同步 Gitee 应用代码更新

.gitignore

@@ -43,6 +43,7 @@ MANIFEST
 .installed.cfg
 
 # Build / packaging
+app.zip
 build/
 dist/
 develop-eggs/
@@ -108,4 +109,4 @@ xlsx/
 OPS_REDIS_MYSQL_OPTIMIZATION_NOTES.md
 架构.md
 *ts.%
-.omc
+.omc

app/.env

@@ -11,9 +11,9 @@ zn_username=%E8%87%AA%E5%8A%A8%E5%8C%96_Robot
 
 client_name=ShuFuAI
 
-# java_api_base=http://47.111.163.154:18080
-java_api_base=http://127.0.0.1:18080
-# java_api_base=http://121.196.149.225:18080
+# java_api_base=http://api.aishufu.top:18080/
+java_api_base=http://api.aishufu.top:18080/
+# java_api_base=http://api.aishufu.top:18080/
 
 # 与 Java 后端共享的 JWT 签名密钥，必须与 backend-java 的 AIIMAGE_JWT_SECRET 完全一致
 AIIMAGE_JWT_SECRET=please-change-this-secret-please-rotate-at-least-32-bytes

app/.env copy

@@ -9,7 +9,7 @@ proxy_mode=2
 client_name=ShuFuAI
 
 
-java_api_base=http://127.0.0.1:18080
-# java_api_base=http://47.111.163.154:18080
+java_api_base=http://api.aishufu.top:18080/
+# java_api_base=http://api.aishufu.top:18080/
 
 

app/.env.example

@@ -11,9 +11,9 @@ zn_username=%E8%87%AA%E5%8A%A8%E5%8C%96_Robot
 client_name=ShuFuAI
 
 
-java_api_base=http://47.111.163.154:18080
-# java_api_base=http://127.0.0.1:18080
-# java_api_base=http://47.111.163.154:18080
+java_api_base=http://api.aishufu.top:18080/
+# java_api_base=http://api.aishufu.top:18080/
+# java_api_base=http://api.aishufu.top:18080/
 
 # 与 Java 后端共享的 JWT 签名密钥，必须与 backend-java 的 AIIMAGE_JWT_SECRET 完全一致
 AIIMAGE_JWT_SECRET=please-change-this-secret-please-rotate-at-least-32-bytes

app/amazon/amazon_base.py

@@ -417,8 +417,8 @@ class AmamzonBase(ZiniaoDriver):
         """
         try:
             timestamp = datetime.now().strftime("%Y-%m-%d %H:%M:%S")
-            if level == "ERROR":
-                show_notification(message, "error")
+            # if level == "ERROR":
+            #     show_notification(message, "error")
             print(f"[{timestamp}] [{self.mark_name}] [{level}] {message}")
         except Exception as e:
             print(f"输出出错,{e}")
@@ -689,8 +689,8 @@ class TaskBase:
         """
         try:
             timestamp = datetime.now().strftime("%Y-%m-%d %H:%M:%S")
-            if level == "ERROR":
-                show_notification(message, "error")
+            # if level == "ERROR":
+            #     show_notification(message, "error")
             print(f"[{timestamp}] [{self.task_name}] [{level}] {message}")
         except Exception as e:
             print(f"输出出错,{e}")
@@ -739,7 +739,7 @@ class TaskBase:
                     if not shop_data:
                         mes = f"获取店铺凭证失败，响应数据: {shop_data.get('message', '未知错误')}"
                         self.log(mes, "ERROR")
-                        show_notification(mes, "ERROR")
+                        # show_notification(mes, "ERROR")
                         continue
 
                     password = shop_data["data"]["password"]

app/amazon/approve.py

@@ -308,8 +308,9 @@ class AmzoneApprove(AmamzonBase):
         self.log("开始执行...")
         num = 0
         retry_num = 0
-
         max_retry_num = 5
+        total_page = 0
+        current_page = 0
         while retry_num < max_retry_num:  # 最多重试3次
             # if num > 3:  #测试
             #     return
@@ -325,13 +326,13 @@ class AmzoneApprove(AmamzonBase):
                     page_pamel = self.tab.eles('xpath://kat-pagination',timeout=5)
                     if len(page_pamel) > 0:
                         current_page = page_pamel[0].sr('xpath:.//ul[@class="pages"]//li[@aria-current="true"]').text
+                        current_page = int(current_page.strip())
 
                         # 总页数
                         total_page = page_pamel[0].sr.eles('xpath:.//ul[@class="pages"]//span[@class="page__inner"][last()]')
                         if len(total_page) > 0:
-                            total_page = total_page[-1].text
-                        else:
-                            total_page = 0
+                            total_page = int(total_page[-1].text.strip())
+
                         self.log(f"当前页码: {current_page} / 总页数: {total_page}")
                 except Exception as e:
                     self.log(f"获取页码失败:{e}")
@@ -543,6 +544,14 @@ class AmzoneApprove(AmamzonBase):
                 self.tab.refresh()
                 self.tab.wait.doc_loaded(raise_err=False,timeout=120)
 
+        #检查页数是否相等，不相等则继续
+        self.log(f"开始检查页数，当前页数 {current_page} / {total_page}")
+        if total_page != 0 and current_page!= 0 and current_page < total_page:
+            self.log(f"检查到页数还未完成，重启浏览器继续")
+            raise RuntimeError(f"与页面的连接已断开，检查到页数还未完成，重启浏览器继续")
+
+
+
 class ApproveTask(TaskBase):
     """审批任务处理类：负责处理产品风险审批任务"""
     task_name = "产品风险审批-TASK"
@@ -674,12 +683,12 @@ class ApproveTask(TaskBase):
                 return
 
             current_url = None
+            max_retries = 200  #只要没有完成，一直重试
             for _ in range(max_retries):
                 try:
                     self.process_country(driver, country_code, task_id, shop_name,risk_listing_filter,target_url=current_url)
                     driver.reset_already_asin()
                     driver.close_store()
-
                     break
                 except Exception as e:
                     self.log(f"处理国家 {country_code} 失败: {str(e)}", "ERROR")
@@ -777,7 +786,7 @@ class ApproveTask(TaskBase):
                 }
             result.append(country_data)
 
