb.debian.org|mirrors.aliyun.com|g' /etc/apt/sources.list.d/debian.sources 2>/dev/null || \
     sed -i 's|deb.debian.org|mirrors.aliyun.com|g' /etc/apt/sources.list 2>/dev/null || true
 
-# 安装最小运行时依赖（opencv-python-headless 需要 libglib2.0-0）
+# 安装系统依赖
 RUN apt-get update && apt-get install -y --no-install-recommends \
-    libglib2.0-0 \
+    ffmpeg \
+    libsm6 \
+    libxext6 \
+    libgl1 \
     && rm -rf /var/lib/apt/lists/*
 
-# 从 builder 复制 ffmpeg 静态二进制
-COPY --from=builder /usr/local/bin/ffmpeg /usr/local/bin/ffmpeg
-COPY --from=builder /usr/local/bin/ffprobe /usr/local/bin/ffprobe
-
-# 从 builder 复制 Python 虚拟环境
-COPY --from=builder /opt/venv /opt/venv
-
 # 设置工作目录
 WORKDIR /app
 
+# ---- 依赖分层：基础依赖（变化少，缓存命中率高）----
+COPY requirements-base.txt /tmp/requirements-base.txt
+
+RUN python -m venv /opt/venv \
+    && /opt/venv/bin/pip install --no-cache-dir -i https://mirrors.aliyun.com/pypi/simple/ --trusted-host mirrors.aliyun.com -r /tmp/requirements-base.txt \
+    && rm /tmp/requirements-base.txt
+
+# ---- 依赖分层：Worker 专属大包（视频处理，变化极少）----
+COPY requirements-worker.txt /tmp/requirements-worker.txt
+
+RUN /opt/venv/bin/pip install --no-cache-dir -i https://mirrors.aliyun.com/pypi/simple/ --trusted-host mirrors.aliyun.com -r /tmp/requirements-worker.txt \
+    && rm /tmp/requirements-worker.txt
+
+# ---- 依赖分层：业务依赖（变化频繁）----
+COPY requirements.txt /tmp/requirements.txt
+
+RUN /opt/venv/bin/pip install --no-cache-dir -i https://mirrors.aliyun.com/pypi/simple/ --trusted-host mirrors.aliyun.com -r /tmp/requirements.txt \
+    && rm /tmp/requirements.txt
+
 # 复制应用代码
 COPY apps/worker/ /app/apps/worker/
 COPY apps/api/app/config.py /app/apps/api/app/config.py
@@ -107,13 +51,13 @@ COPY packages/ /app/packages/
 COPY alembic.ini /app/alembic.ini
 COPY migrations/ /app/migrations/
 
-# 复制 Worker 启动脚本
+# 复制 Worker 启动脚本（支持 WORKER_CONCURRENCY 环境变量）
 COPY infra/docker/entrypoint-worker.sh /usr/local/bin/entrypoint-worker.sh
 RUN chmod +x /usr/local/bin/entrypoint-worker.sh
 
 # 设置 Python 路径
 ENV PATH="/opt/venv/bin:$PATH"
-ENV PYTHONPATH=/app:/app/packages
+ENV PYTHONPATH=/app
 ENV PYTHONUNBUFFERED=1
 ENV APP_VERSION=$APP_VERSION
 

packages/application/tts_job/audio_merger.py

@@ -11,7 +11,7 @@ import shutil
 import tempfile
 from subprocess import CalledProcessError, TimeoutExpired
 
-from packages.shared.ffmpeg_utils import FFMPEG_BIN, run_ffmpeg
+from shared.ffmpeg_utils import FFMPEG_BIN, run_ffmpeg
 
 logger = logging.getLogger(__name__)
 

pyproject.toml

@@ -80,6 +80,7 @@ select = [
 # 与原 setup.cfg flake8 配置对齐，确保不新增阻断
 ignore = [
   "E203",
+  "W503",
   "E501",  # line-too-long（black管）
   "E302",
   "E402",  # module-import-not-at-top（循环导入多）
@@ -96,6 +97,6 @@ ignore = [
 [tool.ruff.lint.per-file-ignores]
 "__init__.py" = ["F401", "F403", "F405"]
 "tests/*" = ["E402", "F401", "F841"]
-"packages/ports/*" = ["E301"]
+"packages/ports/*" = ["E301", "E704"]
 "apps/*/migrations/*" = ["ALL"]
 "alembic/*" = ["ALL"]

scripts/check_migration_safety.py

@@ -123,7 +123,7 @@ def get_new_migrations_via_diff(diff_target: str) -> List[Path]:
         return [REPO_ROOT / f for f in files]
     except subprocess.CalledProcessError as e:
         print(f"⚠️  git diff 失败（{diff_target}）：{e.stderr.strip()}")
-        print("   降级为检查所有迁移文件")
+        print(f"   降级为检查所有迁移文件")
         return sorted(ALEMBIC_VERSIONS_DIR.glob("*.py"))
 
 

scripts/ci_notify.py

@@ -125,7 +125,7 @@ def main() -> int:
     if mode != "start":
         fields.append({"is_short": True, "text": {"tag": "lark_md", "content": f"**耗时**\n{duration}"}})
     else:
-        fields.append({"is_short": True, "text": {"tag": "lark_md", "content": "**状态**\n进行中"}})
+        fields.append({"is_short": True, "text": {"tag": "lark_md", "content": f"**状态**\n进行中"}})
 
     if runner_name:
         fields.append({"is_short": True, "text": {"tag": "lark_md", "content": f"**Runner**\n{runner_name}"}})

