搜索词周/月报表下载

hjm_spider_code/aba_report-download/aba_report.py

+# -*- coding: utf-8 -*-
+"""
+ABA Report 自动化下载框架
+
+流程：
+1. kill 残留 Chrome → 启动浏览器（复用系统用户数据/登录态）
+2. 访问数据页 → 判断登录状态（未登录会跳转登录页，停止执行并通知）
+3. 选择日期范围、站点等筛选条件 → 点击查询
+4. 点击下载按钮 → 弹窗出现 → 点击弹窗中的确认按钮
+5. 等待状态变为 Open → 点击 Open 按钮 → 打开新标签页
+6. 新标签页中找到最新一条下载记录 → 获取下载链接（而非弹窗保存）
+
+定时策略（使用 APScheduler）：
+- 每周日 23:00 执行周报下载，失败则推迟一天重试，最多重试 MAX_RETRY_DAYS 天
+- 登录失效也算失败，推迟到明天重试（等用户手动登录后自动继续）
+- 后续可轻松添加每月任务，只需加一行 scheduler.add_job(...)
+"""
+import time
+import subprocess
+from datetime import datetime, timedelta, date
+
+import requests
+from loguru import logger
+from DrissionPage import ChromiumPage, ChromiumOptions
+from apscheduler.schedulers.blocking import BlockingScheduler
+from apscheduler.triggers.cron import CronTrigger
+
+from config import (
+    WEEKLY_DAY_OF_WEEK, WEEKLY_HOUR, WEEKLY_MINUTE,
+    MONTHLY_DAY, MONTHLY_HOUR, MONTHLY_MINUTE,
+    MAX_RETRY_DAYS, RETRY_INTERVAL_MINUTES, DATA_URL, SITES, SITE_PICKER_MAP, SITE_COUNTRY_NAME, SITE_DM_COUNTRY
+)
+from secure_db_client import get_remote_engine
+import re
+import os
+from urllib.parse import urlparse, parse_qs, unquote
+
+
+# ======================== Chrome 进程管理 ========================
+
+def is_chrome_running():
+    """
+    检测系统中是否有 chrome.exe 进程在运行
+    作用：在启动自动化浏览器之前检测，避免用户数据目录被占用导致启动失败
+    :return: True=有 Chrome 在跑, False=没有
+    """
+    try:
+        # tasklist: Windows 查看进程列表的命令
+        # /FI "IMAGENAME eq chrome.exe": 只过滤 chrome.exe 进程
+        # /NH: 不显示表头
+        output = subprocess.check_output(
+            'tasklist /FI "IMAGENAME eq chrome.exe" /NH',
+            shell=True, stderr=subprocess.DEVNULL
+        ).decode('gbk', errors='ignore')  # Windows 中文系统用 gbk 编码
+        return 'chrome.exe' in output.lower()
+    except Exception:
+        return False
+
+
+def kill_chrome():
+    """
+    强制关闭所有 Chrome 进程，并确认进程完全退出
+    作用：释放用户数据目录的锁，让自动化可以正常启动浏览器
+    """
+    try:
+        # taskkill: Windows 杀进程的命令
+        # /F: 强制终止  /IM: 按进程名匹配  /T: 同时终止子进程
+        subprocess.call('taskkill /F /IM chrome.exe /T', shell=True,
+                        stdout=subprocess.DEVNULL, stderr=subprocess.DEVNULL)
+        # 等待进程完全退出，反复检测直到确认没有 chrome 进程
+        # Chrome 有时会自动重启（崩溃恢复机制），所以需要循环检测
+        for i in range(10):
+            time.sleep(2)
+            if not is_chrome_running():
+                logger.info("Chrome 进程已完全退出")
+                return
+            logger.debug(f"Chrome 仍在运行，再次尝试关闭（第 {i + 1} 次）")
+            subprocess.call('taskkill /F /IM chrome.exe /T', shell=True,
+                            stdout=subprocess.DEVNULL, stderr=subprocess.DEVNULL)
+        logger.warning("多次尝试后 Chrome 仍未完全退出")
+    except Exception as e:
+        logger.error(f"关闭 Chrome 失败: {e}")
+
+
+def ensure_chrome_closed():
+    """
+    确保 Chrome 已关闭，在运行就 kill
+    作用：自动化执行前的前置检查，保证浏览器能正常启动
+    """
+    if is_chrome_running():
+        logger.info("检测到 Chrome 正在运行，强制关闭")
+        kill_chrome()
+
+
+# ======================== 日期计算 ========================
+
+def get_last_week_range():
+    """
+    从数据库 date_20_to_30 表查询上一周的日期范围（周日~周六为一周）
+    返回: (week_start, week_end, date_info)
+        - week_start: 上周日日期，date 对象
+        - week_end: 上周六日期，date 对象
+        - date_info: year_week 字符串，如 "2026-11"
+    """
+    sql = """
+    SELECT
+        MIN(date) AS week_start,
+        MAX(date) AS week_end,
+        year_week AS date_info
+    FROM date_20_to_30
+    WHERE year_week = (
+        SELECT year_week
+        FROM date_20_to_30
+        WHERE date = (
+            SELECT MAX(date) FROM date_20_to_30 WHERE date < (
+                SELECT MIN(date) FROM date_20_to_30 WHERE year_week = (
+                    SELECT year_week FROM date_20_to_30 WHERE date = CURDATE()
+                )
+            )
+        )
+    )
+    """
+    try:
+        engine = get_remote_engine(site_name='us', db_type='mysql')
+        date_df = engine.read_sql(sql)
+
+        if date_df.empty:
+            logger.error("从数据库查询上周日期范围失败：无结果")
+            return None, None, None
+
+        week_start = date_df['week_start'].iloc[0]
+        week_end = date_df['week_end'].iloc[0]
+        date_info = str(date_df['date_info'].iloc[0])
+
+        # 如果返回的是字符串，转为 date 对象
+        if isinstance(week_start, str):
+            week_start = date.fromisoformat(week_start)
+        if isinstance(week_end, str):
+            week_end = date.fromisoformat(week_end)
+        # 如果是 pandas Timestamp，转为 date 对象
+        if hasattr(week_start, 'date'):
+            week_start = week_start.date()
+        if hasattr(week_end, 'date'):
+            week_end = week_end.date()
+
+        logger.info(f"从数据库获取上周范围: {week_start} ~ {week_end} (date_info: {date_info})")
+        return week_start, week_end, date_info
+
+    except Exception as e:
+        logger.error(f"查询上周日期范围异常: {e}")
+        return None, None, None
+
+
+def get_last_month_range():
+    """
+    从数据库 date_20_to_30 表查询上一个月的日期范围
+    返回: (month_start, month_end, date_info)
+        - month_start: 上月第一天，date 对象
+        - month_end: 上月最后一天，date 对象
+        - date_info: year_month 字符串，如 "2026-02"
+    """
+    sql = """
+    SELECT
+        MIN(date) AS month_start,
+        MAX(date) AS month_end,
+        `year_month` AS date_info
+    FROM date_20_to_30
+    WHERE `year_month` = (
+        SELECT `year_month`
+        FROM date_20_to_30
+        WHERE date = (
+            SELECT MAX(date) FROM date_20_to_30 WHERE date < (
+                SELECT MIN(date) FROM date_20_to_30 WHERE `year_month` = (
+                    SELECT `year_month` FROM date_20_to_30 WHERE date = CURDATE()
+                )
+            )
+        )
+    )
+    """
+    try:
+        engine = get_remote_engine(site_name='us', db_type='mysql')
+        date_df = engine.read_sql(sql)
+
+        if date_df.empty:
+            logger.error("从数据库查询上月日期范围失败：无结果")
+            return None, None, None
+
+        month_start = date_df['month_start'].iloc[0]
+        month_end = date_df['month_end'].iloc[0]
+        date_info = str(date_df['date_info'].iloc[0])
+
+        if isinstance(month_start, str):
+            month_start = date.fromisoformat(month_start)
+        if isinstance(month_end, str):
+            month_end = date.fromisoformat(month_end)
+        if hasattr(month_start, 'date'):
+            month_start = month_start.date()
+        if hasattr(month_end, 'date'):
+            month_end = month_end.date()
+
+        logger.info(f"从数据库获取上月范围: {month_start} ~ {month_end} (date_info: {date_info})")
+        return month_start, month_end, date_info
+
+    except Exception as e:
+        logger.error(f"查询上月日期范围异常: {e}")
+        return None, None, None
+
+
+# ======================== 浏览器 ========================
+
+def create_browser():
+    """
+    启动浏览器，使用系统默认用户数据目录
+    作用：打开一个带有用户书签、登录态、插件的 Chrome，而不是全新空白的浏览器
+    前提：Chrome 必须是关闭状态，否则用户数据目录会被占用导致启动失败
+    :return: ChromiumPage 页面对象，失败返回 None
+    """
+    co = ChromiumOptions()
+
+    # 指定 Chrome 浏览器路径
+    # co.set_browser_path(r"C:\Users\Administrator\AppData\Local\google\Chrome\Application\chrome.exe")
+    # co.set_local_port(9333)
+    # co.set_user_data_path(r"C:\Users\Administrator\AppData\Local\Google\Chrome\User Data")
+    # 使用系统 Chrome 的默认用户数据目录（包含登录态、书签、插件等）
+    co.use_system_user_path()
+
+    # Win7 兼容性参数
+    # co.set_argument('--no-sandbox')
+    # co.set_argument('--disable-gpu')
+    # 启动时窗口最大化
+    co.set_argument('--start-maximized')
+    # 禁用自动翻译，强制英文界面，避免页面元素被翻译后选择器匹配不到
+    # co.set_argument('--disable-features=Translate')
+    # co.set_argument('--lang=en-US')
+    try:
+        page = ChromiumPage(co)