-            if len(result) > 20:
+            if len(result) > 10:
                 self.post_result_batch(task_id, shop_name, country_code, result)
                 result = []
 

app/amazon/chrome_base.py

@@ -9,7 +9,7 @@ from datetime import datetime
 from DrissionPage import Chromium, ChromiumOptions
 from collections import defaultdict
 
-from config import base_dir
+from config import base_dir,debug
 from amazon.tool import get_shop_info,show_notification
 
 
@@ -46,7 +46,8 @@ class ChromeAmzoneBase:
         杀死当前谷歌浏览器进程，并使用 drissionpage 启动谷歌浏览器，使用系统安装的浏览器默认用户文件夹
         """
         print("正在关闭现有的chromium浏览器进程...")
-        os.system('taskkill /f /t /im chrome.exe')
+        if not debug:
+            os.system('taskkill /f /t /im chrome.exe')
         time.sleep(2)
 
         print("正在启动chromium浏览器...")
@@ -70,8 +71,8 @@ class ChromeAmzoneBase:
         """
         try:
             timestamp = datetime.now().strftime("%Y-%m-%d %H:%M:%S")
-            if level == "ERROR":
-                show_notification(message, "error")
+            # if level == "ERROR":
+            #     show_notification(message, "error")
             print(f"[{timestamp}] [{self.mark_name}] [{level}] {message}")
         except Exception as e:
             print(f"输出出错,{e}")
@@ -192,4 +193,14 @@ class ChromeAmzoneBase:
                 self.browser.quit()
                 print("浏览器已关闭")
         except Exception as e:
-            print(f"关闭浏览器时出错: {str(e)}")
+            print(f"关闭浏览器时出错: {str(e)}")
+
+
+
+
+
+
+
+
+
+

app/amazon/detail_spider.py

@@ -13,7 +13,7 @@ from amazon.amazon_base import TaskBase
 from amazon.chrome_base import ChromeAmzoneBase
 
 
-from config import runing_task, runing_shop,base_dir,DELETE_BRAND_API_BASE
+from config import runing_task,runing_shop,base_dir,DELETE_BRAND_API_BASE
 
 
 class ChromeAmzone(ChromeAmzoneBase):
@@ -79,7 +79,8 @@ class ChromeAmzone(ChromeAmzoneBase):
         except Exception as e:
             error_msg = f"运行出错: {traceback.format_exc()}"
             self.log(error_msg)
-            show_notification(f"采集失败: {str(e)}", "error")
+            # show_notification(f"采集失败: {str(e)}", "error")
+            raise RuntimeError(error_msg)
             return {}
 
     def _scrape_data(self):
@@ -91,7 +92,8 @@ class ChromeAmzone(ChromeAmzoneBase):
         """
         data = {
             'image_url': "",
-            'title': ""
+            'title': "",
+            "sku" : ""
         }
 
         try:
@@ -100,9 +102,41 @@ class ChromeAmzone(ChromeAmzoneBase):
             title_ele = self.tab.ele('xpath://h1[@id="title"]',timeout=30)
             title = title_ele.text
             data["title"] = title
+            sku_ele_ls = self.tab.eles('xpath://ul[@class="a-unordered-list a-vertical a-spacing-mini"]', timeout=20)
+            if len(sku_ele_ls) > 0:
+                data["sku"] = sku_ele_ls[0].text
             imge_ele = self.tab.ele('xpath://div[@id="imgTagWrapperId"]//img',timeout=20)
-            image_url = imge_ele.attr("src")
+
+            image_url = ""
+            min_image_url = ""
+            data_a_dynamic_image = imge_ele.attr("data-a-dynamic-image")
+            if data_a_dynamic_image:
+                dynamic_image_json = json.loads(data_a_dynamic_image)
+                self.log(f"图片信息:{dynamic_image_json}")
+                max_area = 0
+                min_area = 0
+
+                for url, (width, height) in dynamic_image_json.items():
+                    area = width * height
+                    if area > max_area:
+                        max_area = area
+                        image_url = url
+                    if min_area == 0:
+                        min_area = area
+                        min_image_url = url
+                    if area < min_area:
+                        min_area = area
+                        min_image_url = url
+
+            if not image_url:
+                image_url = imge_ele.attr("src")
+
+            if not min_image_url:
+                min_image_url = imge_ele.attr("src")
+
             data["image_url"] = image_url
+            data["min_image_url"] = min_image_url
+
             return data
 
         except Exception as e:
@@ -246,6 +280,7 @@ class SpiderTask(TaskBase):
                             "country": i.get("country"),
                             "url": return_data.get("image_url"),
                             "title": return_data.get("title"),
+                            "sku": return_data.get("sku")
                         }
                          for i in items
                     ]
@@ -263,7 +298,7 @@ class SpiderTask(TaskBase):
                     result.append(res)
                     print("================")
                     is_done = gp_index == len(groups)-1
-                    if len(result) > 20 or is_done:
+                    if len(result) > 10 or is_done:
                         self.post_result(task_id=task_id,chunkIndex=gp_index+1,chunkTotal=len(groups),
                                          asin=asin,item_data=result,is_done=is_done)
                         result = []

app/amazon/match_action.py

@@ -35,9 +35,12 @@ class AmzoneMatchAction(AmamzonBase):
         self.log("开始执行...")
         num = 0
         retry_num = 0
-        get_page_faild = 0
 
-        while retry_num < 3:  # 最多重试3次
+        max_retry_num = 5
+        total_page = 0
+        current_page = 0
+
+        while retry_num < max_retry_num:
             # if num > 3:  #测试
             #     return
             # 等待加载完成
@@ -45,6 +48,7 @@ class AmzoneMatchAction(AmamzonBase):
                 load_ele = self.tab.eles("xpath://div[contains(@class,'Loader-module__loader')]",timeout=5)
                 if len(load_ele) > 0:
                     load_ele[0].wait.deleted(timeout=3, raise_err=False)
+
                 time.sleep(0.5)
 