scripts/ci_production_healthcheck.sh

@@ -1,483 +0,0 @@
-#!/bin/bash
-# ===========================================
-# CI Production 健康检查 + 自动回滚脚本（SSH 部署模式）
-# ===========================================
-#
-# 在 CI Runner 上执行，通过公网 URL 检查 Production 部署健康状态。
-# 不健康则通过 SSH 自动回滚到上一个版本的镜像。
-#
-# 用法:
-#   ./ci_production_healthcheck.sh
-#
-# 环境变量:
-#   PROD_API_URL          - Production API 公网地址 (默认 https://api.xiaoxiajianji.com)
-#   PROD_WEB_URL          - Production Web 公网地址 (默认 https://saas.xiaoxiajianji.com)
-#   HEALTH_CHECK_TIMEOUT  - 健康检查总超时秒数 (默认 180)
-#   SKIP_ROLLBACK         - 失败时不自动回滚 (true/false, 默认 false)
-#   SKIP_NOTIFY           - 跳过通知 (true/false, 默认 false)
-#   CI_NOTIFY_WEBHOOK     - 通知 Webhook URL
-#
-#   PRODUCTION_SSH_HOST   - 生产服务器 SSH 地址
-#   PRODUCTION_SSH_USER   - SSH 用户名 (默认 root)
-#   PRODUCTION_SSH_PORT   - SSH 端口 (默认 22222)
-#   PRODUCTION_SSH_KEY    - SSH 私钥内容
-#   REGISTRY_TOKEN        - Registry Token（回滚时拉取旧镜像需要）
-#
-#   GITHUB_SHA            - 当前 commit SHA
-#   GITHUB_REF_NAME       - tag 名 (如 v0.1.100)
-#   GITHUB_RUN_ID         - CI Run ID
-#   GITHUB_REPOSITORY     - 仓库名
-#   GITHUB_ACTOR          - 提交者
-
-set -eu
-
-SCRIPT_DIR="$(CDPATH= cd -- "$(dirname -- "$0")" && pwd)"
-
-# 配置
-PROD_API_URL="${PROD_API_URL:-https://api.xiaoxiajianji.com}"
-PROD_WEB_URL="${PROD_WEB_URL:-https://saas.xiaoxiajianji.com}"
-HEALTH_CHECK_TIMEOUT="${HEALTH_CHECK_TIMEOUT:-180}"
-SKIP_ROLLBACK="${SKIP_ROLLBACK:-false}"
-SKIP_NOTIFY="${SKIP_NOTIFY:-false}"
-
-PRODUCTION_SSH_HOST="${PRODUCTION_SSH_HOST:-47.98.113.167}"
-PRODUCTION_SSH_USER="${PRODUCTION_SSH_USER:-root}"
-PRODUCTION_SSH_PORT="${PRODUCTION_SSH_PORT:-22222}"
-
-REGISTRY="${REGISTRY:-git.xiaoxiajianji.com/xiaoxia/xiaoxia-saas}"
-REGISTRY_USER="${REGISTRY_USER:-xiaoxia}"
-
-# 颜色
-RED='\033[0;31m'
-GREEN='\033[0;32m'
-YELLOW='\033[1;33m'
-BLUE='\033[0;34m'
-NC='\033[0m'
-
-log_info()  { echo -e "${GREEN}[INFO]${NC} $1"; }
-log_warn()  { echo -e "${YELLOW}[WARN]${NC} $1"; }
-log_error() { echo -e "${RED}[ERROR]${NC} $1"; }
-log_step()  { echo -e "${BLUE}[STEP]${NC} $1"; }
-
-# ===========================================
-# SSH 工具函数
-# ===========================================
-SSH_KEY_PATH=""
-
-setup_ssh() {
-    # 查找或创建 SSH 密钥
-    if [ -f /root/.ssh/xiaoxia_runtime_builder ]; then
-        SSH_KEY_PATH="/root/.ssh/xiaoxia_runtime_builder"
-    elif [ -f "$HOME/.ssh/xiaoxia_runtime_builder" ]; then
-        SSH_KEY_PATH="$HOME/.ssh/xiaoxia_runtime_builder"
-    elif [ -n "${PRODUCTION_SSH_KEY:-}" ]; then
-        SSH_KEY_PATH="$HOME/.ssh/prod_deploy_key"
-        mkdir -p "$HOME/.ssh"
-        printf '%s\n' "$PRODUCTION_SSH_KEY" > "$SSH_KEY_PATH"
-        chmod 600 "$SSH_KEY_PATH"
-    else
-        log_error "没有可用的 SSH 密钥"
-        return 1
-    fi
-
-    ssh-keyscan -p "$PRODUCTION_SSH_PORT" -H "$PRODUCTION_SSH_HOST" >> ~/.ssh/known_hosts 2>/dev/null || true
-    log_info "SSH 已配置: ${PRODUCTION_SSH_USER}@${PRODUCTION_SSH_HOST}:${PRODUCTION_SSH_PORT}"
-}
-
-run_ssh() {
-    local cmd="$1"
-    ssh -p "$PRODUCTION_SSH_PORT" -i "$SSH_KEY_PATH" -o StrictHostKeyChecking=no \
-        "${PRODUCTION_SSH_USER}@${PRODUCTION_SSH_HOST}" "$cmd"
-}
-
-# ===========================================
-# 1. 记录部署前各服务的镜像版本（用于回滚）
-# ===========================================
-ROLLBACK_API_TAG=""
-ROLLBACK_WORKER_TAG=""
-ROLLBACK_WEB_TAG=""
-
-save_rollback_target() {
-    log_step "记录当前生产环境各服务镜像版本（回滚目标）..."
-
-    # 通过 SSH 获取当前运行的容器镜像
-    local api_image worker_image web_image
-    api_image=$(run_ssh "docker inspect --format '{{.Config.Image}}' xiaoxia-api-production 2>/dev/null || echo ''")
-    worker_image=$(run_ssh "docker inspect --format '{{.Config.Image}}' xiaoxia-worker-production 2>/dev/null || echo ''")
-    web_image=$(run_ssh "docker inspect --format '{{.Config.Image}}' xiaoxia-web-production 2>/dev/null || echo ''")
-
-    # 提取 tag（镜像名是 xiaoxia-saas-api:v0.1.100 格式）
-    ROLLBACK_API_TAG=$(echo "$api_image" | sed 's/.*://' || echo "")
-    ROLLBACK_WORKER_TAG=$(echo "$worker_image" | sed 's/.*://' || echo "")
-    ROLLBACK_WEB_TAG=$(echo "$web_image" | sed 's/.*://' || echo "")
-
-    log_info "  API:     ${ROLLBACK_API_TAG:-未知}"
-    log_info "  Worker:  ${ROLLBACK_WORKER_TAG:-未知}"
-    log_info "  Web:     ${ROLLBACK_WEB_TAG:-未知}"
-
-    # 验证三个服务版本是否一致
-    if [ -n "$ROLLBACK_API_TAG" ] && [ -n "$ROLLBACK_WORKER_TAG" ] && [ -n "$ROLLBACK_WEB_TAG" ]; then
-        if [ "$ROLLBACK_API_TAG" = "$ROLLBACK_WORKER_TAG" ] && [ "$ROLLBACK_API_TAG" = "$ROLLBACK_WEB_TAG" ]; then
-            log_info "  ✅ 三个服务版本一致: $ROLLBACK_API_TAG"
-            export ROLLBACK_TAG="$ROLLBACK_API_TAG"
-        else
-            log_warn "  ⚠️  三个服务版本不一致，回滚时将分别使用各自版本"
-            export ROLLBACK_API_TAG ROLLBACK_WORKER_TAG ROLLBACK_WEB_TAG
-            export ROLLBACK_TAG_MIXED="true"
-        fi
-    else
-        log_warn "  ⚠️  未能获取全部服务版本，回滚功能可能受限"
-    fi
-}
-
-# ===========================================
-# 2. 健康检查（公网视角）
-# ===========================================
-health_check() {
-    local timeout="$HEALTH_CHECK_TIMEOUT"
-    local start_time
-    start_time=$(date +%s)
-
-    log_step "公网健康检查（超时 ${timeout}s）..."
-    log_info "  API: ${PROD_API_URL}/health"
-    log_info "  Web: ${PROD_WEB_URL}/"
-
-    local api_ok=false
-    local web_ok=false
-    local api_docs_ok=false
-    local login_api_ok=false
-
-    while [ $(( $(date +%s) - start_time )) -lt "$timeout" ]; do
-        # 检查 API health
-        if [ "$api_ok" = false ] && curl -sf --max-time 10 "${PROD_API_URL}/health" >/dev/null 2>&1; then
-            log_info "✅ API 健康检查通过"
-            api_ok=true
-        fi
-
-        # 检查 Web 首页
-        if [ "$web_ok" = false ] && curl -sf --max-time 10 "$PROD_WEB_URL/" >/dev/null 2>&1; then
-            log_info "✅ Web 前端检查通过"
-            web_ok=true
-        fi
-
-        # 检查 API docs
-        if [ "$api_docs_ok" = false ]; then
-            HTTP_CODE=$(curl -s -o /dev/null -w "%{http_code}" --max-time 10 "${PROD_API_URL}/docs" 2>/dev/null || echo "000")
-            if [ "$HTTP_CODE" = "200" ]; then
-                log_info "✅ API Docs 检查通过"
-                api_docs_ok=true
-            fi
-        fi
-
-        # 检查登录 API
-        if [ "$login_api_ok" = false ]; then
-            HTTP_CODE=$(curl -s -o /dev/null -w "%{http_code}" --max-time 10 -X POST \
-                "${PROD_API_URL}/api/v1/auth/login" \
-                -H "Content-Type: application/json" \
-                -d '{"email":"smoke@test.com","password":"wrong"}' 2>/dev/null || echo "000")
-            if [ "$HTTP_CODE" = "401" ] || [ "$HTTP_CODE" = "422" ]; then
-                log_info "✅ 登录 API 检查通过（HTTP $HTTP_CODE，符合预期）"
-                login_api_ok=true
-            fi
-        fi
-
-        # 都通过了就退出
-        if [ "$api_ok" = true ] && [ "$web_ok" = true ] && [ "$api_docs_ok" = true ] && [ "$login_api_ok" = true ]; then
-            log_info "🎉 所有健康检查通过！"
-            return 0
-        fi
-
-        sleep 5
-    done
-
-    # 超时了
-    log_error "❌ 健康检查超时 (${timeout}s)"
-    [ "$api_ok" = false ]       && log_error "  - API health 未通过"
-    [ "$web_ok" = false ]       && log_error "  - Web 前端未通过"
-    [ "$api_docs_ok" = false ]  && log_error "  - API Docs 未通过"
-    [ "$login_api_ok" = false ] && log_error "  - 登录 API 未通过"
-
-    return 1
-}
-
-# ===========================================
-# 3. 执行回滚（SSH 重新部署旧版本）
-# ===========================================
-do_rollback() {
-    log_step "执行回滚：通过 SSH 重新部署旧版本镜像..."
-
-    local rollback_tag="${ROLLBACK_TAG:-}"
-    if [ -z "$rollback_tag" ] && [ "${ROLLBACK_TAG_MIXED:-}" != "true" ]; then
-        log_error "没有可回滚的版本记录，无法自动回滚"
-        return 1
-    fi
-
-    # 如果版本不一致，用 API 的版本作为回滚目标
-    if [ -z "$rollback_tag" ]; then
-        rollback_tag="$ROLLBACK_API_TAG"