+        logger.info("浏览器启动成功（使用系统用户数据）")
+        return page
+    except Exception as e:
+        logger.error(f"浏览器启动失败: {e}")
+        return None
+
+
+# ======================== 登录状态 ========================
+
+def check_login_status(page):
+    """
+    判断是否已登录（数据页是否正确打开）
+    主要依据：URL 中包含 brand-analytics
+    辅助验证：页面标题包含 Brand Analytics、h1 标签包含 Top search terms
+    :return: True=已登录, False=未登录（被重定向到了登录页）
+    """
+    # 主要判断：URL 是否还在数据页
+    if "brand-analytics" not in page.url:
+        logger.warning(f"URL 不包含 brand-analytics，当前 URL: {page.url}")
+        return False
+
+    # 辅助验证：页面标题
+    if "Brand Analytics" not in (page.title or ""):
+        logger.warning(f"页面标题不包含 Brand Analytics，当前标题: {page.title}")
+        return False
+
+    # # 辅助验证：h1 标签内容
+    # h1 = page.ele('tag:h1', timeout=3)
+    # if not h1 or "Top search terms" not in (h1.text or ""):
+    #     logger.warning("页面 h1 标签不包含 Top search terms")
+    #     return False
+
+    return True
+
+
+# ======================== 语言切换 ========================
+
+def ensure_english_language(page):
+    """
+    确保页面语言为英文，避免中文/其他语言导致元素选择器匹配失败。
+    通过右上角语言图标判断当前语言，如果不是 EN 则悬浮菜单切换到 English。
+    :return: True=已是英文或切换成功, False=切换失败
+    """
+    locale_wrapper = page.ele('css:.locale-icon-wrapper', timeout=5)
+    if not locale_wrapper:
+        logger.warning("未找到语言图标，跳过语言检查")
+        return True
+
+    current_lang = locale_wrapper.text.strip()
+    if current_lang.startswith('EN'):
+        logger.info(f"当前语言已是英文: {current_lang}")
+        return True
+
+    logger.info(f"当前语言为 {current_lang}，需要切换到英文")
+
+    # 悬浮到语言按钮，触发下拉菜单
+    locale_trigger = page.ele('css:div[data-test-tag="nav-locale-button"] .flyoutTriggerWrap', timeout=5)
+    if not locale_trigger:
+        logger.error("未找到语言切换按钮")
+        return False
+
+    locale_trigger.hover()
+    time.sleep(1)
+
+    # 点击 English 选项（data-test-tag 以 locale-list-item-en_ 开头）
+    en_option = page.ele('css:a[data-test-tag^="locale-list-item-en_"]', timeout=5)
+    if not en_option:
+        logger.error("未找到 English 语言选项")
+        return False
+
+    en_option.click()
+    time.sleep(3)
+    page.wait.doc_loaded()
+
+    # 验证切换结果
+    locale_wrapper = page.ele('css:.locale-icon-wrapper', timeout=5)
+    if locale_wrapper and locale_wrapper.text.strip().startswith('EN'):
+        logger.info("语言已成功切换到英文")
+        return True
+    else:
+        logger.error(f"语言切换后验证失败，当前: {locale_wrapper.text.strip() if locale_wrapper else '未知'}")
+        return False
+
+
+# ======================== 步骤1: 打开数据页 ========================
+
+def navigate_to_data_page(page):
+    """
+    打开数据页面（只调用一次）
+    作用：导航到报告下载的页面，如果没有登录状态会被重定向到登录页
+    """
+    page.get(DATA_URL)
+    page.wait.doc_loaded()
+    logger.info(f"已打开数据页: {DATA_URL}")
+
+
+# ======================== 步骤1.5: 选择国家 ========================
+
+def select_country(page, site):
+    """
+    通过 countryPicker 下拉框切换到目标国家
+    :param site: 站点代码，如 "US", "UK", "DE" 等
+    :return: True=成功, False=失败
+    """
+    picker_value = SITE_PICKER_MAP.get(site)
+    if not picker_value:
+        logger.error(f"[{site}] 未在 SITE_PICKER_MAP 中找到对应的国家代码")
+        return False
+
+    country_picker = page.ele('#countryPicker', timeout=10)
+    if not country_picker:
+        logger.error(f"[{site}] 未找到 countryPicker 下拉框")
+        return False
+
+    # 如果当前已是目标国家，跳过
+    if country_picker.attr('value') == picker_value:
+        logger.info(f"[{site}] 已是目标国家: {picker_value}")
+        return True
+
+    # 展开下拉框
+    country_picker.shadow_root.ele('css:.select-header').click()
+    time.sleep(1)
+
+    # 点击目标国家选项（kat-option 在 light DOM 中）
+    target_option = page.ele(f'css:#countryPicker kat-option[value="{picker_value}"]', timeout=5)
+    if not target_option:
+        logger.error(f"[{site}] 未找到国家选项: {picker_value}")
+        page.ele('tag:body').click()
+        time.sleep(1)
+        return False
+
+    target_option.click()
+    time.sleep(3)  # 切换国家后等待页面数据刷新
+    page.wait.doc_loaded()
+
+    # 验证切换结果
+    if country_picker.attr('value') != picker_value:
+        logger.error(f"[{site}] 国家切换失败，当前值: {country_picker.attr('value')}")
+        return False
+
+    logger.info(f"[{site}] 已切换到国家: {picker_value}")
+    return True
+
+
+# ======================== 步骤2: 选择筛选条件并查询 ========================
+
+def select_filters_and_query(page, week_start, week_end):
+    """
+    选择日期范围、站点等筛选条件，然后点击查询
+    下拉框 value 对应的是周六（week_end）的日期，格式 YYYY-MM-DD
+    操作方式：模拟点击展开下拉框 → 点击目标选项（不用 JS 赋值）
+    :return: True=筛选成功, False=筛选失败
+    """
+    target_value = week_end.isoformat()  # 如 "2026-03-14"
+    logger.info(f"设置筛选条件，目标周: {target_value}")
+
+    # 1. 确认 Reporting range 是 weekly，不是则点击切换
+    reporting_range = page.ele('#reporting-range', timeout=10)
+    if reporting_range.attr('value') != 'weekly':
+        # 通过 shadow_root 点击 header 展开下拉框
+        reporting_range.shadow_root.ele('css:.select-header').click()
+        time.sleep(1)
+        # 通过 shadow_root 点击 Weekly 选项（kat-option 在 shadow DOM 内）
+        weekly_option = reporting_range.shadow_root.ele('css:kat-option[value="weekly"]', timeout=5)
+        if not weekly_option:
+            logger.error("未找到 Weekly 选项")
+            return False
+        weekly_option.click()
+        time.sleep(2)  # 切换后 Select week 的选项列表会刷新，多等一下
+        # 验证
+        if reporting_range.attr('value') != 'weekly':
+            logger.error(f"Reporting range 切换失败，当前值: {reporting_range.attr('value')}")
+            return False
+        logger.info("Reporting range 已切换为 weekly")
+    else:
+        logger.info("Reporting range 已是 weekly")
+
+    # 2. 选择目标周（Select week 下拉框）
+    week_dropdown = page.ele('#weekly-week', timeout=10)
+    current_value = week_dropdown.attr('value')
+    if current_value != target_value:
+        # 通过 shadow_root 点击 header 展开下拉框
+        week_dropdown.shadow_root.ele('css:.select-header').click()
+        time.sleep(1)
+        # 通过 shadow_root 查找目标周选项
+        target_option = week_dropdown.shadow_root.ele(f'css:kat-option[value="{target_value}"]', timeout=5)
+        if not target_option:
+            logger.error(f"未找到目标周选项: {target_value}，下拉框中可能没有该周数据")
+            # 点击空白处收起下拉框，避免影响后续操作
+            page.ele('tag:body').click()
+            time.sleep(1)
+            return False
+        target_option.click()
+        time.sleep(1)
+        logger.info(f"Select week 已切换为 {target_value}（原值: {current_value}）")
+    else:
+        logger.info(f"Select week 已是目标周: {target_value}")
+
+    # 3. 二次验证：确认下拉框的 value 确实变了
+    final_value = page.ele('#weekly-week').attr('value')
+    if final_value != target_value:
+        logger.error(f"选择验证失败！期望: {target_value}，实际: {final_value}")
+        return False
+
+    # 4. 点击 Apply 按钮
+    apply_btn = page.ele('css:kat-button[data-test-id="RequiredFilterApplyButton"][label="Apply"]', timeout=10)
+    apply_btn.click()
+    logger.info("已点击 Apply 按钮")
+
+    # 5. 等待页面加载完成
+    page.wait.doc_loaded()
+    time.sleep(3)  # 额外等待数据渲染
+    logger.info("筛选条件已应用，页面加载完成")
+    return True
+
+
+# ======================== 步骤2-B: 月报筛选条件 ========================
+
+def select_filters_and_query_monthly(page, month_end):
+    """
+    月报筛选：Reporting range 切为 monthly → 选年份 → 选月份 → 点 Apply
+    :param month_end: 上月最后一天，date 对象（如 2026-02-28）
+    :return: True=筛选成功, False=筛选失败
+    """
+    target_year = str(month_end.year)  # 如 "2026"
+    target_value = month_end.isoformat()  # 如 "2026-02-28"
+    month_id = f"{target_year}-month"  # 如 "2026-month"
+    logger.info(f"设置月报筛选条件，目标: {target_year} / {target_value}")
+
+    # 1. 切换 Reporting range 为 monthly