                 # 获取当前页码
@@ -52,23 +56,17 @@ class AmzoneMatchAction(AmamzonBase):
                     page_pamel = self.tab.eles('xpath://kat-pagination',timeout=5)
                     if len(page_pamel) > 0:
                         current_page = page_pamel[0].sr('xpath:.//ul[@class="pages"]//li[@aria-current="true"]').text
+                        current_page = int(current_page.strip())
                         # 测试
                         # if current_page > 1:
                         #     break
                         # 总页数
                         total_page = page_pamel[0].sr.eles('xpath:.//ul[@class="pages"]//span[@class="page__inner"][last()]')
                         if len(total_page) > 0:
-                            total_page = total_page[-1].text
-                        else:
-                            total_page = 0
+                            total_page = int(total_page[-1].text.strip())
                         self.log(f"【{self.mark_name}】当前页码: {current_page} / 总页数: {total_page}")
-                    get_page_faild = 0
                 except Exception as e:
                     self.log(f"【{self.mark_name}】获取页码失败", e)
-                    get_page_faild+= 1
-                    if get_page_faild > 2:
-                        show_notification(f"【{self.mark_name}】获取页码失败超3次停止任务！")
-                        break
 
                 # 保存当前URL，用于失败重试时访问
                 try:
@@ -81,6 +79,15 @@ class AmzoneMatchAction(AmamzonBase):
 
                 sku_ls = self.tab.eles("xpath://div[@data-sku]",timeout=10)
                 self.log(f"【{self.mark_name}】获取到 {len(sku_ls)}")
+                if len(sku_ls) == 0:
+                    retry_num += 1
+                    self.log(f"没有获取到SKU列表，重试{retry_num}/{max_retry_num}")
+                    self.tab.refresh()
+                    self.tab.wait.doc_loaded(raise_err=False, timeout=120)
+                    if retry_num == max_retry_num:
+                        raise RuntimeError(f"与页面的连接已断开，没有获取到SKU列表,重试{retry_num}/{max_retry_num}")
+                    continue
+
                 # for sku_ele in sku_ls[0:2]:
                 for sku_ele in sku_ls:
                     # solve_problem = sku_ele.eles('xpath:.//kat-link[@label="解决商品信息问题"]')
@@ -140,6 +147,12 @@ class AmzoneMatchAction(AmamzonBase):
                 self.tab.refresh()
                 self.tab.wait.doc_loaded(raise_err=False,timeout=120)
 
+        #检查页数是否相等，不相等则继续
+        self.log(f"开始检查页数，当前页数 {current_page} / {total_page}")
+        if total_page != 0 and current_page!= 0 and current_page < total_page:
+            self.log(f"检查到页数还未完成，重启浏览器继续")
+            raise RuntimeError(f"与页面的连接已断开，检查到页数还未完成，重启浏览器继续")
+
 
 class MatchTak(TaskBase):
     task_name = "匹配价格-TASK"
@@ -261,7 +274,7 @@ class MatchTak(TaskBase):
                 self.log(f"检测到任务 {task_id} 的暂停请求，停止处理国家", "WARNING")
                 break
 
-            skip_asin = skip_asins_by_country.get(country_code)
+            skip_asin = skip_asins_by_country.get(country_code,[])
             if skipAsinDetailsByCountry :
                 skipAsinDetails = {i.get("asin"):i.get("minimumPrice") for i in skipAsinDetailsByCountry.get(country_code)}
             else:
@@ -276,6 +289,7 @@ class MatchTak(TaskBase):
                 return
 
             current_url = None
+            max_retries = 200
             for _ in range(max_retries):
                 try:
                     self.process_country(driver, country_code, task_id, shop_name,risk_listing_filter,limit,
@@ -366,8 +380,8 @@ class MatchTak(TaskBase):
             if task_id in runing_task:
                 runing_task[task_id]["current_asin"] = asin
                 runing_task[task_id]["processed_asins"] += 1
-
                 runing_task[task_id]["failed_count"] += 1
+
             minimumPrice = ""
             if "有最低价跳过" in status:
                 minimumPrice = skipAsinDetails.get(asin)
@@ -378,7 +392,7 @@ class MatchTak(TaskBase):
                                 "done": False
                             })
 
-            if len(result) > 20:
+            if len(result) > 10:
                 self.post_result_batch(task_id, shop_name, country_code,result)
                 result = []
 

app/amazon/similar_asin.py

@@ -10,9 +10,10 @@ from datetime import datetime
 from collections import defaultdict
 import requests
 from urllib.parse import quote
-
+import os
 from curl_cffi import requests as requests_frp
 
+from config import base_dir
 
 
 from amazon.tool import show_notification,get_shop_info,remove_special_characters,split_currency_values
@@ -76,7 +77,9 @@ class ChromeAmzone(ChromeAmzoneBase):
         将图片链接（URL 或本地文件路径）转换为 Base64 编码的字符串。
         """
         if image_source.startswith(('http://', 'https://')):
-            response = requests.get(image_source, timeout=10)
+            response = requests.get(image_source, timeout=10,headers={
+                "user-agent":"Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/142.0.0.0 Safari/537.36 Edg/142.0.0.0 18444"
+            })
             response.raise_for_status()  # 非 2xx 状态码将抛出异常
             image_data = response.content
         else:
@@ -153,7 +156,7 @@ class ChromeAmzone(ChromeAmzoneBase):
         # 发送第一次请求
         try:
             response = requests_frp.post(url, headers=headers, params=params, json=data, impersonate="chrome101",
-                                     cookies=cookie, proxies=proxies)
+                                     cookies=cookie, proxies=proxies,verify=False)
             response.encoding = "utf-8"
 