-    fi
-
-    if [ -z "$rollback_tag" ]; then
-        log_error "无法确定回滚版本"
-        return 1
-    fi
-
-    log_info "回滚目标版本: $rollback_tag"
-
-    # 通过 SSH 在生产服务器上执行回滚部署
-    # 复用 Registry 方式部署脚本的逻辑，用旧版本 tag 重新部署
-    local rollback_script=$(cat << 'ROLLBACK_EOF'
-#!/bin/sh
-set -eu
-
-IMAGE_TAG="$1"
-REGISTRY_TOKEN="$2"
-REGISTRY="${REGISTRY:-git.xiaoxiajianji.com/xiaoxia/xiaoxia-saas}"
-REGISTRY_USER="${REGISTRY_USER:-xiaoxia}"
-
-ENV_FILE="${ENV_FILE:-/var/lib/xiaoxia-saas-production/.env}"
-GENERATED_DIR="${GENERATED_DIR:-/var/lib/xiaoxia-saas-production/generated}"
-LEGACY_ASSETS_DIR="${LEGACY_ASSETS_DIR:-/var/lib/xiaoxia-saas-production/legacy-assets}"
-
-echo "=== Rollback to $IMAGE_TAG ==="
-
-# 登录 Registry
-if [ -n "$REGISTRY_TOKEN" ]; then
-    REGISTRY_HOST=$(echo "$REGISTRY" | cut -d/ -f1)
-    printf %s "$REGISTRY_TOKEN" | docker login "$REGISTRY_HOST" -u "$REGISTRY_USER" --password-stdin 2>/dev/null || true
-fi
-
-# Pull 旧版本镜像
-LOCAL_API="xiaoxia-saas-api:${IMAGE_TAG}"
-LOCAL_WORKER="xiaoxia-saas-worker:${IMAGE_TAG}"
-LOCAL_WEB="xiaoxia-saas-web:${IMAGE_TAG}"
-
-docker pull "${REGISTRY}/xiaoxia-saas-api:${IMAGE_TAG}"
-docker pull "${REGISTRY}/xiaoxia-saas-worker:${IMAGE_TAG}"
-docker pull "${REGISTRY}/xiaoxia-saas-web:${IMAGE_TAG}"
-
-docker tag "${REGISTRY}/xiaoxia-saas-api:${IMAGE_TAG}" "$LOCAL_API"
-docker tag "${REGISTRY}/xiaoxia-saas-worker:${IMAGE_TAG}" "$LOCAL_WORKER"
-docker tag "${REGISTRY}/xiaoxia-saas-web:${IMAGE_TAG}" "$LOCAL_WEB"
-
-echo "Rollback images pulled."
-
-# 停止当前容器
-docker rm -f xiaoxia-api-production 2>/dev/null || true
-docker rm -f xiaoxia-worker-production 2>/dev/null || true
-docker rm -f xiaoxia-web-production 2>/dev/null || true
-
-LOG_OPTS="--log-driver json-file --log-opt max-size=50m --log-opt max-file=3"
-
-# 启动 API（回滚不跑 migration，因为新版本可能加了字段，回滚后代码是旧的但数据还在）
-echo "Starting API (rollback)..."
-docker run -d \
-  --name xiaoxia-api-production \
-  --env-file "$ENV_FILE" \
-  --network xiaoxia-net-production \
-  -p 127.0.0.1:8001:8000 \
-  -e APP_ENV=production \
-  -e APP_VERSION="$IMAGE_TAG" \
-  -e GENERATED_FILES_DIR=/app/generated \
-  -e GENERATED_FILES_URL_PREFIX=/generated-files \
-  -e PUBLIC_API_BASE_URL=https://api.xiaoxiajianji.com \
-  -v "$GENERATED_DIR:/app/generated" \
-  --restart unless-stopped \
-  --cpus 2 \
-  --memory 2g \
-  --health-cmd "python -c \"import urllib.request; urllib.request.urlopen('http://localhost:8000/health', timeout=5)\"" \
-  --health-interval 30s \
-  --health-timeout 10s \
-  --health-retries 3 \
-  --health-start-period 40s \
-  $LOG_OPTS \
-  "$LOCAL_API"
-
-# 启动 Worker
-echo "Starting Worker (rollback)..."
-docker run -d \
-  --name xiaoxia-worker-production \
-  --env-file "$ENV_FILE" \
-  --network xiaoxia-net-production \
-  -e APP_ENV=production \
-  -e APP_VERSION="$IMAGE_TAG" \
-  -e WORKER_CONCURRENCY=1 \
-  -e WORKER_MAX_TASKS_PER_CHILD=100 \
-  -e GENERATED_FILES_DIR=/app/generated \
-  -e GENERATED_FILES_URL_PREFIX=/generated-files \
-  -e PUBLIC_API_BASE_URL=https://api.xiaoxiajianji.com \
-  -v "$GENERATED_DIR:/app/generated" \
-  --restart unless-stopped \
-  --cpus 2 \
-  --memory 2g \
-  --health-cmd "sh -c \"grep -q celery /proc/1/cmdline || exit 1\"" \
-  --health-interval 30s \
-  --health-timeout 10s \
-  --health-retries 3 \
-  --health-start-period 30s \
-  $LOG_OPTS \
-  "$LOCAL_WORKER"
-
-# 启动 Web
-LEGACY_VOLUME=""
-if [ -d "$LEGACY_ASSETS_DIR" ] && [ "$(ls -A "$LEGACY_ASSETS_DIR" 2>/dev/null)" ]; then
-  LEGACY_VOLUME="-v ${LEGACY_ASSETS_DIR}:/usr/share/nginx/html/assets-legacy/assets:ro"
-fi
-
-echo "Starting Web (rollback)..."
-docker run -d \
-  --name xiaoxia-web-production \
-  --network xiaoxia-net-production \
-  -p 127.0.0.1:3002:80 \
-  --restart unless-stopped \
-  --cpus 0.5 \
-  --memory 512m \
-  $LEGACY_VOLUME \
-  --health-cmd "wget --spider -q http://127.0.0.1:80" \
-  --health-interval 30s \
-  --health-timeout 5s \
-  --health-retries 3 \
-  $LOG_OPTS \
-  "$LOCAL_WEB"
-
-# 等待 API 健康
-echo "Waiting for API (rollback)..."
-i=0
-while [ "$i" -lt 40 ]; do
-  if curl -sf --max-time 5 http://127.0.0.1:8001/health >/dev/null 2>&1; then
-    echo "API healthy (rollback)."
-    break
-  fi
-  i=$((i + 1))
-  sleep 3
-done
-
-# 等待 Web 健康
-echo "Waiting for Web (rollback)..."
-i=0
-while [ "$i" -lt 15 ]; do
-  if curl -sf --max-time 5 http://127.0.0.1:3002/ >/dev/null 2>&1; then
-    echo "Web healthy (rollback)."
-    break
-  fi
-  i=$((i + 1))
-  sleep 2
-done
-
-echo "=== Rollback complete: $IMAGE_TAG ==="
-docker ps --format "table {{.Names}}\t{{.Status}}\t{{.Image}}" | grep production
-ROLLBACK_EOF
-)
-
-    # 将脚本 base64 编码后通过 SSH 执行
-    local script_b64
-    script_b64=$(echo "$rollback_script" | base64 -w 0)
-
-    log_info "在生产服务器上执行回滚脚本..."
-    if run_ssh "echo '$script_b64' | base64 -d | sh -s -- '$rollback_tag' '${REGISTRY_TOKEN:-}'" 2>&1; then
-        log_info "✅ 回滚命令执行完成"
-        return 0
-    else
-        log_error "❌ 回滚命令执行失败"
-        return 1
-    fi
-}
-
-# ===========================================
-# 4. 发送通知
-# ===========================================
-send_notification() {
-    local status="$1"   # success / failure / rollback
-    local detail="$2"
-
-    if [ "${SKIP_NOTIFY:-false}" = "true" ]; then
-        log_info "跳过通知（SKIP_NOTIFY=true）"
-        return 0
-    fi
-
-    local webhook="${CI_NOTIFY_WEBHOOK:-}"
-    if [ -z "$webhook" ]; then
-        log_warn "未配置 CI_NOTIFY_WEBHOOK，跳过通知"
-        return 0
-    fi
-
-    if [ -f "$SCRIPT_DIR/deploy_notify.py" ]; then
-        python3 "$SCRIPT_DIR/deploy_notify.py" \
-            --status "$status" \
-            --detail "$detail" \
-            --webhook "$webhook" \
-            --env production \
-            2>/dev/null || log_warn "通知发送失败（非致命）"
-    else
-        log_warn "找不到 deploy_notify.py，跳过通知"
-    fi
-
-    # 标记：通知已由健康检查脚本发出，避免 CI 兜底通知重复发送
-    echo "$status" > /tmp/prod_deploy_notification_sent
-}
-
-# ===========================================
-# 主流程
-# ===========================================
-main() {
-    echo ""
-    echo "=========================================="
-    echo "  CI Production 健康检查 + 自动回滚"
-    echo "=========================================="
-    echo ""
-
-    local deploy_status="success"
-    local deploy_detail=""
-
-    # 1. 设置 SSH
-    if ! setup_ssh; then
-        log_error "SSH 配置失败，无法执行回滚"
-    fi
-
-    # 2. 记录部署前状态（回滚目标）
-    save_rollback_target || true
-
-    # 3. 健康检查（公网视角）
-    if ! health_check; then
-        log_error "健康检查失败"
-        deploy_status="failure"
-        deploy_detail="公网健康检查超时，部署后服务未正常响应"
-
-        # 自动回滚
-        if [ "${SKIP_ROLLBACK:-false}" != "true" ]; then
-            log_warn "开始自动回滚..."
-            if do_rollback; then
-                deploy_status="rollback"
-                deploy_detail="健康检查失败，已自动回滚到上一版本 (${ROLLBACK_TAG:-未知})"
-
-                # 回滚后再检查一下公网状态
-                log_info "回滚完成，重新检查公网健康状态..."
-                if health_check; then
-                    log_info "✅ 回滚后服务已恢复"
-                    deploy_detail="${deploy_detail}，回滚后服务已恢复"
-                else
-                    log_error "⚠️  回滚后健康检查仍未通过，请手动排查"
-                    deploy_detail="${deploy_detail}，但回滚后仍未恢复，请紧急排查"
-                fi
-            else
-                deploy_detail="健康检查失败且回滚失败，请手动排查"
-            fi
-        fi
-
-        send_notification "$deploy_status" "$deploy_detail"
-
-        # 失败时退出非零，让 CI Job 标记为失败
-        exit 1
-    fi
-
-    # 4. 成功
-    log_info ""
-    log_info "=================================="
-    log_info "  ✅ Production 部署成功！"
-    log_info "=================================="
-
-    deploy_detail="部署成功，所有健康检查通过 (${GITHUB_REF_NAME:-未知版本})"
-    send_notification "success" "$deploy_detail"
-}
-
-main "$@"