+    reporting_range = page.ele('#reporting-range', timeout=10)
+    if reporting_range.attr('value') != 'monthly':
+        reporting_range.shadow_root.ele('css:.select-header').click()
+        time.sleep(1)
+        monthly_option = reporting_range.shadow_root.ele('css:kat-option[value="monthly"]', timeout=5)
+        if not monthly_option:
+            logger.error("未找到 Monthly 选项")
+            return False
+        monthly_option.click()
+        time.sleep(2)
+        if reporting_range.attr('value') != 'monthly':
+            logger.error(f"Reporting range 切换失败，当前值: {reporting_range.attr('value')}")
+            return False
+        logger.info("Reporting range 已切换为 monthly")
+    else:
+        logger.info("Reporting range 已是 monthly")
+
+    # 2. 选择年份（Select year）
+    year_dropdown = page.ele('#monthly-year', timeout=10)
+    if year_dropdown.attr('value') != target_year:
+        year_dropdown.shadow_root.ele('css:.select-header').click()
+        time.sleep(1)
+        year_option = year_dropdown.shadow_root.ele(f'css:kat-option[value="{target_year}"]', timeout=5)
+        if not year_option:
+            logger.error(f"未找到年份选项: {target_year}")
+            page.ele('tag:body').click()
+            time.sleep(1)
+            return False
+        year_option.click()
+        time.sleep(2)  # 等待 Select month 下拉框加载
+        logger.info(f"Select year 已切换为 {target_year}")
+    else:
+        logger.info(f"Select year 已是 {target_year}")
+
+    # 3. 选择月份（Select month，ID 为 #{年份}-month）
+    month_dropdown = page.ele(f'#{month_id}', timeout=10)
+    if not month_dropdown:
+        logger.error(f"未找到 Select month 下拉框: #{month_id}")
+        return False
+
+    if month_dropdown.attr('value') != target_value:
+        month_dropdown.shadow_root.ele('css:.select-header').click()
+        time.sleep(1)
+        month_option = month_dropdown.shadow_root.ele(f'css:kat-option[value="{target_value}"]', timeout=5)
+        if not month_option:
+            logger.error(f"未找到月份选项: {target_value}")
+            page.ele('tag:body').click()
+            time.sleep(1)
+            return False
+        month_option.click()
+        time.sleep(1)
+        logger.info(f"Select month 已切换为 {target_value}")
+    else:
+        logger.info(f"Select month 已是 {target_value}")
+
+    # 4. 二次验证
+    final_value = page.ele(f'#{month_id}').attr('value')
+    if final_value != target_value:
+        logger.error(f"月份选择验证失败！期望: {target_value}，实际: {final_value}")
+        return False
+
+    # 5. 点击 Apply
+    apply_btn = page.ele('css:kat-button[data-test-id="RequiredFilterApplyButton"][label="Apply"]', timeout=10)
+    apply_btn.click()
+    logger.info("已点击 Apply 按钮")
+
+    # 6. 等待页面加载完成
+    page.wait.doc_loaded()
+    time.sleep(3)
+    logger.info("月报筛选条件已应用，页面加载完成")
+    return True
+
+
+# ======================== 步骤3: 批量提交下载请求 ========================
+
+def submit_download_request(page, site, date_type, week_start, week_end, wait_timeout=60):
+    """
+    提交单站点的下载请求（不打开 Download Manager）：
+    1. 切换国家
+    2. 选择筛选条件
+    3. 点 Generate Download → 弹窗中再点 Generate Download
+    4. 等待按钮变为 Open Download Manager（表示已进入后端队列）
+    5. 关闭弹窗（关键：不跳转 Manager，准备处理下一个站点）
+
+    :param page: 数据页的页面对象
+    :param site: 站点代码,如 "US","UK"
+    :param date_type: 'week' 或 'month'
+    :param week_start: 周期开始日期
+    :param week_end: 周期结束日期
+    :param wait_timeout: 等待按钮变为 Open Download Manager 的超时时间（秒）
+    :return: True=提交成功, False=失败
+    """
+    logger.info(f"[{site}] === 提交下载请求 ===")
+
+    # 1. 切换国家
+    if not select_country(page, site):
+        logger.error(f"[{site}] 国家切换失败")
+        return False
+
+    # 2. 选筛选条件
+    if date_type == 'month':
+        filter_ok = select_filters_and_query_monthly(page, week_end)
+    else:
+        filter_ok = select_filters_and_query(page, week_start, week_end)
+    if not filter_ok:
+        logger.error(f"[{site}] 筛选条件设置失败")
+        return False
+
+    # 等待 Apply 后数据渲染
+    time.sleep(2)
+
+    # 3. 点击页面上的 Generate Download 按钮
+    generate_btn = page.ele('#GenerateDownloadButton', timeout=10)
+    if not generate_btn:
+        logger.error(f"[{site}] 未找到 Generate Download 按钮")
+        return False
+    generate_btn.click()
+    logger.info(f"[{site}] 已点击 Generate Download 按钮")
+    time.sleep(3)
+
+    # 4. 弹窗出现,定位弹窗按钮
+    modal_btn = page.ele('#downloadModalGenerateDownloadButton', timeout=10)
+    if not modal_btn:
+        logger.error(f"[{site}] 未找到弹窗中的 Generate Download 按钮")
+        return False
+
+    # 等弹窗动画结束
+    time.sleep(2)
+    current_label = modal_btn.attr('label')
+    logger.info(f"[{site}] 弹窗按钮 label: {current_label}")
+
+    # 5. 点击弹窗 Generate Download(如果不是已就绪状态)
+    if current_label != 'Open Download Manager':
+        modal_btn.click()
+        logger.info(f"[{site}] 已点击弹窗 Generate Download")
+        time.sleep(2)
+
+        # 6. 等待按钮变为 Open Download Manager(表示已进队列)
+        waited = 0
+        interval = 3
+        submitted = False
+        while waited < wait_timeout:
+            time.sleep(interval)
+            waited += interval
+            btn = page.ele('#downloadModalGenerateDownloadButton', timeout=2)
+            if btn and btn.attr('label') == 'Open Download Manager':
+                logger.info(f"[{site}] 按钮已变为 Open Download Manager,提交成功")
+                submitted = True
+                break
+            logger.debug(f"[{site}] 等待中,已 {waited}/{wait_timeout}s")
+        if not submitted:
+            logger.error(f"[{site}] 等待 Open Download Manager 超时,提交失败")
+            return False
+    else:
+        logger.info(f"[{site}] 弹窗按钮已是 Open Download Manager(可能上次提交残留),视为成功")
+
+    # 7. 关闭弹窗,准备处理下一个站点
+    try:
+        close_btn = page.ele('#downloadModalCloseButton', timeout=3)
+        if close_btn:
+            close_btn.click()
+            logger.debug(f"[{site}] 已关闭弹窗")
+            time.sleep(1)
+    except Exception as e:
+        logger.debug(f"[{site}] 关闭弹窗异常(可忽略): {e}")
+
+    return True
+
+
+def open_download_manager(page):
+    """
+    通过独立路由 /brand-analytics/download-manager 直接打开 Download Manager 标签页
+    跳过"再点 Generate → 弹窗 → 等 label 变 Open Download Manager → 点击跳转"的繁琐流程
+
+    :param page: 数据页的页面对象
+    :return: Download Manager 标签页对象,失败返回 None
+    """
+    # 从 DATA_URL 提取域名,拼接 Download Manager 完整 URL
+    # 例: https://sellercentral.amazon.com/brand-analytics/download-manager
+    parsed = urlparse(DATA_URL)
+    manager_url = f"{parsed.scheme}://{parsed.netloc}/brand-analytics/download-manager"
+
+    try:
+        # 新开标签页访问 Download Manager
+        new_tab = page.new_tab(manager_url)
+        new_tab.wait.doc_loaded()
+        time.sleep(2)  # 额外等待表格 DOM 渲染
+        logger.info(f"[open_manager] 已打开 Download Manager: {manager_url}")
+        return new_tab
+    except Exception as e:
+        logger.error(f"[open_manager] 打开 Download Manager 失败: {e}")
+        return None
+
+
+# ======================== 步骤4: 在 Download Manager 批量等待并获取下载链接 ========================
+
+def _download_from_row(tab, row, timeout=30):
+    """
+    从指定行触发下载并拦截 report/download 请求,提取 preSignedUrl(内部辅助函数)
+
+    :param tab: Download Manager 标签页
+    :param row: 已匹配到的目标行元素
+    :param timeout: 等待网络响应的超时时间(秒)
+    :return: 下载链接 URL 或 None
+    """
+    # 1. 启动网络监听
+    tab.listen.start('report/download')
+
+    # 2. 找下载按钮
+    download_btn = row.ele('css:.css-1oickjg', timeout=5)
+    if not download_btn:
+        logger.error("未找到 Download 按钮元素(.css-1oickjg)")
+        tab.listen.stop()
+        return None
+
+    # 3. 记录当前标签页,点击下载
+    tabs_before = tab.browser.tab_ids
+    download_btn.click()
+    logger.debug("已点击 Download 按钮,等待拦截 report/download 响应...")