             # 检查响应状态码和数据有效性
@@ -176,6 +179,7 @@ class ChromeAmzone(ChromeAmzoneBase):
                     need_retry = True
 
             # 如果需要重试且配置了代理URL
+            # need_retry = True
             if need_retry and CONFIG_PROXY_URL and not proxies:
                 try:
                     proxy_resp = requests.get(CONFIG_PROXY_URL, timeout=10)
@@ -204,7 +208,7 @@ class ChromeAmzone(ChromeAmzoneBase):
                         if proxy_ip:
                             proxies = {
                                 "http": f"http://{proxy_ip}",
-                                "https": f"http://{proxy_ip}",
+                                "https": f"https://{proxy_ip}",
                             }
 
                     if proxies:
@@ -251,8 +255,37 @@ class ChromeAmzone(ChromeAmzoneBase):
             title = title_ele.text
             data["title"] = title
             imge_ele = self.tab.ele('xpath://div[@id="imgTagWrapperId"]//img',timeout=20)
-            image_url = imge_ele.attr("src")
+
+            image_url = ""
+            min_image_url = ""
+            data_a_dynamic_image = imge_ele.attr("data-a-dynamic-image")
+            if data_a_dynamic_image:
+                dynamic_image_json = json.loads(data_a_dynamic_image)
+                self.log(f"图片信息:{dynamic_image_json}")
+                max_area = 0
+                min_area = 0
+
+                for url, (width, height) in dynamic_image_json.items():
+                    area = width * height
+                    if area > max_area:
+                        max_area = area
+                        image_url = url
+                    if min_area == 0:
+                        min_area = area
+                        min_image_url = url
+                    if area < min_area:
+                        min_area = area
+                        min_image_url = url
+
+            if not image_url:
+                image_url = imge_ele.attr("src")
+
+            if not min_image_url:
+                min_image_url = imge_ele.attr("src")
+
+
             data["image_url"] = image_url
+            data["min_image_url"] = min_image_url
 
             # sku
             sku_ele_ls = self.tab.eles('xpath://ul[@class="a-unordered-list a-vertical a-spacing-mini"]',timeout=20)
@@ -277,12 +310,13 @@ class ChromeAmzone(ChromeAmzoneBase):
         Returns:
             dict: 包含采集到的数据
         """
+        return_data = {}
         try:
             # 验证国家是否支持
             if country not in self.country_info:
                 error_msg = f"不支持的国家: {country}，支持的国家有: {list(self.country_info.keys())}"
                 print(error_msg)
-                show_notification(error_msg, "error")
+                # show_notification(error_msg, "error")
                 return None
 
             # 获取国家配置
@@ -314,29 +348,35 @@ class ChromeAmzone(ChromeAmzoneBase):
             self.log("正在抓取商品数据...")
             data = self._scrape_data()
 
+            return_data.update(data)
+
             # 判断是否采集标题出错
             new_size = "220,220"
             pattern = r"\._(?:[A-Z]+)?(\d+)_\."
-            replacement = f"._{new_size}_.jpg"
+            # replacement = f"._{new_size}_.jpg"
             # 使用正则替换
-            image_new_url = re.sub(pattern, lambda m: replacement, data["image_url"])
+            # image_new_url = re.sub(pattern, lambda m: replacement, data["image_url"])
+            image_new_url = data["min_image_url"]
+
+            print(image_new_url)
 
             iamge_base64 = self.image_to_base64(image_new_url)
 
             similar_data = []
             # 4、请求获取插件的数据
-            for page_num in range(total_page):
-                resp_data = self.get_aliprice_data(
-                    page=page_num,
-                    title = data["title"],
-                    domain=domain,
-                    category=data["category"],
-                    imageBase64 = iamge_base64
-                )
-                self.log(f"Aliprice 扩展数据获取:{str(resp_data)[0:200]}")
-                if len(resp_data.get("data",[])) > 0:
-                    for i in resp_data.get("data"):
-                        similar_data.append(i)
+            if image_new_url:
+                for page_num in range(total_page):
+                    resp_data = self.get_aliprice_data(
+                        page=page_num+1,
+                        title = data["title"],
+                        domain=domain,
+                        category=data["category"],
+                        imageBase64 = iamge_base64
+                    )
+                    self.log(f"Aliprice 扩展数据获取:{str(resp_data)[0:200]}")
+                    if len(resp_data.get("data",[])) > 0:
+                        for i in resp_data.get("data"):
+                            similar_data.append(i)
 
             data["similar_data"] = similar_data
 
@@ -346,14 +386,17 @@ class ChromeAmzone(ChromeAmzoneBase):
             data['url'] = product_url
             data['timestamp'] = datetime.now().strftime('%Y-%m-%d %H:%M:%S')
 
-            print(f"数据抓取完成: {json.dumps(data)}")
-            return data
+            print(f"数据抓取完成: {json.dumps(return_data)}")
+            return_data.update(data)
+
+            return return_data
 
         except Exception as e:
             error_msg = f"运行出错: {traceback.format_exc()}"
             print(error_msg)
-            show_notification(f"采集失败: {str(e)}", "error")
-            return {}
+            # show_notification(f"采集失败: {str(e)}", "error")
+            raise RuntimeError(f"{traceback.format_exc()}")
+            return return_data
 
 
 class SimilarAsinTask(TaskBase):
@@ -478,10 +521,9 @@ class SimilarAsinTask(TaskBase):
                 result = []
                 for gp_index,gp in enumerate(groups):
                     items = gp.get("items", [])
-                    return_data = {}
-
+                    group_item = []
                     for index,value in enumerate(items):
-
+                        print(value)
                         return_data = {
                             'image_url': "",
                             'title': "",
@@ -494,37 +536,36 @@ class SimilarAsinTask(TaskBase):
                         country = value.get("country")
                         for _ in range(max_retry):
                             try:
-                                return_data = chrome.run(country, asin,total_page=2) or {}
+                                return_data = chrome.run(country, asin,total_page=1) or {}
                                 self.log(f"抓取结果->{return_data}")
                                 break
                             except Exception as e:
                                 # if "与页面的连接已断开" in str(e):
                                 chrome = ChromeAmzone()
                                 self.log(f"{asin}抓取数据报错,{e}")
-                        if not isinstance(return_data, dict):
-                            return_data = {}
-                        if return_data.get("image_url"):
-                            break
+                        # if not isinstance(return_data, dict):
+                        #     return_data = {}
+                        # if return_data.get("image_url"):
+                        #     break
+
+                        group_item.append({
+                                "sourceFileKey": value.get("sourceFileKey",""),
+                                "sourceFilename": value.get("sourceFilename",""),
+                                "rowToken": value.get("rowToken",""),
+                                "groupKey": value.get("groupKey",""),
+                                "id": value.get("values",{}).get("id",""),
+                                "asin": value.get("asin",""),
+                                "sku" : return_data.get("sku",""),
+                                "country": value.get("country",""),
+                                "url": return_data.get("image_url",""),
+                                "title": return_data.get("title",""),
+                                "done": False,
+                                "urls" : [i.get("ori_picture") for i in return_data.get("similar_data",[])[:16]]
+                            })
+                        # task_id: int, chunkIndex:int,chunkTotal: int, country_code: str, asin: str, status: dict,error:str="",
+                        #                     item_data:dict={},
+
 