scripts/ci_staging_deploy.sh

@@ -1,250 +0,0 @@
-#!/bin/sh
-# ===========================================
-# Staging 部署脚本（SSH 模式）
-# ===========================================
-# 通过 SSH 在 staging 服务器上执行
-#
-# 环境变量:
-#   IMAGE_TAG       - 镜像版本 tag（如 commit SHA 或分支名）
-#   REGISTRY_TOKEN  - Registry 访问令牌
-#   REGISTRY        - Registry 地址（默认 git.xiaoxiajianji.com/xiaoxia/xiaoxia-saas）
-#   REGISTRY_USER   - Registry 用户名（默认 xiaoxia）
-#   ENV_FILE        - 环境变量文件路径
-#   GENERATED_DIR   - 生成文件目录
-#   SKIP_MIGRATION  - 跳过数据库迁移（true/false，默认 false）
-
-set -eu
-
-IMAGE_TAG="${IMAGE_TAG:-}"
-REGISTRY="${REGISTRY:-git.xiaoxiajianji.com/xiaoxia/xiaoxia-saas}"
-REGISTRY_USER="${REGISTRY_USER:-xiaoxia}"
-REGISTRY_TOKEN="${REGISTRY_TOKEN:-}"
-
-ENV_FILE="${ENV_FILE:-/var/lib/xiaoxia-saas-staging/.env}"
-GENERATED_DIR="${GENERATED_DIR:-/var/lib/xiaoxia-saas-staging/generated}"
-LEGACY_ASSETS_DIR="${LEGACY_ASSETS_DIR:-/var/lib/xiaoxia-saas-staging/legacy-assets}"
-
-SKIP_MIGRATION="${SKIP_MIGRATION:-false}"
-
-if [ -z "$IMAGE_TAG" ]; then
-  echo "ERROR: IMAGE_TAG is required"
-  exit 1
-fi
-
-test -f "$ENV_FILE"
-mkdir -p "$GENERATED_DIR"
-mkdir -p "$LEGACY_ASSETS_DIR"
-
-echo "=========================================="
-echo "  Staging 部署 - $IMAGE_TAG"
-echo "=========================================="
-
-# ---- 登录 Registry ----
-if [ -n "$REGISTRY_TOKEN" ]; then
-  echo "Logging in to registry: $REGISTRY"
-  REGISTRY_HOST=$(echo "$REGISTRY" | cut -d/ -f1)
-  printf %s "$REGISTRY_TOKEN" | docker login "$REGISTRY_HOST" -u "$REGISTRY_USER" --password-stdin 2>/dev/null || {
-    echo "WARN: docker login failed, will try to pull anyway"
-  }
-fi
-
-# ---- Pull 新版本镜像 ----
-REGISTRY_API="${REGISTRY}/xiaoxia-saas-api:${IMAGE_TAG}"
-REGISTRY_WORKER="${REGISTRY}/xiaoxia-saas-worker:${IMAGE_TAG}"
-REGISTRY_WEB="${REGISTRY}/xiaoxia-saas-web:${IMAGE_TAG}"
-
-LOCAL_API="xiaoxia-saas-api:${IMAGE_TAG}"
-LOCAL_WORKER="xiaoxia-saas-worker:${IMAGE_TAG}"
-LOCAL_WEB="xiaoxia-saas-web:${IMAGE_TAG}"
-
-echo "Pulling API image..."
-docker pull "$REGISTRY_API"
-echo "Pulling Worker image..."
-docker pull "$REGISTRY_WORKER"
-echo "Pulling Web image..."
-docker pull "$REGISTRY_WEB"
-
-# Re-tag 成本地名
-docker tag "$REGISTRY_API" "$LOCAL_API"
-docker tag "$REGISTRY_WORKER" "$LOCAL_WORKER"
-docker tag "$REGISTRY_WEB" "$LOCAL_WEB"
-echo "All images pulled and tagged."
-
-# ---- 备份 legacy assets ----
-echo "Backing up legacy assets from current web container..."
-if docker inspect xiaoxia-web-staging >/dev/null 2>&1; then
-  _tmpdir="/tmp/legacy-assets-$$"
-  rm -rf "$_tmpdir"
-  mkdir -p "$_tmpdir"
-  docker cp xiaoxia-web-staging:/usr/share/nginx/html/assets/. "$_tmpdir/" 2>/dev/null || true
-  # 只有目录非空才拷贝，避免覆盖有内容的 legacy assets
-  if [ -d "$_tmpdir" ] && [ "$(ls -A "$_tmpdir" 2>/dev/null)" ]; then
-    cp -an "$_tmpdir"/. "$LEGACY_ASSETS_DIR"/ 2>/dev/null || true
-    echo "Legacy assets backed up: $(ls "$_tmpdir" | wc -l) files"
-  fi
-  rm -rf "$_tmpdir"
-else
-  echo "No existing web container, skipping legacy assets backup"
-fi
-
-# 清理 7 天前的 legacy assets
-if [ -d "$LEGACY_ASSETS_DIR" ]; then
-  find "$LEGACY_ASSETS_DIR" -type f -mtime +7 -delete 2>/dev/null || true
-  echo "Legacy assets cleanup done (retain 7 days)"
-fi
-
-# ---- 检查基础设施容器 ----
-echo "Checking infrastructure containers..."
-for c in xiaoxia-postgres-staging xiaoxia-redis-staging; do
-  if ! docker inspect "$c" >/dev/null 2>&1; then
-    echo "ERROR: Required container not found: $c"
-    exit 1
-  fi
-  state=$(docker inspect -f '{{.State.Status}}' "$c")
-  if [ "$state" != "running" ]; then
-    echo "ERROR: Container not running: $c ($state)"
-    exit 1
-  fi
-done
-
-# ---- 创建网络（不存在则创建） ----
-docker network create xiaoxia-net-staging 2>/dev/null || true
-
-# ---- 数据库迁移 ----
-if [ "$SKIP_MIGRATION" != "true" ]; then
-  echo "Running database migrations..."
-  docker run --rm \
-    --env-file "$ENV_FILE" \
-    --network xiaoxia-net-staging \
-    -e APP_ENV=staging \
-    "$LOCAL_API" sh -c "cd /app && alembic upgrade head"
-  echo "Migrations completed."
-else
-  echo "Skipping migrations (SKIP_MIGRATION=true)"
-fi
-
-# ---- 停止旧容器 ----
-echo "Stopping old containers..."
-docker rm -f xiaoxia-api-staging 2>/dev/null || true
-docker rm -f xiaoxia-worker-staging 2>/dev/null || true
-docker rm -f xiaoxia-web-staging 2>/dev/null || true
-
-LOG_OPTS="--log-driver json-file --log-opt max-size=50m --log-opt max-file=3"
-
-# ---- 启动 API ----
-echo "Starting API container..."
-docker run -d \
-  --name xiaoxia-api-staging \
-  --env-file "$ENV_FILE" \
-  --network xiaoxia-net-staging \
-  -p 127.0.0.1:8000:8000 \
-  -e APP_ENV=staging \
-  -e APP_VERSION="$IMAGE_TAG" \
-  -e GENERATED_FILES_DIR=/app/generated \
-  -e GENERATED_FILES_URL_PREFIX=/generated-files \
-  -e PUBLIC_API_BASE_URL=https://staging-api.xiaoxiajianji.com \
-  -v "$GENERATED_DIR:/app/generated" \
-  --restart unless-stopped \
-  --health-cmd "python -c \"import urllib.request; urllib.request.urlopen('http://localhost:8000/health', timeout=5)\"" \
-  --health-interval 30s \
-  --health-timeout 10s \
-  --health-retries 3 \
-  --health-start-period 40s \
-  $LOG_OPTS \
-  "$LOCAL_API"
-
-# ---- 启动 Worker ----
-echo "Starting Worker container..."
-docker run -d \
-  --name xiaoxia-worker-staging \
-  --env-file "$ENV_FILE" \
-  --network xiaoxia-net-staging \
-  -e APP_ENV=staging \
-  -e APP_VERSION="$IMAGE_TAG" \
-  -e WORKER_CONCURRENCY=1 \
-  -e WORKER_MAX_TASKS_PER_CHILD=100 \
-  -e GENERATED_FILES_DIR=/app/generated \
-  -e GENERATED_FILES_URL_PREFIX=/generated-files \
-  -e PUBLIC_API_BASE_URL=https://staging-api.xiaoxiajianji.com \
-  -v "$GENERATED_DIR:/app/generated" \
-  --restart unless-stopped \
-  --health-cmd "sh -c \"grep -q celery /proc/1/cmdline || exit 1\"" \
-  --health-interval 30s \
-  --health-timeout 10s \
-  --health-retries 3 \
-  --health-start-period 30s \
-  $LOG_OPTS \
-  "$LOCAL_WORKER"
-
-# ---- 启动 Web ----
-LEGACY_VOLUME=""
-if [ -d "$LEGACY_ASSETS_DIR" ] && [ "$(ls -A "$LEGACY_ASSETS_DIR" 2>/dev/null)" ]; then
-  LEGACY_VOLUME="-v ${LEGACY_ASSETS_DIR}:/usr/share/nginx/html/assets-legacy/assets:ro"
-  echo "Web container: legacy assets mounted (fallback)"
-else
-  echo "Web container: no legacy assets to mount"
-fi
-
-echo "Starting Web container..."
-docker run -d \
-  --name xiaoxia-web-staging \
-  --network xiaoxia-net-staging \
-  -p 127.0.0.1:3001:80 \
-  --restart unless-stopped \
-  $LEGACY_VOLUME \
-  --health-cmd "wget --spider -q http://127.0.0.1:80" \
-  --health-interval 30s \
-  --health-timeout 5s \
-  --health-retries 3 \
-  $LOG_OPTS \
-  "$LOCAL_WEB"
-
-# ---- 等待 API 健康 ----
-echo "Waiting for API to become healthy..."
-i=0
-while [ "$i" -lt 40 ]; do
-  if curl -sf --max-time 5 http://127.0.0.1:8000/health >/dev/null 2>&1; then
-    echo "API is healthy!"
-    break
-  fi
-  i=$((i + 1))
-  echo "  Waiting... ($i/40)"
-  sleep 3
-done
-
-if [ "$i" -ge 40 ]; then
-  echo "ERROR: API did not become healthy within 120s"
-  docker logs --tail 50 xiaoxia-api-staging
-  exit 1
-fi
-
-# ---- 等待 Web 健康 ----
-echo "Waiting for Web to become healthy..."
-i=0
-while [ "$i" -lt 15 ]; do
-  if curl -sf --max-time 5 http://127.0.0.1:3001/ >/dev/null 2>&1; then
-    echo "Web is healthy!"
-    break
-  fi
-  i=$((i + 1))
-  echo "  Waiting... ($i/15)"
-  sleep 2
-done
-
-if [ "$i" -ge 15 ]; then
-  echo "ERROR: Web did not become healthy within 30s"
-  docker logs --tail 30 xiaoxia-web-staging
-  exit 1
-fi
-
-# ---- 清理旧镜像 ----
-echo "Cleaning up old images..."
-docker image prune -af --filter "until=168h" 2>/dev/null || true
-docker builder prune -af --filter "until=168h" 2>/dev/null || true
-
-echo ""
-echo "=== Staging deployment complete ==="
-echo "API:      http://127.0.0.1:8000"
-echo "Web:      http://127.0.0.1:3001"
-echo "Version:  $IMAGE_TAG"
-docker ps --format "table {{.Names}}\t{{.Status}}\t{{.Image}}" | grep staging