+
+    # 4. 等待响应
+    resp = tab.listen.wait(timeout=timeout)
+    tab.listen.stop()
+
+    download_url = None
+    if not resp:
+        logger.error("未拦截到 report/download 请求")
+    else:
+        try:
+            download_url = resp.response.body.get('preSignedUrl')
+            if not download_url:
+                logger.error(f"响应体中未提取到 preSignedUrl: {resp.response.body}")
+        except Exception as e:
+            logger.error(f"提取 preSignedUrl 异常: {e}")
+
+    # 5. 关闭下载触发的新标签页(浏览器有时会因为 download 链接弹一个新 tab)
+    time.sleep(2)
+    tabs_after = tab.browser.tab_ids
+    new_tabs = [t for t in tabs_after if t not in tabs_before]
+    for t in new_tabs:
+        try:
+            tab.browser.get_tab(t).close()
+            logger.debug("已关闭下载触发的新标签页")
+        except Exception:
+            pass
+
+    return download_url
+
+
+def get_download_urls_batch(tab, pending_tasks, max_wait_minutes=20, poll_interval=5):
+    """
+    批量等待已提交的下载任务在 Download Manager 中变为 Ready,并依次取下载链接
+
+    工作原理：
+    - 已提交 N 个站点,前 N 行就是本次新提交的记录(按时间倒序,最新在最上)
+    - 不刷新页面(Manager 页面会自动更新 Action 列状态 Pending → Download)
+    - 每 poll_interval 秒重新读 DOM,扫描前 N 行
+    - 对每个 pending site 用 country + timeframe + dates 在前 N 行找匹配
+    - 找到且 Action 含 'Download' → 触发下载,从 pending 移除
+    - 找到但 Action 还是 Pending → 跳过,下轮再来
+    - 没找到匹配 → 当成"还没出现",下轮再来
+    - pending 空 或 超过 max_wait_minutes → 退出
+
+    :param tab: Download Manager 标签页
+    :param pending_tasks: 待处理任务列表,每项为 dict:
+        {'site': 'US', 'date_type': 'week', 'period_start': date, 'period_end': date}
+    :param max_wait_minutes: 总等待上限(分钟)
+    :param poll_interval: 轮询间隔(秒)
+    :return: dict {site: download_url 或 None}
+    """
+    N = len(pending_tasks)  # 扫描范围(固定不变)
+    results = {task['site']: None for task in pending_tasks}
+    pending = list(pending_tasks)  # 副本,处理过的会被 remove
+
+    max_seconds = max_wait_minutes * 60
+    T0 = time.time()
+
+    logger.info(
+        f"开始批量等待 {N} 个站点 Action 就绪,扫描范围: 前 {N} 行,"
+        f"上限 {max_wait_minutes} 分钟,每 {poll_interval}s 检查一次"
+    )
+
+    # 首次进入循环前,等表格 DOM 稳定 + 新提交的 N 条全部出现
+    time.sleep(8)
+
+    while pending and (time.time() - T0) < max_seconds:
+        # 1. 读取当前所有行(不刷新页面,Manager 自动更新)
+        try:
+            rows = tab.eles('css:div[role="rowgroup"] div[role="row"]', timeout=10)
+        except Exception as e:
+            logger.warning(f"读取行元素异常: {e}")
+            rows = []
+
+        if not rows:
+            logger.warning("未读到任何数据行,等下轮重试")
+        else:
+            # 2. 只扫前 N 行(本次新提交的)
+            target_rows = rows[:N]
+
+            # 3. 遍历每个 pending site,在前 N 行里找属于自己的那一行
+            for task in pending[:]:  # 用副本遍历方便 remove
+                site = task['site']
+                expected_country = SITE_DM_COUNTRY.get(site, site)
+                expected_time_frame = 'Monthly' if task['date_type'] == 'month' else 'Weekly'
+                ps = task['period_start']
+                pe = task['period_end']
+                start_formats = {
+                    ps.strftime('%d/%m/%Y'),
+                    ps.strftime('%m/%d/%Y'),
+                    ps.strftime('%Y-%m-%d'),
+                }
+                end_formats = {
+                    pe.strftime('%d/%m/%Y'),
+                    pe.strftime('%m/%d/%Y'),
+                    pe.strftime('%Y-%m-%d'),
+                }
+
+                matched_row = None
+                matched_action = None
+
+                for row in target_rows:
+                    try:
+                        cells = row.eles('css:div')
+                    except Exception:
+                        continue
+                    if len(cells) < 8:
+                        continue
+                    try:
+                        row_country = cells[1].text.strip()
+                        row_time_frame = cells[2].text.strip()
+                        row_start = cells[3].text.strip()
+                        row_end = cells[4].text.strip()
+                        row_action = cells[7].text.strip()
+                    except Exception:
+                        continue
+
+                    if (row_country == expected_country and
+                            row_time_frame == expected_time_frame and
+                            row_start in start_formats and
+                            row_end in end_formats):
+                        matched_row = row
+                        matched_action = row_action
+                        break  # country 是唯一识别,找到自己的行就停
+
+                if matched_row is None:
+                    logger.debug(f"[{site}] 前 {N} 行无匹配,下轮再扫")
+                    continue
+
+                # 4. 检查 Action 列状态
+                if 'Download' in matched_action:
+                    logger.info(f"[{site}] Action 已就绪 ('{matched_action}'),触发下载")
+                    url = _download_from_row(tab, matched_row)
+                    if url:
+                        results[site] = url
+                        logger.success(f"[{site}] 下载链接获取成功: {url}")
+                    else:
+                        logger.error(f"[{site}] 下载触发失败,放弃该站点(外层会重试)")
+                    pending.remove(task)
+                else:
+                    logger.debug(f"[{site}] 状态: {matched_action},下轮再来")
+
+        # 5. 判断退出
+        if not pending:
+            logger.success(f"所有站点已完成,总耗时 {int(time.time() - T0)}s")
+            break
+
+        # 6. 睡眠 poll_interval,进入下一轮
+        elapsed = int(time.time() - T0)
+        logger.info(
+            f"剩余 pending: {[t['site'] for t in pending]},已等待 {elapsed}s / {max_seconds}s"
+        )
+        time.sleep(poll_interval)
+
+        # 兜底逻辑:如果实测发现 Manager 页面不会自动更新 Action 列状态,
+        # 解开下面三行注释,改为每轮刷新页面.