-                    group_item = [
-                        {
-                            "sourceFileKey": i.get("sourceFileKey"),
-                            "sourceFilename": i.get("sourceFilename"),
-                            "rowToken": i.get("rowToken"),
-                            "groupKey": i.get("groupKey"),
-                            "id": i.get("id"),
-                            "asin": i.get("asin"),
-                            "sku" : return_data.get("sku"),
-                            "country": i.get("country"),
-                            "url": return_data.get("image_url"),
-                            "title": return_data.get("title"),
-                            "done": False,
-                            "urls" : [i.get("ori_picture") for i in return_data.get("similar_data")]
-                        }
-                         for i in items
-                    ]
-                    # task_id: int, chunkIndex:int,chunkTotal: int, country_code: str, asin: str, status: dict,error:str="",
-                    #                     item_data:dict={},
                     res = {
                       "sourceFileKey": gp.get("sourceFileKey"),
                       "sourceFilename": gp.get("sourceFilename"),
@@ -629,7 +670,7 @@ class SimilarAsinTask(TaskBase):
                 time.sleep(2)
 
         self.log(f"已达到最大重试次数，结果回传最终失败", "ERROR")
-        raise RuntimeError("已达到最大重试次数，结果回传最终失败")
+        # raise RuntimeError("已达到最大重试次数，结果回传最终失败")
 
 if __name__ == '__main__':
     spide = ChromeAmzone()

app/assets/brand-COze15GJ.js

@@ -0,0 +1 @@
+import{bN as r}from"./java-modules-B8c-YG5x.js";const n="";function s(e){return r(`${n}/api/brand/expand-folder-recursive`,{folder:e})}export{s as e};

app/assets/brand-D__LhhPM.js

@@ -1 +0,0 @@
-import{b8 as r}from"./pywebview-3bFwh9Cg.js";const n="";function s(e){return r(`${n}/api/brand/expand-folder-recursive`,{folder:e})}export{s as e};

app/assets/similar-asin-B5CvMmwV.css

@@ -1 +0,0 @@
-.module-page[data-v-49f8fcb3]{min-height:100vh;background:#1a1a1a}.main-content[data-v-49f8fcb3]{display:flex;height:calc(100vh - 56px);min-height:calc(100vh - 56px)}.left-panel[data-v-49f8fcb3]{width:400px;background:#1e1e1e;padding:20px;overflow-y:auto;border-right:1px solid #2a2a2a}.right-panel[data-v-49f8fcb3]{flex:1;min-width:0;background:#1a1a1a;display:flex;flex-direction:column}.section-title[data-v-49f8fcb3],.subsection-title[data-v-49f8fcb3]{font-size:13px;color:#bbb;margin-bottom:10px}.upload-zone[data-v-49f8fcb3]{border:1px dashed #3a3a3a;border-radius:10px;padding:18px;background:#252525;margin-bottom:18px}.hint[data-v-49f8fcb3],.loading-msg[data-v-49f8fcb3],.files[data-v-49f8fcb3],.muted[data-v-49f8fcb3]{color:#888;font-size:12px;line-height:1.5}.link[data-v-49f8fcb3]{color:#6ea8fe;text-decoration:none}.link[data-v-49f8fcb3]:hover{color:#9fc5ff}.btns[data-v-49f8fcb3],.run-row[data-v-49f8fcb3]{display:flex;gap:10px;flex-wrap:wrap}.opt-btn[data-v-49f8fcb3],.btn-run[data-v-49f8fcb3],.btn-delete[data-v-49f8fcb3],.download[data-v-49f8fcb3]{border:none;cursor:pointer;border-radius:7px}.opt-btn[data-v-49f8fcb3]{padding:8px 14px;color:#ccc;background:#2a2a2a;border:1px solid #3a3a3a}.btn-run[data-v-49f8fcb3]{padding:10px 18px;color:#fff;background:#3498db;font-weight:600}.btn-queue[data-v-49f8fcb3]{background:#27ae60}.btn-run[data-v-49f8fcb3]:disabled{opacity:.55;cursor:not-allowed}.selected-files[data-v-49f8fcb3]{margin-top:14px;color:#999;font-size:12px;word-break:break-all}.selected-files span[data-v-49f8fcb3]{display:block;margin:4px 0}.parse-card[data-v-49f8fcb3],.queue-payload[data-v-49f8fcb3]{margin-top:14px;padding:12px;border:1px solid #2a2a2a;border-radius:8px;background:#202020;color:#b8c1cc;font-size:12px}.queue-payload[data-v-49f8fcb3]{max-height:220px;overflow:auto;color:#8fd3ff;white-space:pre-wrap}.panel-header[data-v-49f8fcb3]{padding:16px 20px;border-bottom:1px solid #2a2a2a;font-size:15px;font-weight:600;color:#ddd}.task-list-wrap[data-v-49f8fcb3]{flex:1;padding:16px 20px;overflow:auto}.clean-result-summary[data-v-49f8fcb3]{display:grid;grid-template-columns:repeat(4,minmax(0,1fr));gap:12px;margin-bottom:16px}.summary-card[data-v-49f8fcb3]{padding:14px 16px;border:1px solid #2a2a2a;border-radius:8px;background:#1e1e1e}.summary-card strong[data-v-49f8fcb3]{display:block;margin-top:8px;color:#eaf4ff;font-size:22px}.summary-label[data-v-49f8fcb3]{color:#8d8d8d;font-size:12px}.result-list-wrap[data-v-49f8fcb3]{border:1px solid #2a2a2a;border-radius:8px;background:#1e1e1e;min-height:180px;margin:0 0 16px}.result-list-header[data-v-49f8fcb3]{display:flex;justify-content:space-between;padding:12px 16px;border-bottom:1px solid #2a2a2a;color:#ddd;font-size:14px}.empty-tasks[data-v-49f8fcb3]{color:#666;font-size:13px;padding:18px;text-align:center}.task-list[data-v-49f8fcb3]{list-style:none;margin:0;padding:12px}.task-item[data-v-49f8fcb3]{display:flex;align-items:flex-start;justify-content:space-between;gap:12px;padding:12px 14px;border:1px solid #2a2a2a;border-radius:8px;margin-bottom:8px;background:#222}.left[data-v-49f8fcb3]{flex:1;min-width:0}.id[data-v-49f8fcb3]{color:#e0e0e0;font-size:13px;font-weight:600}.task-right[data-v-49f8fcb3]{display:flex;gap:8px;align-items:center;flex-wrap:wrap}.status[data-v-49f8fcb3]{padding:4px 10px;border-radius:6px;font-size:12px}.status.success[data-v-49f8fcb3]{background:#2ecc712e;color:#2ecc71}.status.failed[data-v-49f8fcb3]{background:#e74c3c2e;color:#ff6b6b}.status.running[data-v-49f8fcb3]{background:#3498db2e;color:#3498db}.status.pending[data-v-49f8fcb3]{background:#95a5a62e;color:#bdc3c7}.result-hint[data-v-49f8fcb3]{margin-top:6px;color:#e0b96d}.file-progress[data-v-49f8fcb3]{margin-top:8px;max-width:520px}.file-progress-meta[data-v-49f8fcb3]{display:flex;justify-content:space-between;gap:12px;color:#d8c278;font-size:12px}.file-progress-track[data-v-49f8fcb3]{margin-top:5px;height:8px;border-radius:999px;overflow:hidden;background:#303030;border:1px solid #3b3b3b}.file-progress-bar[data-v-49f8fcb3]{height:100%;border-radius:inherit;background:linear-gradient(90deg,#4aa3ff,#f0c75e);transition:width .25s ease}.file-progress-count[data-v-49f8fcb3]{margin-top:4px;color:#858585;font-size:11px}.download[data-v-49f8fcb3]{padding:6px 10px;color:#d6ecff;background:#3498db2e}.btn-delete[data-v-49f8fcb3]{padding:6px 10px;color:#ff8f8f;background:#e74c3c1f}@media(max-width:1100px){.main-content[data-v-49f8fcb3]{flex-direction:column;height:auto}.left-panel[data-v-49f8fcb3]{width:100%;border-right:none;border-bottom:1px solid #2a2a2a}.clean-result-summary[data-v-49f8fcb3]{grid-template-columns:repeat(2,minmax(0,1fr))}}