scripts/ci_staging_healthcheck.sh

@@ -1,10 +1,10 @@
 #!/bin/bash
 # ===========================================
-# CI Staging 健康检查 + 自动回滚脚本（SSH 部署模式）
+# CI Staging 健康检查 + 自动回滚脚本（Watchtower 模式）
 # ===========================================
 #
 # 在 CI Runner 上执行，通过公网 URL 检查 Staging 部署健康状态。
-# 不健康则通过 SSH 自动回滚到上一个版本的镜像。
+# 不健康则自动回滚（把 :staging tag 指回旧版本镜像，Watchtower 检测到后自动回滚）。
 #
 # 用法:
 #   ./ci_staging_healthcheck.sh
@@ -13,15 +13,14 @@
 #   STAGING_API_URL       - Staging API 地址 (默认 https://staging-api.xiaoxiajianji.com)
 #   STAGING_WEB_URL       - Staging Web 地址 (默认 https://staging.xiaoxiajianji.com)
 #   HEALTH_CHECK_TIMEOUT  - 健康检查总超时秒数 (默认 120)
+#   WATCHTOWER_WAIT       - 等待 Watchtower 检测更新的秒数 (默认 90)
 #   SKIP_ROLLBACK         - 失败时不自动回滚 (true/false, 默认 false)
 #   SKIP_NOTIFY           - 跳过通知 (true/false, 默认 false)
 #   CI_NOTIFY_WEBHOOK     - 通知 Webhook URL
 #
-#   STAGING_SSH_HOST      - Staging 服务器 SSH 地址 (默认 47.98.113.167)
-#   STAGING_SSH_USER      - SSH 用户名 (默认 root)
-#   STAGING_SSH_PORT      - SSH 端口 (默认 22222)
-#   STAGING_SSH_KEY       - SSH 私钥内容
-#   REGISTRY_TOKEN        - Registry Token（回滚时拉取旧镜像需要）
+#   REGISTRY_TOKEN        - Gitea Registry Token（回滚时需要）
+#   REGISTRY_HOST         - Registry 地址 (默认 git.xiaoxiajianji.com)
+#   REGISTRY_REPO         - 仓库路径 (默认 xiaoxia/xiaoxia-saas)
 #
 #   GITHUB_SHA            - 当前 commit SHA
 #   GITHUB_REF_NAME       - 分支名
@@ -37,15 +36,12 @@ SCRIPT_DIR="$(CDPATH= cd -- "$(dirname -- "$0")" && pwd)"
 STAGING_API_URL="${STAGING_API_URL:-https://staging-api.xiaoxiajianji.com}"
 STAGING_WEB_URL="${STAGING_WEB_URL:-https://staging.xiaoxiajianji.com}"
 HEALTH_CHECK_TIMEOUT="${HEALTH_CHECK_TIMEOUT:-120}"
+WATCHTOWER_WAIT="${WATCHTOWER_WAIT:-90}"
 SKIP_ROLLBACK="${SKIP_ROLLBACK:-false}"
 SKIP_NOTIFY="${SKIP_NOTIFY:-false}"
 
-STAGING_SSH_HOST="${STAGING_SSH_HOST:-47.98.113.167}"
-STAGING_SSH_USER="${STAGING_SSH_USER:-root}"
-STAGING_SSH_PORT="${STAGING_SSH_PORT:-22222}"
-
-REGISTRY="${REGISTRY:-git.xiaoxiajianji.com/xiaoxia/xiaoxia-saas}"
-REGISTRY_USER="${REGISTRY_USER:-xiaoxia}"
+REGISTRY_HOST="${REGISTRY_HOST:-git.xiaoxiajianji.com}"
+REGISTRY_REPO="${REGISTRY_REPO:-xiaoxia/xiaoxia-saas}"
 
 # 颜色
 RED='\033[0;31m'
@@ -60,85 +56,59 @@ log_error() { echo -e "${RED}[ERROR]${NC} $1"; }
 log_step()  { echo -e "${BLUE}[STEP]${NC} $1"; }
 
 # ===========================================
-# SSH 工具函数
+# 1. 记录部署前 :staging 镜像 digest（用于回滚）
 # ===========================================
-SSH_KEY_PATH=""
-
-setup_ssh() {
-    # 查找或创建 SSH 密钥
-    if [ -f /root/.ssh/xiaoxia_runtime_builder ]; then
-        SSH_KEY_PATH="/root/.ssh/xiaoxia_runtime_builder"
-    elif [ -f "$HOME/.ssh/xiaoxia_runtime_builder" ]; then
-        SSH_KEY_PATH="$HOME/.ssh/xiaoxia_runtime_builder"
-    elif [ -n "${STAGING_SSH_KEY:-}" ]; then
-        SSH_KEY_PATH="$HOME/.ssh/staging_deploy_key"
-        mkdir -p "$HOME/.ssh"
-        printf '%s\n' "$STAGING_SSH_KEY" > "$SSH_KEY_PATH"
-        chmod 600 "$SSH_KEY_PATH"
-    else
-        log_error "没有可用的 SSH 密钥"
-        return 1
-    fi
-
-    ssh-keyscan -p "$STAGING_SSH_PORT" -H "$STAGING_SSH_HOST" >> ~/.ssh/known_hosts 2>/dev/null || true
-    log_info "SSH 已配置: ${STAGING_SSH_USER}@${STAGING_SSH_HOST}:${STAGING_SSH_PORT}"
-}
-
-run_ssh() {
-    local cmd="$1"
-    ssh -p "$STAGING_SSH_PORT" -i "$SSH_KEY_PATH" -o StrictHostKeyChecking=no \
-        "${STAGING_SSH_USER}@${STAGING_SSH_HOST}" "$cmd"
-}
-
-# ===========================================
-# 1. 记录部署前各服务的镜像版本（用于回滚）
-# ===========================================
-ROLLBACK_API_TAG=""
-ROLLBACK_WORKER_TAG=""
-ROLLBACK_WEB_TAG=""
-
 save_rollback_target() {
-    log_step "记录当前 staging 各服务镜像版本（回滚目标）..."
+    log_step "记录当前 :staging 镜像 digest（回滚目标）..."
 
-    # 通过 SSH 获取当前运行的容器镜像
-    local api_image worker_image web_image
-    api_image=$(run_ssh "docker inspect --format '{{.Config.Image}}' xiaoxia-api-staging 2>/dev/null || echo ''")
-    worker_image=$(run_ssh "docker inspect --format '{{.Config.Image}}' xiaoxia-worker-staging 2>/dev/null || echo ''")
-    web_image=$(run_ssh "docker inspect --format '{{.Config.Image}}' xiaoxia-web-staging 2>/dev/null || echo ''")
-
-    # 提取 tag（镜像名是 xiaoxia-saas-api:abc123 或 git.xiaoxiajianji.com/.../xiaoxia-saas-api:staging 格式）
-    ROLLBACK_API_TAG=$(echo "$api_image" | sed 's/.*://' || echo "")
-    ROLLBACK_WORKER_TAG=$(echo "$worker_image" | sed 's/.*://' || echo "")
-    ROLLBACK_WEB_TAG=$(echo "$web_image" | sed 's/.*://' || echo "")
-
-    log_info "  API:     ${ROLLBACK_API_TAG:-未知}"
-    log_info "  Worker:  ${ROLLBACK_WORKER_TAG:-未知}"
-    log_info "  Web:     ${ROLLBACK_WEB_TAG:-未知}"
-
-    # 验证三个服务版本是否一致
-    if [ -n "$ROLLBACK_API_TAG" ] && [ -n "$ROLLBACK_WORKER_TAG" ] && [ -n "$ROLLBACK_WEB_TAG" ]; then
-        if [ "$ROLLBACK_API_TAG" = "$ROLLBACK_WORKER_TAG" ] && [ "$ROLLBACK_API_TAG" = "$ROLLBACK_WEB_TAG" ]; then
-            log_info "  ✅ 三个服务版本一致: $ROLLBACK_API_TAG"
-            export ROLLBACK_TAG="$ROLLBACK_API_TAG"
-        else
-            log_warn "  ⚠️  三个服务版本不一致，回滚时将分别使用各自版本"
-            export ROLLBACK_API_TAG ROLLBACK_WORKER_TAG ROLLBACK_WEB_TAG
-            export ROLLBACK_TAG_MIXED="true"
-        fi
-    else
-        log_warn "  ⚠️  未能获取全部服务版本，回滚功能可能受限"
+    # 登录 Registry
+    if [ -n "${REGISTRY_TOKEN:-}" ]; then
+        printf '%s' "${REGISTRY_TOKEN}" | docker login "${REGISTRY_HOST}" -u xiaoxia --password-stdin 2>/dev/null || true
     fi
+
+    ROLLBACK_IMAGES=""
+
+    for svc in api worker web; do
+        image="${REGISTRY_HOST}/${REGISTRY_REPO}/xiaoxia-saas-${svc}:staging"
+        # pull 当前 :staging 镜像获取 digest
+        if docker pull "$image" 2>/dev/null; then
+            digest=$(docker inspect -f '{{index .RepoDigests 0}}' "$image" 2>/dev/null || echo "")
+            if [ -n "$digest" ]; then
+                ROLLBACK_IMAGES="${ROLLBACK_IMAGES}${svc}:${digest} "
+                log_info "  $svc: $digest"
+            else
+                log_warn "  $svc: 无法获取 digest，该服务将不参与回滚"
+            fi
+        else
+            log_warn "  $svc: 拉取 :staging 镜像失败，该服务将不参与回滚"
+        fi
+    done
+
+    if [ -z "$ROLLBACK_IMAGES" ]; then
+        log_warn "没有获取到任何可回滚镜像，回滚功能将不可用"
+    fi
+
+    export ROLLBACK_IMAGES
 }
 