+        # try:
+        #     tab.refresh()
+        #     tab.wait.doc_loaded()
+        #     time.sleep(2)
+        # except Exception as e:
+        #     logger.warning(f"刷新失败: {e}")
+
+    # 循环结束,标记未完成的任务
+    if pending:
+        for task in pending:
+            logger.error(f"[{task['site']}] 等待 {max_wait_minutes} 分钟超时,未获取下载链接")
+
+    return results
+
+
+# ======================== 消息通知 ========================
+
+def send_notification(event, **kwargs):
+    """
+    通用消息通知函数，根据事件类型组装消息
+
+    :param event: 事件类型
+    :param kwargs: 事件相关参数
+
+    事件类型：
+    - login_failed: 登录失败（需要人工介入）
+    - task_summary: 最终任务汇总（所有重试结束后只发一次）
+    """
+    period_type = "周报" if kwargs.get('date_type') == 'week' else "月报"
+    date_info = kwargs.get('date_info', '')
+    week_start = kwargs.get('week_start', '')
+    week_end = kwargs.get('week_end', '')
+    now = datetime.now().strftime('%Y-%m-%d %H:%M:%S')
+
+    if event == 'login_failed':
+        sites = kwargs.get('sites', [])
+        msg = (
+            f"【ABA {period_type} - 登录失败】\n"
+            f"站点: {', '.join(sites) if sites else '未知'}\n"
+            f"周期: {date_info} ({week_start} ~ {week_end})\n"
+            f"原因: 未检测到登录状态，请手动登录后重试\n"
+            f"时间: {now}"
+        )
+    elif event == 'task_summary':
+        success_sites = kwargs.get('success_sites', [])
+        failed_sites = kwargs.get('failed_sites', [])
+        attempt = kwargs.get('attempt', 0)
+        status = "全部成功" if not failed_sites else "部分失败"
+        retry_info = f"（第 {attempt + 1} 次执行）" if attempt > 0 else ""
+        msg = (
+            f"【ABA {period_type} - 任务汇总】{retry_info}\n"
+            f"状态: {status}\n"
+            f"周期: {date_info} ({week_start} ~ {week_end})\n"
+            f"成功: {', '.join(success_sites) if success_sites else '无'}\n"
+            f"失败: {', '.join(failed_sites) if failed_sites else '无'}\n"
+            f"时间: {now}"
+        )
+    else:
+        msg = f"【ABA 通知】未知事件: {event}，参数: {kwargs}"
+
+    logger.info(f"通知消息:\n{msg}")
+    user = ['hejiangming', 'fangxingjun']
+    title = f"ABA {period_type} 周期 {date_info} 通知"
+
+    send_wx_msg(user, title, content=msg)
+
+
+def send_wx_msg(users: list, title: str, content: str, msgtype: str = "textcard"):
+    """
+    :param users: 填写需要推送的微信用户名list
+    :param title: 推送的标题（如果msgtype采用markdown形式，则不附带标题）
+    :param content: 推送的主体内容
+    :param msgtype: 推送的消息类型（textcard：默认卡片类型；markdown：markdaown结构）
+    """
+    if users is not None:
+        accounts = ",".join(users)
+        # 排除users_list=[''] 无需发送
+        if bool(accounts):
+            host = "http://120.79.147.190:8080"
+            url = f'{host}/soundasia_selection/dolphinScheduler/sendMessage'
+            data = {
+                'account': accounts,
+                'title': title,
+                'content': content,
+                'msgtype': msgtype
+            }
+            try:
+                requests.post(url=url, data=data, timeout=15)
+            except:
+                pass
+    return True
+
+
+# ======================== 工具函数 ========================
+
+def extract_filename(url):
+    """
+    从预签名 URL 中提取文件名
+    优先从 response-content-disposition 参数中提取 filename
+    """
+    parsed = urlparse(url)
+    params = parse_qs(parsed.query)
+    disposition = params.get("response-content-disposition", [""])[0]
+    match = re.search(r"filename[*]?=['\"]?(?:UTF-8'')?([^;\"']+)", unquote(disposition))
+    if match:
+        return match.group(1)
+    return os.path.basename(parsed.path)
+
+
+# ======================== 保存下载记录到数据库 ========================
+
+def save_download_record(site, date_type, week_start, week_end, date_info, download_url):
+    """
+    将下载记录保存到 MySQL
+
+    :param site: 站点代码，如 "DE", "UK"
+    :param date_type: 日期类型，如 "week", "month"
+    :param week_start: 目标周开始日期
+    :param week_end: 目标周结束日期
+    :param date_info: 周期标识，周报如 "2026-11"，月报如 "2026-02"
+    :param download_url: 预签名下载链接
+    :return: True=保存成功, False=失败
+    """
+    try:
+        engine = get_remote_engine(site_name='us', db_type='mysql')
+
+        site_name = site.lower()
+        country_name = SITE_COUNTRY_NAME.get(site, site)
+        date_start = week_start.isoformat()
+        date_end = week_end.isoformat()
+        file_name = extract_filename(download_url)
+        hex_url = download_url.encode().hex()
+        state = 1
+
+        final_sql = (
+            f"INSERT INTO aba_report_download "
+            f"(site_name, date_type, date_info, country_name, date_start, date_end, "
+            f"file_name, download_url, state, created_time, updated_time) VALUES "
+            f"('{site_name}', '{date_type}', '{date_info}', '{country_name}', "
+            f"'{date_start}', '{date_end}', '{file_name}', UNHEX('{hex_url}'), "
+            f"{state}, NOW(), NOW()) "
+            f"ON DUPLICATE KEY UPDATE "
+            f"download_url=UNHEX('{hex_url}'), file_name='{file_name}', "
+            f"state={state}, updated_time=NOW()"
+        )
+
+        with engine.begin() as conn:
+            conn.execute(final_sql)
+
+        logger.info(f"[{site}] 下载记录已保存到数据库")
+        return True
+
+    except Exception as e:
+        logger.error(f"[{site}] 保存数据库失败: {e}")
+        return False
+
+
+def get_already_downloaded_sites(sites, date_type, date_info):
+    """
+    查询 aba_report_download 表，返回该周期内已有下载记录的站点列表
+    只判断记录是否存在，不看 state（state 是流转状态: 1未开始/2下载中/3下载完成/4同步完成，会变化）
+    记录存在 = 当时已成功拿到下载链接并写库 = 不需要重复下载
+
+    用途：任务计划程序在 2/3/4 号多次触发月报时，已下载成功的周期直接跳过，避免重复下载
+
+    :param sites: 站点代码列表，如 ["US"] 或 ["UK", "DE"]
+    :param date_type: 日期类型，"week" 或 "month"
+    :param date_info: 周期标识，周报如 "2026-11"，月报如 "2026-02"
+    :return: 已有记录的站点列表（大写），如 ["US"]；查询失败返回 []（照常执行，宁可重复也不漏）
+    """
+    site_names = ','.join(f"'{s.lower()}'" for s in sites)
+    sql = f"""
+    SELECT site_name FROM aba_report_download
+    WHERE date_type = '{date_type}'
+      AND date_info = '{date_info}'
+      AND site_name IN ({site_names})
+    """
+    try:
+        engine = get_remote_engine(site_name='us', db_type='mysql')
+        df = engine.read_sql(sql)
+        return [s.upper() for s in df['site_name'].tolist()] if not df.empty else []
+    except Exception as e:
+        logger.error(f"查询已下载站点异常: {e}")
+        return []
+
+
+# ======================== 推送接口 ========================
+
+def push_to_api(download_infos, date_type, date_info, week_start):
+    """
+    将下载成功的站点信息推送到接口
+
+    :param download_infos: 下载信息列表，每项包含 site, download_url, file_name
+    :param date_type: 日期类型，"week" 或 "month"
+    :param date_info: 周期标识，如 "2026-11"
+    :param week_start: 起始日期
+    """
+    api_url = "https://selection.yswg.com.cn/soundasia_selection/bigfile/importByUrl"
+
+    type_map = {"week": "1", "month": "2"}
+    payload = []
+    for info in download_infos:
+        payload.append({
+            "fileUrl": info['download_url'],
+            "fileName": info['file_name'],
+            "site": info['site'],
+            "type": type_map.get(date_type, "1"),
+            "dateInfo": str(date_info),
+            "startDate": str(week_start),
+        })
+
+    try:
+        import json
+        headers = {"Content-Type": "application/json"}
+        resp = requests.post(api_url, data=json.dumps(payload), headers=headers, timeout=30)
+        logger.info(f"接口推送响应{resp.text}")
+        if resp.status_code == 200:
+            logger.success(f"接口推送成功，共 {len(payload)} 条记录")
+        else:
+            logger.error(f"接口推送失败，状态码: {resp.status_code}，响应: {resp.text}")
+    except Exception as e:
+        logger.error(f"接口推送异常: {e}")
+
+
+# ======================== 主流程 ========================
+
+def run_weekly_download(sites, date_type, week_start, week_end, date_info):
+    """
+    打开数据页(一次) → 检查登录(一次) → 三阶段执行：
+      阶段 1: 遍历 sites 批量提交下载请求(不打开 Manager)
+      阶段 2: 打开 Download Manager,批量等待所有 Action 就绪,依次取链接
+      阶段 3: 保存数据库 + 返回结果
+
+    :param sites: 站点代码列表,如 ["US"] 或 ["UK", "DE", "ES", "FR", "IT", "NL"]
+    :param date_type: 日期类型,"week" 或 "month"
+    :param week_start: 目标周开始日期
+    :param week_end: 目标周结束日期
+    :param date_info: 周期标识,周报如 "2026-11",月报如 "2026-02"
+    :return: (success_sites, failed_sites, success_download_infos) 或 False(浏览器/登录失败)
+    """
+    # 打开数据页,失败则关闭浏览器重试,最多 3 次
+    page = None
+    max_open_retries = 3
+    for open_attempt in range(1, max_open_retries + 1):
+        ensure_chrome_closed()
+        page = create_browser()
+        if not page:
+            logger.error(f"浏览器启动失败(第 {open_attempt}/{max_open_retries} 次)")
+            continue
+
+        navigate_to_data_page(page)
+
+        if check_login_status(page):
+            logger.info(f"数据页访问成功(第 {open_attempt} 次)")
+            break
+
+        logger.warning(f"数据页访问失败(第 {open_attempt}/{max_open_retries} 次),关闭浏览器重试...")