app/blueprints/auth.py

@@ -28,6 +28,7 @@ def api_auth_sync():
     """前端拿到 Java 返回的 JWT 后调用，把 token 写进 Python 同源 cookie。"""
     data = request.get_json(silent=True) or {}
     token = (data.get('token') or '').strip()
+
     if not token:
         return jsonify({'success': False, 'error': '缺少 token'}), 400
     payload, reason = parse_token_with_reason(token)

app/config.py

@@ -22,7 +22,7 @@ proxy_url = os.getenv("proxy_url")
 proxy_mode = int(os.getenv("proxy_mode",1))
 
 client_name=os.getenv("client_name") + ".exe"
-JAVA_API_BASE = os.getenv("java_api_base", "http://127.0.0.1:18080")
+JAVA_API_BASE = os.getenv("java_api_base", "http://api.aishufu.top:18080/").rstrip("/")
 
 # 紫鸟浏览器配置
 from urllib.parse import unquote
@@ -32,7 +32,7 @@ ZN_USERNAME = unquote(ZN_USERNAME)
 ZN_PASSWORD = os.getenv("zn_password", "#20zsg25")
 
 # 删除品牌API配置
-DELETE_BRAND_API_BASE = os.getenv("DELETE_BRAND_API_BASE", JAVA_API_BASE)
+DELETE_BRAND_API_BASE = os.getenv("DELETE_BRAND_API_BASE", JAVA_API_BASE).rstrip("/")
 
 
 cache_path = "./user_data"

app/jwt_util.py

@@ -4,6 +4,7 @@ JWT 工具：解析 Java 端签发的 token、从请求中提取 token。
 """
 import os
 import sys
+import time
 from typing import Optional, Tuple
 
 from flask import request
@@ -25,6 +26,14 @@ except ImportError as _e:
 _DEFAULT_SECRET = "please-change-this-secret-please-rotate-at-least-32-bytes"
 COOKIE_NAME = os.getenv("AIIMAGE_AUTH_COOKIE_NAME", "aiimage_token")
 
+# 桌面端与 Java 服务器时钟可能漂移（笔记本休眠、用户手动改时间、跨时区等），
+# 给 JWT exp/nbf 校验留出容差，避免 /api/auth/sync 因为时间不同步而 401。
+# 默认 5 分钟，可通过 AIIMAGE_JWT_LEEWAY_SECONDS 覆盖。
+try:
+    _JWT_LEEWAY_SECONDS = int(os.getenv("AIIMAGE_JWT_LEEWAY_SECONDS", "300"))
+except ValueError:
+    _JWT_LEEWAY_SECONDS = 300
+
 
 def _signing_key() -> bytes:
     """与 Java JwtService.signingKey 保持一致：UTF-8 字节，不足 32 字节右侧补 0。"""
@@ -35,6 +44,26 @@ def _signing_key() -> bytes:
     return key_bytes
 