 # ===========================================
-# 2. 健康检查（公网视角）
+# 2. 等待 Watchtower 更新
+# ===========================================
+wait_for_watchtower() {
+    log_step "等待 Watchtower 检测新镜像并更新（${WATCHTOWER_WAIT}s）..."
+    sleep "$WATCHTOWER_WAIT"
+    log_info "等待结束，开始健康检查"
+}
+
+# ===========================================
+# 3. 健康检查
 # ===========================================
 health_check() {
     local timeout="$HEALTH_CHECK_TIMEOUT"
     local start_time
     start_time=$(date +%s)
 
-    log_step "公网健康检查（超时 ${timeout}s）..."
+    log_step "健康检查（超时 ${timeout}s）..."
     log_info "  API: ${STAGING_API_URL}/health"
     log_info "  Web: ${STAGING_WEB_URL}/"
 
@@ -201,183 +171,69 @@ health_check() {
 }
 
 # ===========================================
-# 3. 执行回滚（SSH 重新部署旧版本）
+# 4. 执行回滚（把 :staging tag 改回旧版本）
 # ===========================================
 do_rollback() {
-    log_step "执行回滚：通过 SSH 重新部署旧版本镜像..."
+    log_step "执行回滚：将 :staging tag 指回旧版本镜像..."
 
-    local rollback_tag="${ROLLBACK_TAG:-}"
-    if [ -z "$rollback_tag" ] && [ "${ROLLBACK_TAG_MIXED:-}" != "true" ]; then
-        log_error "没有可回滚的版本记录，无法自动回滚"
+    if [ -z "${ROLLBACK_IMAGES:-}" ]; then
+        log_error "没有可回滚的镜像记录，无法自动回滚"
         return 1
     fi
 
-    # 如果版本不一致，用 API 的版本作为回滚目标
-    if [ -z "$rollback_tag" ]; then
-        rollback_tag="$ROLLBACK_API_TAG"
-    fi
+    local rollback_ok=true
 
-    if [ -z "$rollback_tag" ]; then
-        log_error "无法确定回滚版本"
-        return 1
-    fi
+    for entry in $ROLLBACK_IMAGES; do
+        svc="${entry%%:*}"
+        digest="${entry#*:}"
+        # digest 是 host/repo/image@sha256:xxx 格式，entry 里只取了第一部分，需要重新组合
+        # 实际上 ROLLBACK_IMAGES 存的是 svc:repo/image@sha256:xxx
+        # 让我们重新从完整digest中提取
+        image_digest="${entry#*:}:${entry#*:*:}"
+        # 不对，格式是 svc:REGISTRY_HOST/REGISTRY_REPO/xiaoxia-saas-svc@sha256:xxx
+        # 让我重新组织一下
+        :
+    done
 
-    log_info "回滚目标版本: $rollback_tag"
+    # 上面的解析有问题，重新来
+    # ROLLBACK_IMAGES 格式: "api:digest1 worker:digest2 web:digest3"
+    # 但digest本身带冒号(@sha256:xxx)，所以分隔有问题
+    # 改用数组方式
+    log_info "开始回滚各服务..."
 
-    # 构建回滚脚本（直接部署旧版本镜像，不跑 migration）
-    local rollback_script=$(cat << 'ROLLBACK_EOF'
-#!/bin/sh
-set -eu
+    local IFS=' '
+    for pair in $ROLLBACK_IMAGES; do
+        svc="${pair%%:*}"
+        # 剩余部分是完整的 digest 字符串（含 @sha256:xxx）
+        digest="${pair#*:}"
 
-IMAGE_TAG="$1"
-REGISTRY_TOKEN="$2"
-REGISTRY="${REGISTRY:-git.xiaoxiajianji.com/xiaoxia/xiaoxia-saas}"
-REGISTRY_USER="${REGISTRY_USER:-xiaoxia}"
+        image_name="${REGISTRY_HOST}/${REGISTRY_REPO}/xiaoxia-saas-${svc}"
 
-ENV_FILE="${ENV_FILE:-/var/lib/xiaoxia-saas-staging/.env}"
-GENERATED_DIR="${GENERATED_DIR:-/var/lib/xiaoxia-saas-staging/generated}"
-LEGACY_ASSETS_DIR="${LEGACY_ASSETS_DIR:-/var/lib/xiaoxia-saas-staging/legacy-assets}"
+        log_info "  回滚 $svc 到: $digest"
 
-echo "=== Rollback to $IMAGE_TAG ==="
+        # 通过 digest 拉取旧镜像
+        if docker pull "$digest" 2>/dev/null; then
+            # 重新打 :staging tag
+            docker tag "$digest" "${image_name}:staging"
+            docker push "${image_name}:staging"
+            log_info "  ✅ $svc 已回滚"
+        else
+            log_error "  ❌ $svc 回滚失败：无法拉取旧镜像"
+            rollback_ok=false
+        fi
+    done
 
-# 登录 Registry
-if [ -n "$REGISTRY_TOKEN" ]; then
-    REGISTRY_HOST=$(echo "$REGISTRY" | cut -d/ -f1)
-    printf %s "$REGISTRY_TOKEN" | docker login "$REGISTRY_HOST" -u "$REGISTRY_USER" --password-stdin 2>/dev/null || true
-fi
-
-# Pull 旧版本镜像
-LOCAL_API="xiaoxia-saas-api:${IMAGE_TAG}"
-LOCAL_WORKER="xiaoxia-saas-worker:${IMAGE_TAG}"
-LOCAL_WEB="xiaoxia-saas-web:${IMAGE_TAG}"
-
-docker pull "${REGISTRY}/xiaoxia-saas-api:${IMAGE_TAG}"
-docker pull "${REGISTRY}/xiaoxia-saas-worker:${IMAGE_TAG}"
-docker pull "${REGISTRY}/xiaoxia-saas-web:${IMAGE_TAG}"
-
-docker tag "${REGISTRY}/xiaoxia-saas-api:${IMAGE_TAG}" "$LOCAL_API"
-docker tag "${REGISTRY}/xiaoxia-saas-worker:${IMAGE_TAG}" "$LOCAL_WORKER"
-docker tag "${REGISTRY}/xiaoxia-saas-web:${IMAGE_TAG}" "$LOCAL_WEB"
-
-echo "Rollback images pulled."
-
-# 停止当前容器（回滚不跑 migration，避免数据问题）
-docker rm -f xiaoxia-api-staging 2>/dev/null || true
-docker rm -f xiaoxia-worker-staging 2>/dev/null || true
-docker rm -f xiaoxia-web-staging 2>/dev/null || true
-
-LOG_OPTS="--log-driver json-file --log-opt max-size=50m --log-opt max-file=3"
-
-# 启动 API（回滚不跑 migration）
-echo "Starting API (rollback)..."
-docker run -d \
-  --name xiaoxia-api-staging \
-  --env-file "$ENV_FILE" \
-  --network xiaoxia-net-staging \
-  -p 127.0.0.1:8000:8000 \
-  -e APP_ENV=staging \
-  -e APP_VERSION="$IMAGE_TAG" \
-  -e GENERATED_FILES_DIR=/app/generated \
-  -e GENERATED_FILES_URL_PREFIX=/generated-files \
-  -e PUBLIC_API_BASE_URL=https://staging-api.xiaoxiajianji.com \
-  -v "$GENERATED_DIR:/app/generated" \
-  --restart unless-stopped \
-  --health-cmd "python -c \"import urllib.request; urllib.request.urlopen('http://localhost:8000/health', timeout=5)\"" \
-  --health-interval 30s \
-  --health-timeout 10s \
-  --health-retries 3 \
-  --health-start-period 40s \
-  $LOG_OPTS \
-  "$LOCAL_API"
-
-# 启动 Worker
-echo "Starting Worker (rollback)..."
-docker run -d \
-  --name xiaoxia-worker-staging \
-  --env-file "$ENV_FILE" \
-  --network xiaoxia-net-staging \
-  -e APP_ENV=staging \
-  -e APP_VERSION="$IMAGE_TAG" \
-  -e WORKER_CONCURRENCY=1 \
-  -e WORKER_MAX_TASKS_PER_CHILD=100 \
-  -e GENERATED_FILES_DIR=/app/generated \
-  -e GENERATED_FILES_URL_PREFIX=/generated-files \
-  -e PUBLIC_API_BASE_URL=https://staging-api.xiaoxiajianji.com \
-  -v "$GENERATED_DIR:/app/generated" \
-  --restart unless-stopped \
-  --health-cmd "sh -c \"grep -q celery /proc/1/cmdline || exit 1\"" \
-  --health-interval 30s \
-  --health-timeout 10s \
-  --health-retries 3 \
-  --health-start-period 30s \
-  $LOG_OPTS \
-  "$LOCAL_WORKER"
-
-# 启动 Web
-LEGACY_VOLUME=""
-if [ -d "$LEGACY_ASSETS_DIR" ] && [ "$(ls -A "$LEGACY_ASSETS_DIR" 2>/dev/null)" ]; then
-  LEGACY_VOLUME="-v ${LEGACY_ASSETS_DIR}:/usr/share/nginx/html/assets-legacy/assets:ro"
-fi
-
-echo "Starting Web (rollback)..."
-docker run -d \
-  --name xiaoxia-web-staging \
-  --network xiaoxia-net-staging \
-  -p 127.0.0.1:3001:80 \
-  --restart unless-stopped \
-  $LEGACY_VOLUME \
-  --health-cmd "wget --spider -q http://127.0.0.1:80" \
-  --health-interval 30s \
-  --health-timeout 5s \
-  --health-retries 3 \
-  $LOG_OPTS \
-  "$LOCAL_WEB"
-
-# 等待 API 健康
-echo "Waiting for API (rollback)..."
-i=0
-while [ "$i" -lt 40 ]; do
-  if curl -sf --max-time 5 http://127.0.0.1:8000/health >/dev/null 2>&1; then
-    echo "API healthy (rollback)."
-    break
-  fi
-  i=$((i + 1))
-  sleep 3
-done
-
-# 等待 Web 健康
-echo "Waiting for Web (rollback)..."
-i=0
-while [ "$i" -lt 15 ]; do
-  if curl -sf --max-time 5 http://127.0.0.1:3001/ >/dev/null 2>&1; then
-    echo "Web healthy (rollback)."
-    break
-  fi
-  i=$((i + 1))
-  sleep 2
-done
-
-echo "=== Rollback complete: $IMAGE_TAG ==="
-docker ps --format "table {{.Names}}\t{{.Status}}\t{{.Image}}" | grep staging
-ROLLBACK_EOF
-)
-
-    # 将脚本 base64 编码后通过 SSH 执行
-    local script_b64
-    script_b64=$(echo "$rollback_script" | base64 -w 0)
-
-    log_info "在 staging 服务器上执行回滚脚本..."
-    if run_ssh "echo '$script_b64' | base64 -d | sh -s -- '$rollback_tag' '${REGISTRY_TOKEN:-}'" 2>&1; then
-        log_info "✅ 回滚命令执行完成"
+    if [ "$rollback_ok" = true ]; then
+        log_info "所有服务回滚命令执行完成，Watchtower 将在 ~60s 内检测到并更新"
         return 0
     else
-        log_error "❌ 回滚命令执行失败"
+        log_error "部分服务回滚失败"
         return 1
     fi
 }
 