+        try:
+            page.quit()
+        except Exception:
+            pass
+        page = None
+        time.sleep(5)
+
+    if not page:
+        logger.error("多次重试后仍无法访问数据页,判定为登录失效")
+        send_notification('login_failed', date_type=date_type, date_info=date_info, week_start=week_start,
+                          week_end=week_end, sites=sites)
+        return False
+
+    try:
+        # 确保页面语言为英文(只一次)
+        if not ensure_english_language(page):
+            logger.error("语言切换失败,可能导致后续元素定位异常")
+            return False
+
+        # ============ 阶段 1: 批量提交所有站点的下载请求 ============
+        logger.info("=" * 60)
+        logger.info(f"阶段 1: 批量提交 {len(sites)} 个站点的下载请求")
+        logger.info(f"目标周期: {week_start} ~ {week_end} ({date_type})")
+        logger.info("=" * 60)
+
+        submitted_sites = []
+        failed_submit_sites = []
+        for site in sites:
+            try:
+                if submit_download_request(page, site, date_type, week_start, week_end):
+                    submitted_sites.append(site)
+                    logger.success(f"[{site}] 提交成功")
+                else:
+                    failed_submit_sites.append(site)
+                    logger.error(f"[{site}] 提交失败")
+            except Exception as e:
+                logger.error(f"[{site}] 提交异常: {e}")
+                failed_submit_sites.append(site)
+
+        if not submitted_sites:
+            logger.error("所有站点提交均失败,流程终止")
+            return [], list(sites), []
+
+        logger.info(f"提交成功 {len(submitted_sites)} 个: {submitted_sites}")
+        if failed_submit_sites:
+            logger.warning(f"提交失败 {len(failed_submit_sites)} 个: {failed_submit_sites}")
+
+        # ============ 阶段 2: 打开 Download Manager 批量等待 ============
+        logger.info("=" * 60)
+        logger.info(f"阶段 2: 打开 Download Manager,批量等待 {len(submitted_sites)} 个站点 Action 就绪")
+        logger.info("=" * 60)
+
+        manager_tab = open_download_manager(page)
+        if not manager_tab:
+            logger.error("未能打开 Download Manager,所有已提交站点视为失败")
+            return [], list(sites), []
+
+        pending_tasks = [
+            {
+                'site': site,
+                'date_type': date_type,
+                'period_start': week_start,
+                'period_end': week_end,
+            }
+            for site in submitted_sites
+        ]
+
+        results = get_download_urls_batch(
+            manager_tab, pending_tasks, max_wait_minutes=20, poll_interval=5
+        )
+
+        # 关闭 Manager 标签页
+        try:
+            manager_tab.close()
+        except Exception:
+            pass
+
+        # ============ 阶段 3: 处理结果(存库 + 整理返回) ============
+        logger.info("=" * 60)
+        logger.info("阶段 3: 处理下载结果")
+        logger.info("=" * 60)
+
+        success_sites = []
+        failed_sites = list(failed_submit_sites)  # 提交失败的算作失败
+        success_download_infos = []
+
+        for site, url in results.items():
+            if url:
+                success_sites.append(site)
+                save_download_record(site, date_type, week_start, week_end, date_info, url)
+                success_download_infos.append({
+                    'site': site.lower(),
+                    'download_url': url,
+                    'file_name': extract_filename(url),
+                })
+            else:
+                failed_sites.append(site)
+
+        if failed_sites:
+            logger.warning(f"以下站点下载失败: {', '.join(failed_sites)}")
+        else:
+            logger.success(f"所有站点下载完成: {', '.join(sites)}")
+
+        return success_sites, failed_sites, success_download_infos
+
+    finally:
+        try:
+            page.quit()
+            logger.info("浏览器已关闭")
+        except Exception:
+            pass
+
+
+# ======================== 带重试的任务包装 ========================
+
+def weekly_task_with_retry(sites, date_type='week'):
+    """
+    定时任务入口（由 APScheduler 触发，或手动测试调用）
+
+    执行逻辑：
+    1. 首次执行 run_weekly_download()
+    2. 成功 → 结束，等下次触发
+    3. 失败（登录失效/数据未更新等） → 推迟一天再试，最多重试 MAX_RETRY_DAYS 天
+
+    :param sites: 站点代码列表，如 ["US"] 或 ["UK", "DE", "ES", "FR", "IT"]
+    :param date_type: 日期类型，周任务传 "week"，月任务传 "month"
+
+    注意：重试期间会阻塞当前任务，但不会影响调度器中的其他任务
+    因为 APScheduler 默认每个任务在独立线程中执行
+    """
+    weekday_names = ['周一', '周二', '周三', '周四', '周五', '周六', '周日']
+
+    # 首次执行时从数据库获取目标周期，重试时复用同一个周期
+    if date_type == 'month':
+        week_start, week_end, date_info = get_last_month_range()
+    else:
+        week_start, week_end, date_info = get_last_week_range()
+    if not week_start or not week_end:
+        logger.error(f"获取{'上月' if date_type == 'month' else '上周'}日期范围失败，终止执行")
+        return
+
+    # 检查本周期是否已有下载记录，已成功的站点跳过
+    # 场景：任务计划程序 2/3/4 号多次触发月报，3 号成功后 4 号触发时直接退出
+    done_sites = get_already_downloaded_sites(sites, date_type, date_info)
+    remaining_sites = [s for s in sites if s not in done_sites]  # 待下载站点，重试时只包含失败的
+    if not remaining_sites:
+        logger.success(f"周期 {date_info} 所有站点已有下载记录（{', '.join(done_sites)}），本次跳过执行")
+        return
+    if done_sites:
+        logger.info(f"周期 {date_info} 已下载站点: {', '.join(done_sites)}，本次只处理: {', '.join(remaining_sites)}")
+
+    logger.info(f"待下载站点: {', '.join(remaining_sites)}")
+    all_success_sites = []  # 累计所有成功的站点
+    all_download_infos = []  # 累计所有成功站点的下载信息，用于推送接口
+
+    for attempt in range(MAX_RETRY_DAYS + 1):
+        today_name = weekday_names[datetime.now().weekday()]
+        if attempt == 0:
+            logger.info(f"首次执行（当前: {today_name}），站点: {', '.join(remaining_sites)}")
+        else:
+            logger.info(f"第 {attempt} 次重试（当前: {today_name}），站点: {', '.join(remaining_sites)}")
+
+        result = run_weekly_download(remaining_sites, date_type, week_start, week_end, date_info)
+
+        # run_weekly_download 返回 (success_sites, failed_sites, success_download_infos) 或 False（浏览器/登录失败）
+        if result is False:
+            # 登录/浏览器失败，login_failed 通知已在内部发送，这里不重复发汇总
+            failed_sites = list(remaining_sites)
+        else:
+            success_sites, failed_sites, success_download_infos = result
+            all_success_sites.extend(success_sites)
+            all_download_infos.extend(success_download_infos)
+            remaining_sites = failed_sites  # 下次只重试失败的站点
+
+            # 发送任务汇总通知（每次执行都发，不管成功失败）
+            send_notification('task_summary', date_type=date_type, date_info=date_info,
+                              week_start=week_start, week_end=week_end,
+                              success_sites=all_success_sites, failed_sites=failed_sites, attempt=attempt)
+
+            # 全部成功 → 推送接口 → 结束
+            if not failed_sites:
+                logger.success("所有站点数据下载成功")
+                push_to_api(all_download_infos, date_type, date_info, week_start)
+                return
+
+        # 最后一次机会也失败了 → 推送已成功的部分 → 放弃
+        if attempt >= MAX_RETRY_DAYS:
+            logger.error(f"已重试 {MAX_RETRY_DAYS} 次仍失败，放弃本周期")
+            if all_download_infos:
+                push_to_api(all_download_infos, date_type, date_info, week_start)
+            return
+
+        # 还有机会，等待指定间隔后重试
+        sleep_seconds = RETRY_INTERVAL_MINUTES * 60
+        next_retry_time = datetime.now() + timedelta(minutes=RETRY_INTERVAL_MINUTES)
+        logger.warning(
+            f"本次失败站点: {', '.join(remaining_sites)}，"
+            f"将在 {next_retry_time.strftime('%Y-%m-%d %H:%M')} 重试"
+            f"（已尝试 {attempt + 1}/{MAX_RETRY_DAYS + 1} 次）"
+        )
+        time.sleep(sleep_seconds)
+
+
+# ======================== 调度器 ========================
+
+def start(sites):
+    """
+    启动 APScheduler 定时调度器
+
+    :param sites: 站点代码列表，如 ["US"] 或 ["UK", "DE", "ES", "FR", "IT"]
+
+    APScheduler 核心概念：
+    - BlockingScheduler: 阻塞式调度器，启动后主线程就一直等待，适合脚本独占运行
+    - CronTrigger: 类似 Linux crontab 的触发器，可以指定 星期几/几号/几点 执行