 
+def _describe_clock_skew(token: str) -> str:
+    """解析 token 里的 iat/exp，与本地时钟比较，返回 ' iat=.. exp=.. now=.. skew=..s' 字符串。
+    仅供日志使用，不做安全决策；解析失败返回空串，不影响主流程。"""
+    if pyjwt is None:
+        return ""
+    try:
+        unverified = pyjwt.decode(token, options={"verify_signature": False, "verify_exp": False, "verify_iat": False, "verify_nbf": False})
+    except Exception:
+        return ""
+    iat = unverified.get("iat")
+    exp = unverified.get("exp")
+    now = int(time.time())
+    parts = [f"now={now}"]
+    if isinstance(iat, (int, float)):
+        parts.append(f"iat={int(iat)} skew_iat={now - int(iat)}s")
+    if isinstance(exp, (int, float)):
+        parts.append(f"exp={int(exp)} skew_exp={now - int(exp)}s")
+    return ", " + " ".join(parts)
+
+
 def parse_token_with_reason(token: str) -> Tuple[Optional[dict], Optional[str]]:
     """解析 JWT，返回 (payload, error_reason)。
     payload 命中时 error_reason 为 None；失败时 payload 为 None，error_reason 描述根因。"""
@@ -49,10 +78,17 @@ def parse_token_with_reason(token: str) -> Tuple[Optional[dict], Optional[str]]:
         return None, f"无法解析 token header: {type(e).__name__}: {e}"
     alg = header.get("alg") or "?"
     try:
-        payload = pyjwt.decode(token, _signing_key(), algorithms=["HS256", "HS384", "HS512"])
+        payload = pyjwt.decode(
+            token,
+            _signing_key(),
+            algorithms=["HS256", "HS384", "HS512"],
+            leeway=_JWT_LEEWAY_SECONDS,
+        )
     except Exception as e:
         # 常见：ExpiredSignatureError / InvalidSignatureError / DecodeError / InvalidAlgorithmError
-        return None, f"{type(e).__name__}: {e} (token alg={alg})"
+        # 把 token 内 exp/iat 与本地时钟一起打出来，定位"时间漂移"类失败更直接
+        skew_info = _describe_clock_skew(token)
+        return None, f"{type(e).__name__}: {e} (token alg={alg}{skew_info})"
     sub = payload.get("sub")
     try:
         user_id = int(sub) if sub is not None else None

app/main.py

@@ -4,6 +4,8 @@ import json
 import sys
 import shutil
 import os
+import uuid
+import ctypes
 os.makedirs(cache_path,exist_ok=True)
 if not debug:
     today = datetime.datetime.now().strftime("%Y_%m_%d")
@@ -186,17 +188,7 @@ class WindowAPI:
         if not result:
             return {'success': False, 'error': '用户取消'}
         path = result[0] if isinstance(result, (list, tuple)) else result
-        try:
-            import requests
-            resp = requests.get(url.strip(), timeout=120, stream=True)
-            resp.raise_for_status()
-            with open(path, 'wb') as f:
-                for chunk in resp.iter_content(chunk_size=65536):
-                    if chunk:
-                        f.write(chunk)
-            return {'success': True, 'path': path}
-        except Exception as e:
-            return {'success': False, 'error': str(e)}
+        return self._download_url_to_path(url, path)
 
     def save_file_from_url_new(self, url, default_filename='download.bin'):
         """弹窗选择保存位置，从 url 下载文件并保存。根据文件后缀动态设置保存类型。"""
@@ -230,18 +222,124 @@ class WindowAPI:
             return {'success': False, 'error': '用户取消'}
 
         path = result[0] if isinstance(result, (list, tuple)) else result
+        return self._download_url_to_path(url, path)
+
+    def save_file_from_url_with_progress(self, url, default_filename='download.bin', download_id=''):
+        """弹窗选择保存位置，从 url 下载文件并通过前端事件上报下载进度。"""
+        print("调用到save_file_from_url_with_progress方法")
+        if not url or not url.strip():
+            return {'success': False, 'error': '下载地址为空'}
+
+        filename = str(default_filename or 'download.bin').strip() or 'download.bin'
+        ext = os.path.splitext(filename)[1].lower()
+        if ext == '.zip':
+            file_types = ('ZIP 压缩包 (*.zip)', '所有文件 (*.*)')
+        elif ext in ('.xlsx', '.xls'):
+            file_types = ('Excel 文件 (*.xlsx;*.xls)', '所有文件 (*.*)')
+        elif ext == '.csv':
+            file_types = ('CSV 文件 (*.csv)', '所有文件 (*.*)')
+        elif ext == '.txt':
+            file_types = ('文本文件 (*.txt)', '所有文件 (*.*)')
+        elif ext == '.json':
+            file_types = ('JSON 文件 (*.json)', '所有文件 (*.*)')
+        else:
+            file_types = ('所有文件 (*.*)',)
+
+        result = self._window.create_file_dialog(
+            webview.SAVE_DIALOG,
+            save_filename=filename,
+            file_types=file_types
+        )
+        if not result:
+            return {'success': False, 'error': '用户取消'}
+
+        path = result[0] if isinstance(result, (list, tuple)) else result
+        progress_id = str(download_id or uuid.uuid4())
+        return self._download_url_to_path(url, path, progress_id)
+
+    def _download_url_to_path(self, url, path, download_id=''):
+        """先下载到临时文件，完整下载后再替换为最终文件，避免用户打开半成品。"""
+        final_path = os.path.abspath(path)
+        final_dir = os.path.dirname(final_path) or os.getcwd()
+        final_name = os.path.basename(final_path)
+        temp_path = os.path.join(final_dir, f".{final_name}.crdownload")
+        last_progress_emit = 0.0
+        last_progress_percent = -1
         try:
-            import requests
+            os.makedirs(final_dir, exist_ok=True)
+            if os.path.exists(temp_path):
+                os.remove(temp_path)
             resp = requests.get(url.strip(), timeout=120, stream=True)
             resp.raise_for_status()
-            with open(path, 'wb') as f:
+            total = int(resp.headers.get('Content-Length') or 0)
+            downloaded = 0
+            self._emit_download_progress(download_id, 'running', final_path, downloaded, total)
+            with open(temp_path, 'wb') as f:
+                self._hide_file(temp_path)
                 for chunk in resp.iter_content(chunk_size=65536):
                     if chunk:
                         f.write(chunk)
-            return {'success': True, 'path': path}
+                        downloaded += len(chunk)
+                        percent = int(downloaded * 100 / total) if total else 0
+                        now_time = time.time()
+                        if percent != last_progress_percent or now_time - last_progress_emit >= 0.2:
+                            self._emit_download_progress(download_id, 'running', final_path, downloaded, total)
+                            last_progress_emit = now_time
+                            last_progress_percent = percent
+                f.flush()
+                os.fsync(f.fileno())
+            self._show_file(temp_path)
+            os.replace(temp_path, final_path)
+            self._emit_download_progress(download_id, 'success', final_path, downloaded, total)
+            return {'success': True, 'path': final_path}
         except Exception as e:
+            try:
+                if os.path.exists(temp_path):
+                    os.remove(temp_path)
+            except Exception:
+                pass
+            self._emit_download_progress(download_id, 'failed', final_path, 0, 0, str(e))
             return {'success': False, 'error': str(e)}
 