 # ===========================================
-# 4. 发送通知
+# 5. 发送通知
 # ===========================================
 send_notification() {
     local status="$1"   # success / failure / rollback
@@ -394,12 +250,12 @@ send_notification() {
         return 0
     fi
 
+    # 调用通知脚本
     if [ -f "$SCRIPT_DIR/deploy_notify.py" ]; then
         python3 "$SCRIPT_DIR/deploy_notify.py" \
             --status "$status" \
             --detail "$detail" \
             --webhook "$webhook" \
-            --env staging \
             2>/dev/null || log_warn "通知发送失败（非致命）"
     else
         log_warn "找不到 deploy_notify.py，跳过通知"
@@ -411,42 +267,41 @@ send_notification() {
 # ===========================================
 main() {
     echo ""
-    echo "==========================================="
-    echo "  CI Staging 健康检查 + 自动回滚（SSH模式）"
-    echo "==========================================="
+    echo "=========================================="
+    echo "  CI Staging 健康检查 + 自动回滚"
+    echo "=========================================="
     echo ""
 
     local deploy_status="success"
     local deploy_detail=""
 
-    # 1. 设置 SSH
-    if ! setup_ssh; then
-        log_error "SSH 配置失败，无法执行回滚"
-    fi
+    # 1. 记录部署前状态（回滚目标）
+    save_rollback_target
 
-    # 2. 记录部署前状态（回滚目标）
-    save_rollback_target || true
+    # 2. 等待 Watchtower 更新
+    wait_for_watchtower
 
-    # 3. 健康检查（公网视角）
+    # 3. 健康检查
     if ! health_check; then
         log_error "健康检查失败"
         deploy_status="failure"
-        deploy_detail="公网健康检查超时，部署后服务未正常响应"
+        deploy_detail="健康检查超时，部署后服务未正常启动"
 
         # 自动回滚
         if [ "${SKIP_ROLLBACK:-false}" != "true" ]; then
             log_warn "开始自动回滚..."
             if do_rollback; then
                 deploy_status="rollback"
-                deploy_detail="健康检查失败，已自动回滚到上一版本 (${ROLLBACK_TAG:-未知})"
+                deploy_detail="健康检查失败，已自动回滚到上一版本"
 
-                # 回滚后再检查一下公网状态
-                log_info "回滚完成，重新检查公网健康状态..."
+                # 回滚后再检查一下回滚后的状态
+                log_info "等待 Watchtower 应用回滚（90s）..."
+                sleep 90
                 if health_check; then
                     log_info "✅ 回滚后服务已恢复"
                     deploy_detail="${deploy_detail}，回滚后服务已恢复"
                 else
-                    log_error "⚠️  回滚后健康检查仍未通过，请手动排查"
+                    log_error "⚠️ 回滚后健康检查仍未通过，请手动排查"
                     deploy_detail="${deploy_detail}，但回滚后仍未恢复，请紧急排查"
                 fi
             else
@@ -466,7 +321,7 @@ main() {
     log_info "  ✅ Staging 部署成功！"
     log_info "=================================="
 
-    deploy_detail="部署成功，所有健康检查通过 (${GITHUB_SHA:-未知版本})"
+    deploy_detail="部署成功，所有健康检查通过"
     send_notification "success" "$deploy_detail"
 }
 

scripts/deploy_notify.py

@@ -1,17 +1,13 @@
 #!/usr/bin/env python3
 """
-部署通知脚本（支持 Staging / Production）
+Staging 部署通知脚本
 
-与 CI 通知（ci_notify_success.py / ci_notify_failure.py）卡片格式对齐。
 发送部署结果通知到飞书 webhook（卡片格式）。
-
 支持三种状态：success / failure / rollback
-支持两种环境：staging / production
 
 用法:
-    python3 deploy_notify.py --status success --detail "部署成功" --env staging
-    python3 deploy_notify.py --status rollback --detail "健康检查失败，已回滚" --env production
-    python3 deploy_notify.py --status failure --detail "部署过程出错" --env production --failed-step "构建镜像"
+    python3 deploy_notify.py --status success --detail "部署成功" --webhook <url>
+    python3 deploy_notify.py --status rollback --detail "健康检查失败，已回滚"
 """
 
 import argparse
@@ -23,139 +19,73 @@ import urllib.request
 STATUS_CONFIG = {
     "success": {
         "emoji": "✅",
-        "title_suffix": "部署成功",
+        "title": "Staging 部署成功",
         "color": "green",
-        "button_text": "查看构建详情",
-        "button_type": "primary",
     },
     "failure": {
         "emoji": "❌",
-        "title_suffix": "部署失败",
+        "title": "Staging 部署失败",
         "color": "red",
-        "button_text": "查看失败日志",
-        "button_type": "danger",
     },
     "rollback": {
         "emoji": "↩️",
-        "title_suffix": "部署已回滚",
+        "title": "Staging 部署已回滚",
         "color": "yellow",
-        "button_text": "查看构建详情",
-        "button_type": "primary",
-    },
-}
-
-ENV_CONFIG = {
-    "staging": {
-        "label": "Staging",
-        "url_web": "https://staging.xiaoxiajianji.com",
-        "url_api": "https://staging-api.xiaoxiajianji.com",
-    },
-    "production": {
-        "label": "Production",
-        "url_web": "https://saas.xiaoxiajianji.com",
-        "url_api": "https://api.xiaoxiajianji.com",
     },
 }
 
 
-def build_card(
-    status: str,
-    detail: str,
-    env: str = "staging",
-    duration: str = "",
-    failed_step: str = "",
-    pr_url: str = "",
-) -> dict:
-    """构建飞书卡片消息（与 ci_notify_*.py 风格一致）。"""
+def build_card(status: str, detail: str) -> dict:
+    """构建飞书卡片消息。"""
     cfg = STATUS_CONFIG.get(status, STATUS_CONFIG["failure"])
-    env_cfg = ENV_CONFIG.get(env, ENV_CONFIG["staging"])
-    title = f"{cfg['emoji']} {env_cfg['label']} {cfg['title_suffix']}"
 
     commit = os.environ.get("GITHUB_SHA", "unknown")[:8]
-    ref = os.environ.get("GITHUB_REF_NAME", "unknown")
+    branch = os.environ.get("GITHUB_REF_NAME", "unknown")
     actor = os.environ.get("GITHUB_ACTOR", "system")
     run_id = os.environ.get("GITHUB_RUN_ID", "-")
     repo = os.environ.get("GITHUB_REPOSITORY", "xiaoxia/xiaoxia-saas")
 