+    - add_job(函数, 触发器): 注册一个定时任务
+
+    CronTrigger 常用参数：
+    - day_of_week: 星期几触发，'sun'=周日, 'mon'=周一, '0'=周一, '6'=周日
+    - day: 每月几号触发，1=1号, 'last'=最后一天
+    - hour/minute: 几点几分触发
+    - month: 几月触发，不写就是每月
+
+    添加新任务示例：
+    # 每月1号 09:00 执行月报下载
+    # scheduler.add_job(
+    #     monthly_task_with_retry,           # 要执行的函数
+    #     CronTrigger(day=1, hour=9, minute=0),  # 每月1号 09:00
+    #     id='monthly_report',               # 任务唯一ID，用于管理（暂停/删除等）
+    #     name='月报下载',
+    #     args=[sites]
+    # )
+    """
+    scheduler = BlockingScheduler()
+
+    # --- 每周任务：周日 23:00 执行周报下载 ---
+    scheduler.add_job(
+        weekly_task_with_retry,
+        CronTrigger(day_of_week=WEEKLY_DAY_OF_WEEK, hour=WEEKLY_HOUR, minute=WEEKLY_MINUTE),
+        args=[sites, 'week'],
+        id='weekly_aba_report',
+        name='ABA周报下载',
+        misfire_grace_time=None
+    )
+
+    # --- 每月任务：每月1号 23:00 执行月报下载 ---
+    scheduler.add_job(
+        weekly_task_with_retry,
+        CronTrigger(day=MONTHLY_DAY, hour=MONTHLY_HOUR, minute=MONTHLY_MINUTE),
+        args=[sites, 'month'],
+        id='monthly_aba_report',
+        name='ABA月报下载',
+        misfire_grace_time=None
+    )
+
+    logger.info(f"定时调度器启动，站点: {', '.join(sites)}，已注册任务：")
+    for job in scheduler.get_jobs():
+        logger.info(f"  - {job.name} (ID: {job.id}, 触发器: {job.trigger})")
+
+    # 启动调度器（阻塞主线程，程序会一直运行等待触发）
+    scheduler.start()
+
+
+def manual_download(sites, date_type, start_date_str, end_date_str, date_info):
+    """
+    手动指定日期范围导出，用于补导历史漏掉的周期
+
+    :param sites: 站点代码列表，如 ["US"] 或 ["UK", "DE"]
+    :param date_type: 日期类型，如 "week", "month"
+    :param start_date_str: 开始日期，格式 'YYYY-MM-DD'，比如 '2026-03-01'
+    :param end_date_str: 结束日期，格式 'YYYY-MM-DD'，比如 '2026-03-07'
+    :param date_info: 周期标识，周报如 "2026-11"，月报如 "2026-02"
+
+    用法示例（在 __main__ 里调用）：
+        manual_download(sites, 'week', '2026-03-01', '2026-03-07', '2026-10')
+        manual_download(sites, 'month', '2026-02-01', '2026-02-28', '2026-02')
+    """
+    week_start = date.fromisoformat(start_date_str)
+    week_end = date.fromisoformat(end_date_str)
+    logger.info(
+        f"手动导出模式（{date_type}），站点: {', '.join(sites)}，目标周期: {week_start} ~ {week_end} (date_info: {date_info})")
+    remaining_sites = list(sites)  # 待下载站点，重试时只包含失败的
+    all_success_sites = []  # 累计所有成功的站点
+    all_download_infos = []  # 累计所有成功站点的下载信息，用于推送接口
+
+    result = run_weekly_download(remaining_sites, date_type, week_start, week_end, date_info)
+    if result is False:
+        # 登录/浏览器失败，login_failed 通知已在内部发送
+        failed_sites = list(sites)
+    else:
+        success_sites, failed_sites, success_download_infos = result
+        all_success_sites.extend(success_sites)
+        all_download_infos.extend(success_download_infos)
+
+    # 发送任务汇总通知
+    send_notification('task_summary', date_type=date_type, date_info=date_info,
+                      week_start=week_start, week_end=week_end,
+                      success_sites=all_success_sites, failed_sites=failed_sites, attempt=0)
+
+    # 推送已成功的部分到接口
+    if all_download_infos:
+        push_to_api(all_download_infos, date_type, date_info, week_start)
+
+    if not failed_sites:
+        logger.success("所有站点数据下载成功")
+    else:
+        logger.warning(f"手动导出部分失败，失败站点: {', '.join(failed_sites)}")
+
+
+if __name__ == "__main__":
+    logger.add(
+        "logs/aba_report_{time:YYYY-MM-DD}.log",
+        rotation="1 day",
+        retention="30 days",
+        encoding="utf-8",
+        level="INFO"
+    )
+
+    # 站点列表从 config.py 中读取，部署时改 config.py 即可
+    import sys
+
+    # python aba_report.py week   → 执行周报
+    # python aba_report.py month  → 执行月报
+    # python aba_report.py        → 不带参数，走下面原有逻辑
+    if len(sys.argv) > 1:
+        task_type = sys.argv[1]
+        if task_type == 'week':
+            weekly_task_with_retry(SITES, 'week')
+        elif task_type == 'month':
+            weekly_task_with_retry(SITES, 'month')
+        else:
+            logger.error(f"未知的任务类型: {task_type}，请使用 week 或 month")
+    else:
+        # --- 手动补导历史数据（指定日期范围） ---
+        # manual_download(SITES, 'week', '2026-03-08', '2026-03-14', '2026-11')
+        # manual_download(SITES, 'month', '2026-02-01', '2026-02-28', '2026-02')
+
+        # --- 手动执行当前周期（带重试） ---
+        weekly_task_with_retry(SITES, 'week')
+        # weekly_task_with_retry(SITES, 'month')
+
+        # --- 启动定时调度（每周日自动执行） ---
+        # start(SITES)

hjm_spider_code/aba_report-download/config.py

+# -*- coding: utf-8 -*-
+"""
+ABA Report 配置文件
+部署时只需修改 DATA_URL 和 SITES
+"""
+
+# ======================== 定时任务配置 ========================
+# 每周任务：周日 23:00 执行（day_of_week: 'sun'=周日, 'mon'=周一, 'tue', 'wed', 'thu', 'fri', 'sat'）
+WEEKLY_DAY_OF_WEEK = 'sun'
+WEEKLY_HOUR = 18
+WEEKLY_MINUTE = 0
+
+# 每月任务：每月1号 22:00 执行（比周任务 23:00 早一小时，避免冲突）
+MONTHLY_DAY = 1
+MONTHLY_HOUR = 22
+MONTHLY_MINUTE = 0
+
+# 失败后最多重试次数（周和月通用）
+MAX_RETRY_DAYS = 5
+# 每次重试间隔（分钟）
+RETRY_INTERVAL_MINUTES = 30
+
+# ======================== 部署配置========================
+# 电脑A:
+DATA_URL = "https://sellercentral.amazon.com/brand-analytics/dashboard/top-search-terms"
+SITES = ["US"]
+# 电脑B:
+# DATA_URL = "https://sellercentral.amazon.co.uk/brand-analytics/dashboard/top-search-terms"
+# SITES = ["UK", "DE", "ES", "FR", "IT"]
+
+
+# ======================== 站点映射配置 ========================
+# 站点代码 → countryPicker 下拉框的 value（小写，用于页面切换国家）
+SITE_PICKER_MAP = {
+    "US": "us",
+    "UK": "gb",
+    "DE": "de",
+    "ES": "es",
+    "FR": "fr",
+    "IT": "it",
+    "NL": "nl",
+}
+
+# 站点代码 → 国家英文全称（用于存数据库 country_name 字段）
+SITE_COUNTRY_NAME = {
+    "US": "United States",
+    "UK": "United Kingdom",
+    "DE": "Germany",
+    "ES": "Spain",
+    "FR": "France",
+    "IT": "Italy",
+    "NL": "Netherlands",
+}
+
+# 站点代码 → Download Manager 页面 Country 列显示的国家缩写（大写）
+SITE_DM_COUNTRY = {
+    "US": "US",
+    "UK": "GB",
+    "DE": "DE",
+    "ES": "ES",
+    "FR": "FR",
+    "IT": "IT",
+    "NL": "NL",
+}

hjm_spider_code/aba_report-download/requirements.txt

+loguru
+DrissionPage
+APScheduler

hjm_spider_code/aba_report-download/secure_db_client.py

+import json
+
+import pandas as pd
+import numpy as np
+import orjson, requests, time
+from typing import List
+
+
+# -------- 映射字典 --------
+site_name_db_dict = {
+    "us": "selection",
+    "uk": "selection_uk",
+    "de": "selection_de",
+    "es": "selection_es",
+    "fr": "selection_fr",
+    "it": "selection_it",
+}
+
+db_type_alias_map = {
+    "mysql": "mysql",  # 阿里云mysql
+    "postgresql_14": "postgresql_14",  # pg14爬虫库-内网
+    "postgresql_14_outer": "postgresql_14_outer",  # pg14爬虫库-外网
+    "postgresql_15": "postgresql_15",  # pg15正式库-内网
+    "postgresql_15_outer": "postgresql_15_outer",  # pg15正式库-外网
+    "postgresql_cluster": "postgresql_cluster",  # pg集群-内网
+    "postgresql_cluster_outer": "postgresql_cluster_outer",  # pg集群-外网
+    "doris": "doris",  # doris集群-内网
+}
+
+DEFAULT_SERVERS = [
+    # "http://192.168.200.210:7777",   # 内网
+    # "http://192.168.10.217:7777",   # 内网-h7
+    # "http://61.145.136.61:7777",   # 外网
+    # "http://61.145.136.61:7779",   # 外网
+    "http://61.145.136.61:7780",   # 外网
+]
+# ---------------------------
+
+def df_to_json_records(df: pd.DataFrame) -> list:
+    """保证 DataFrame 可安全序列化为 JSON records（处理 NaN / ±Inf）"""
+    df_clean = df.copy()
+
+    # 1️⃣ 替换 ±Inf -> NaN
+    num_cols = df_clean.select_dtypes(include=[np.number]).columns
+    if len(num_cols):
+        df_clean[num_cols] = df_clean[num_cols].replace([np.inf, -np.inf], np.nan)
+
+    # 2️⃣ 替换 NaN -> None（注意：有时 astype(object) 不彻底，需用 applymap）
+    df_clean = df_clean.applymap(lambda x: None if pd.isna(x) else x)
+
+    # 3️⃣ 转为 dict records
+    return df_clean.to_dict("records")
+
+