-    # 版本信息：tag 部署显示 tag，分支部署显示分支
-    if ref.startswith("v"):
-        version_info = f"版本 {ref}"
-    else:
-        version_info = ref
-
-    # 构建内容行（与 ci_notify_*.py 风格一致：**标签**: 值）
-    lines = []
-
-    # 详情行（部署特有）
-    if detail:
-        lines.append(f"**详情**: {detail}")
-
-    lines.append(f"**环境**: {env_cfg['label']}")
-    lines.append(f"**版本**: {version_info}")
-
-    # 失败阶段（失败/回滚时显示）
-    if failed_step and status in ("failure", "rollback"):
-        lines.append(f"**失败阶段**: {failed_step}")
-
-    # 耗时（可选）
-    if duration:
-        lines.append(f"**耗时**: {duration}")
-
-    lines.append(f"**提交**: {commit}")
-    lines.append(f"**提交者**: {actor}")
-    lines.append(f"**Run ID**: {run_id}")
-
     elements = [
         {
             "tag": "div",
             "text": {
                 "tag": "lark_md",
-                "content": "\n".join(lines),
+                "content": (
+                    f"**状态**: {cfg['emoji']} {cfg['title']}\n"
+                    f"**详情**: {detail}\n"
+                    f"**分支**: {branch}\n"
+                    f"**提交**: {commit}\n"
+                    f"**提交者**: {actor}\n"
+                ),
             },
         },
     ]
 
-    # 查看详情按钮
-    run_url = f"https://git.xiaoxiajianji.com/{repo}/actions/runs/{run_id}"
+    # 如果有 run_id，加一个查看详情按钮
     if run_id and run_id != "-":
+        run_url = f"https://git.xiaoxiajianji.com/{repo}/actions/runs/{run_id}"
         elements.append(
             {
                 "tag": "action",
                 "actions": [
                     {
                         "tag": "button",
-                        "text": {"tag": "plain_text", "content": cfg["button_text"]},
+                        "text": {"tag": "plain_text", "content": "查看构建详情"},
                         "url": run_url,
-                        "type": cfg["button_type"],
+                        "type": "primary",
                     }
                 ],
             }
         )
 
-    # PR 链接（可选）
-    if pr_url:
-        elements.append(
-            {
-                "tag": "action",
-                "actions": [
-                    {
-                        "tag": "button",
-                        "text": {"tag": "plain_text", "content": "查看 PR"},
-                        "url": pr_url,
-                        "type": "default",
-                    }
-                ],
-            }
-        )
-
-    # 访问地址（部署特有）
+    # 加 Staging 访问链接
     elements.append(
         {
             "tag": "note",
             "elements": [
                 {
                     "tag": "plain_text",
-                    "content": f"Web: {env_cfg['url_web']}  |  API: {env_cfg['url_api']}",
+                    "content": "Staging: https://staging.xiaoxiajianji.com",
                 }
             ],
         }
@@ -167,7 +97,7 @@ def build_card(
             "header": {
                 "title": {
                     "tag": "plain_text",
-                    "content": title,
+                    "content": f"{cfg['emoji']} {cfg['title']}",
                 },
                 "status": cfg["color"],
             },
@@ -176,17 +106,9 @@ def build_card(
     }
 
 
-def send_notification(
-    webhook: str,
-    status: str,
-    detail: str,
-    env: str = "staging",
-    duration: str = "",
-    failed_step: str = "",
-    pr_url: str = "",
-) -> bool:
+def send_notification(webhook: str, status: str, detail: str) -> bool:
     """发送通知到 webhook。"""
-    payload = build_card(status, detail, env, duration, failed_step, pr_url)
+    payload = build_card(status, detail)
     data = json.dumps(payload).encode("utf-8")
 
     req = urllib.request.Request(
@@ -199,7 +121,7 @@ def send_notification(
     try:
         with urllib.request.urlopen(req, timeout=10) as resp:
             resp.read()
-        print(f"通知已发送: {env} {status}")
+        print(f"通知已发送: {status}")
         return True
     except Exception as e:
         print(f"通知发送失败: {e}", file=sys.stderr)
@@ -207,7 +129,7 @@ def send_notification(
 
 
 def main():
-    parser = argparse.ArgumentParser(description="部署通知脚本")
+    parser = argparse.ArgumentParser(description="Staging 部署通知")
     parser.add_argument(
         "--status",
         required=True,
@@ -215,15 +137,6 @@ def main():
         help="部署状态",
     )
     parser.add_argument("--detail", default="", help="详情描述")
-    parser.add_argument(
-        "--env",
-        default="staging",
-        choices=["staging", "production"],
-        help="部署环境 (默认 staging)",
-    )
-    parser.add_argument("--duration", default="", help="部署耗时")
-    parser.add_argument("--failed-step", default="", help="失败阶段")
-    parser.add_argument("--pr-url", default="", help="PR 链接")
     parser.add_argument(
         "--webhook",
         default=os.environ.get("CI_NOTIFY_WEBHOOK", ""),
@@ -233,19 +146,10 @@ def main():
     args = parser.parse_args()
 
     if not args.webhook:
-        print("未配置 CI_NOTIFY_WEBHOOK，跳过通知")
-        print("如需启用，请在仓库 Settings -> Secrets and variables -> Actions 中添加 CI_NOTIFY_WEBHOOK")
+        print("未配置 webhook URL，跳过通知")
         return 0
 
-    send_notification(
-        webhook=args.webhook,
-        status=args.status,
-        detail=args.detail,
-        env=args.env,
-        duration=args.duration,
-        failed_step=args.failed_step,
-        pr_url=args.pr_url,
-    )
+    send_notification(args.webhook, args.status, args.detail)
     return 0
 
 

scripts/init_phase6_tasks.py

@@ -378,17 +378,17 @@ def init_phase6_tasks():
     conn.commit()
     conn.close()
 
-    print("\n[SUCCESS] Phase 6 任务初始化完成！")
+    print(f"\n[SUCCESS] Phase 6 任务初始化完成！")
     print(f"📊 总计 {len(PHASE6_TASKS)} 个任务")
     print(f"✅ 成功创建 {created_count} 个任务")
-    print("\n任务分布：")
-    print("  Week 1-2: 基础搭建 - 7 个任务")
-    print("  Week 3: 认证页面 - 5 个任务")
-    print("  Week 4: 工作空间管理 - 6 个任务")
-    print("  Week 5: 订阅管理 - 5 个任务")
-    print("  Week 6: Admin 后台 - 5 个任务")
-    print("  Week 7: 个人中心 - 4 个任务")
-    print("  Week 8: 测试和优化 - 8 个任务")
+    print(f"\n任务分布：")
+    print(f"  Week 1-2: 基础搭建 - 7 个任务")
+    print(f"  Week 3: 认证页面 - 5 个任务")
+    print(f"  Week 4: 工作空间管理 - 6 个任务")
+    print(f"  Week 5: 订阅管理 - 5 个任务")
+    print(f"  Week 6: Admin 后台 - 5 个任务")
+    print(f"  Week 7: 个人中心 - 4 个任务")
+    print(f"  Week 8: 测试和优化 - 8 个任务")
     print(f"\n预计总工时：{sum(t['estimated_hours'] for t in PHASE6_TASKS)} 小时")
 
 

scripts/init_tracker_data.py

@@ -294,8 +294,8 @@ def main():
     print("\n" + "=" * 60)
     print("[OK] 数据初始化完成！")
     print("=" * 60)
-    print("\n访问推进器: http://47.98.113.167:8088/projects")
-    print("访问 API 文档: http://47.98.113.167:8089/docs\n")
+    print(f"\n访问推进器: http://47.98.113.167:8088/projects")
+    print(f"访问 API 文档: http://47.98.113.167:8089/docs\n")
 
 
 if __name__ == "__main__":

tests/unit/test_config_schemas_and_ai_endpoints.py

@@ -173,9 +173,7 @@ class TestConfigSchemas:
         cfg = BGMConfig(volume=0.5)
         assert cfg.volume == 0.5
 
-        from pydantic import ValidationError
-
-        with pytest.raises(ValidationError):
+        with pytest.raises(Exception):
             BGMConfig(volume=1.5)  # > 1.0 应该校验失败
 
     def test_edit_plan_config_schema_full(self):

tests/unit/test_url_security.py

@@ -233,7 +233,7 @@ class TestSafeDownload(unittest.TestCase):
         with unittest.mock.patch("urllib.request.build_opener") as mock_opener:
             mock_resp = unittest.mock.MagicMock()
             mock_resp.headers = {"Content-Type": "audio/mpeg"}
-            mock_resp.read.side_effect = [b"ID3audio_data", b""]
+            mock_resp.read.side_effect = [b"audio_data", b""]
             mock_resp.geturl.return_value = "https://example.com/test.mp3"
             mock_opener.return_value.open.return_value = mock_resp
             size = safe_download_file(
@@ -242,7 +242,7 @@ class TestSafeDownload(unittest.TestCase):
                 purpose="test",
                 allowed_mime_types=ALLOWED_AUDIO_MIME_TYPES,
             )
-            self.assertEqual(size, 13)
+            self.assertEqual(size, 10)
             self.assertTrue(os.path.exists(dest))
 
     def test_safe_download_file_stream_size_limit(self):
@@ -274,7 +274,7 @@ class TestSafeDownload(unittest.TestCase):
 
     def test_safe_download_bytes_returns_content(self):
         """safe_download_bytes 应该返回文件内容."""
-        test_data = b"ID3hello world test audio"
+        test_data = b"hello world test audio"
         with unittest.mock.patch("urllib.request.build_opener") as mock_opener:
             mock_resp = unittest.mock.MagicMock()
             mock_resp.headers = {"Content-Type": "audio/mpeg"}