+def clean_json_field_for_orjson(v):
+    """清洗单个 JSON 字段的值，使其符合 orjson 要求并避免空字典入库"""
+    if v is None or pd.isna(v):
+        return None
+
+    # 1️⃣ 如果是空字典对象，返回 None
+    if isinstance(v, dict) and not v:
+        return None
+
+    # 2️⃣ 如果是空字符串或仅为 "{}"，返回 None
+    if isinstance(v, str):
+        stripped = v.strip()
+        if not stripped or stripped == "{}":
+            return None
+        try:
+            parsed = json.loads(stripped)
+            if isinstance(parsed, dict) and not parsed:
+                return None
+            return json.dumps(parsed, ensure_ascii=False)
+        except Exception:
+            return v  # 非 JSON 字符串则原样保留
+
+    return v
+
+
+def fully_clean_for_orjson(df: pd.DataFrame) -> pd.DataFrame:
+    """全面清洗 DataFrame 以符合 orjson 要求"""
+    df = df.replace([np.inf, -np.inf], np.nan)
+    df = df.applymap(lambda x: None if pd.isna(x) else x)
+
+    # 找出所有可能为 JSON 字符串的字段
+    json_like_cols = [col for col in df.columns if col.endswith('_json')]
+
+    # 针对每个 JSON-like 字段，应用清洗函数
+    for col in json_like_cols:
+        df[col] = df[col].apply(clean_json_field_for_orjson)
+
+    return df
+
+
+class RemoteTransaction:
+
+    def __init__(self, db: str, database: str,
+                 session: requests.Session, urls: List[str]):
+        self.db = db
+        self.database = database
+        self.session = session
+        self.urls = urls
+        self.sql_queue = []
+
+    # def execute(self, sql: str):
+    #     self.sql_queue.append(sql)
+    def execute(self, sql: str, params=None):
+        """
+        params 可取：
+        • None → 纯文本 SQL
+        • dict → 单条参数化   e.g. {"id":1,"name":"a"}
+        • list/tuple → 批量 executemany
+            - list[dict]  ↔ INSERT .. VALUES (:id,:name)
+            - list[tuple] ↔ INSERT .. VALUES (%s,%s)
+        """
+        self.sql_queue.append({"sql": sql, "params": params})
+
+    def __enter__(self): return self
+
+    def __exit__(self, exc_type, exc, tb):
+        for url in self.urls:
+            try:
+                self.session.post(
+                    url + "/transaction",
+                    json={"db_type": self.db,
+                          "sql_list": self.sql_queue,
+                          "database": self.database},  # site_name not needed on server, kept for clarity
+                    timeout=3000,
+                ).raise_for_status()
+                return
+            except Exception as e:
+                print(f"[WARN] 事务失败 {url}: {e}")
+        raise RuntimeError("All servers failed for transaction")
+
+
+class RemoteEngine:
+    def __init__(self, db: str, database: str,
+                 server_urls: List[str], retries: int = 2):
+        """
+        :param db: db_type--数据库类型
+        :param database: 数据库名称
+        :param server_urls:
+        :param retries:
+        """
+        self.db = db
+        self.database = database
+        self.urls = [u.rstrip("/") for u in server_urls]
+        self.session = requests.Session()
+        self.retries = retries
+
+    def _request(self, endpoint: str, payload):
+        for url in self.urls:
+            for _ in range(self.retries):
+                try:
+                    json_bytes = orjson.dumps(payload)
+                    r = self.session.post(f"{url}/{endpoint}",
+                                          data=json_bytes,
+                                          headers={"Content-Type": "application/json"},
+                                          timeout=3000)
+
+                    # r = self.session.post(f"{url}/{endpoint}",
+                    #                       json=payload, timeout=10)
+                    r.raise_for_status()
+                    return r.json()
+                except Exception as e:
+                    print(f"[WARN] {endpoint} fail @ {url}: {e}")
+                    time.sleep(1)
+        raise RuntimeError(f"All servers failed for {endpoint}")
+    # def _request(self, endpoint: str, payload):
+    #     # 用 orjson，“allow_nan” 会把 NaN/Inf 写成 null
+    #     # json_bytes = orjson.dumps(payload,
+    #     #                           option=orjson.OPT_NON_STR_KEYS | orjson.OPT_NAIVE_UTC | orjson.OPT_OMIT_MICROSECOND | orjson.OPT_ALLOW_INF_AND_NAN)
+    #     json_bytes = orjson.dumps(
+    #         payload,
+    #         option=orjson.OPT_NON_STR_KEYS | orjson.OPT_NAIVE_UTC | orjson.OPT_ALLOW_INF_AND_NAN
+    #     )
+    #
+    #     headers = {"Content-Type": "application/json"}
+    #
+    #     for url in self.urls:
+    #         for _ in range(self.retries):
+    #             try:
+    #                 r = self.session.post(f"{url}/{endpoint}",
+    #                                       data=json_bytes, headers=headers,
+    #                                       timeout=15)
+    #                 r.raise_for_status()
+    #                 return r.json()
+    #             except Exception as e:
+    #                 print(f"[WARN] {endpoint} fail @ {url}: {e}")
+    #                 time.sleep(1)
+    #     raise RuntimeError(f"All servers failed for {endpoint}")
+
+    # ---------- 公共 API ----------
+    def read_sql(self, sql: str) -> pd.DataFrame:
+        data = self._request("query",
+                             {"db_type": self.db,
+                              "sql": sql,
+                              "database": self.database})
+        return pd.DataFrame(data["result"])
+
+    def to_sql(self, df: pd.DataFrame, table: str, if_exists="append"):
+
+        return self._request("insert",
+                             {"db_type": self.db,
+                              "table": table,
+                              "if_exists": if_exists,
+                              "data": fully_clean_for_orjson(df=df).to_dict("records"),
+                              # "data": df_to_json_records(df),   # ← 清洗后的 records
+                              "database": self.database})
+
+    def read_then_update(
+            self,
+            select_sql: str,
+            update_table: str,
+            set_values: dict,
+            where_keys: List[str],
+            error_if_empty: bool = False,
+    ):
+        """
+        动态生成 UPDATE：把 select_sql 读到的行，按 where_keys 精准更新 set_values
+        返回 (DataFrame, rows_updated)
+        """
+        payload = {
+            "db_type": self.db,
+            "database": self.database,
+            "select_sql": select_sql,
+            "update_table": update_table,
+            "set_values": set_values,
+            "where_keys": where_keys,
+            "error_if_empty": error_if_empty,
+        }
+        resp = self._request("read_then_update", payload)
+        df = pd.DataFrame(resp["read_result"])
+        rows_updated = resp.get("rows_updated", 0)
+        return df
+
+    def begin(self):
+        return RemoteTransaction(self.db, self.database,
+                                 self.session, self.urls)
+# ---------------------------------
+
+# db -- 数据库类型
+# database -- 站点
+
+def get_remote_engine(site_name: str, db_type: str, database: str = None,
+                      servers: List[str] = None) -> RemoteEngine:
+    """
+    :param site_name: 站点
+    :param db_type: 数据库类型
+    :param database: 数据库名称-默认为None， 否则通过站点来匹配
+    :param servers: 服务器url地址
+    :return: 数据库连接对象
+    """
+    if not database:
+        database = site_name_db_dict[site_name]
+    if site_name not in site_name_db_dict:
+        raise ValueError(f"Unknown site_name: {site_name}")
+    if db_type not in db_type_alias_map:
+        raise ValueError(f"Unknown db_type: {db_type}")
+    # print(f"db_type: {db_type_alias_map[db_type]}, database: {database}")
+    return RemoteEngine(
+        db=db_type_alias_map[db_type],
+        database=database,
+        server_urls=servers or DEFAULT_SERVERS,
+    )