+    def _emit_download_progress(self, download_id, status, path, downloaded, total, error=''):
+        if not download_id:
+            return
+        payload = {
+            'id': download_id,
+            'status': status,
+            'path': path,
+            'downloaded': downloaded,
+            'total': total,
+            'percent': round(downloaded * 100 / total, 1) if total else 0,
+            'error': error,
+        }
+        script = (
+            "window.dispatchEvent(new CustomEvent('pywebview-download-progress', "
+            f"{{ detail: {json.dumps(payload, ensure_ascii=False)} }}));"
+        )
+        try:
+            self._window.evaluate_js(script)
+        except Exception:
+            pass
+
+    def _hide_file(self, path):
+        """Windows 下隐藏临时下载文件，减少用户误打开半成品的机会。"""
+        if os.name != 'nt':
+            return
+        try:
+            ctypes.windll.kernel32.SetFileAttributesW(path, 0x02)
+        except Exception:
+            pass
+
+    def _show_file(self, path):
+        """恢复普通文件属性，避免最终保存文件被隐藏。"""
+        if os.name != 'nt':
+            return
+        try:
+            ctypes.windll.kernel32.SetFileAttributesW(path, 0x80)
+        except Exception:
+            pass
+
     def save_template_xlsx(self):
         """弹窗选择保存位置，将品牌文档格式模板 xlsx 保存到用户选择的位置。"""
         template_name = '品牌文档格式_模板.xlsx'
@@ -440,7 +538,7 @@ def main():
     
     api = WindowAPI(window)
     window.expose(api.close, api.minimize, api.maximize, api.toggle_maximize, generate_images, api.save_image, api.save_image_to_folder, api.select_folder, api.select_brand_xlsx_files, api.select_brand_folder,
-                  api.save_file_from_url, api.save_template_xlsx,api.save_template_zip,api.upload_file_to_java,api.open_external_url,api.enqueue_json,api.get_detail_del,api.save_file_from_url_new,api.get_device_id)
+                  api.save_file_from_url, api.save_template_xlsx,api.save_template_zip,api.upload_file_to_java,api.open_external_url,api.enqueue_json,api.get_detail_del,api.save_file_from_url_new,api.save_file_from_url_with_progress,api.get_device_id)
     webview.start(
         debug=True,
         storage_path=cache_path,

app/new_web_source/convert.html

@@ -5,10 +5,10 @@
     <meta name="viewport" content="width=device-width, initial-scale=1.0" />
     <title>格式转换 - 数富AI</title>
     <script type="module" crossorigin src="/assets/convert.js"></script>
-    <link rel="modulepreload" crossorigin href="/assets/pywebview-3bFwh9Cg.js">
-    <link rel="modulepreload" crossorigin href="/assets/brand-D__LhhPM.js">
-    <link rel="stylesheet" crossorigin href="/assets/pywebview-B7UKcPVJ.css">
-    <link rel="stylesheet" crossorigin href="/assets/convert-BKSNvX8i.css">
+    <link rel="modulepreload" crossorigin href="/assets/java-modules-B8c-YG5x.js">
+    <link rel="modulepreload" crossorigin href="/assets/brand-COze15GJ.js">
+    <link rel="stylesheet" crossorigin href="/assets/java-modules-l5anrOZ2.css">
+    <link rel="stylesheet" crossorigin href="/assets/convert-6oxZNMye.css">
   </head>
   <body>
     <div id="app"></div>

app/new_web_source/dedupe.html

@@ -5,10 +5,10 @@
     <meta name="viewport" content="width=device-width, initial-scale=1.0" />
     <title>数据去重 - 数富AI</title>
     <script type="module" crossorigin src="/assets/dedupe.js"></script>
-    <link rel="modulepreload" crossorigin href="/assets/pywebview-3bFwh9Cg.js">
-    <link rel="modulepreload" crossorigin href="/assets/brand-D__LhhPM.js">
-    <link rel="stylesheet" crossorigin href="/assets/pywebview-B7UKcPVJ.css">
-    <link rel="stylesheet" crossorigin href="/assets/dedupe-DNlVfFj-.css">
+    <link rel="modulepreload" crossorigin href="/assets/java-modules-B8c-YG5x.js">
+    <link rel="modulepreload" crossorigin href="/assets/brand-COze15GJ.js">
+    <link rel="stylesheet" crossorigin href="/assets/java-modules-l5anrOZ2.css">
+    <link rel="stylesheet" crossorigin href="/assets/dedupe-DpfPYQDv.css">
   </head>
   <body>
     <div id="app"></div>

app/new_web_source/delete-brand.html

@@ -5,11 +5,11 @@
     <meta name="viewport" content="width=device-width, initial-scale=1.0" />
     <title>删除品牌 - 数富AI</title>
     <script type="module" crossorigin src="/assets/delete-brand.js"></script>
-    <link rel="modulepreload" crossorigin href="/assets/pywebview-3bFwh9Cg.js">
-    <link rel="modulepreload" crossorigin href="/assets/brand-D__LhhPM.js">
+    <link rel="modulepreload" crossorigin href="/assets/java-modules-B8c-YG5x.js">
+    <link rel="modulepreload" crossorigin href="/assets/brand-COze15GJ.js">
     <link rel="modulepreload" crossorigin href="/assets/categorized-timers-JPA-olTr.js">
-    <link rel="stylesheet" crossorigin href="/assets/pywebview-B7UKcPVJ.css">
-    <link rel="stylesheet" crossorigin href="/assets/delete-brand-BJod9K65.css">
+    <link rel="stylesheet" crossorigin href="/assets/java-modules-l5anrOZ2.css">
+    <link rel="stylesheet" crossorigin href="/assets/delete-brand-DZ3x5hOX.css">
   </head>
   <body>
     <div id="app"></div>

app/new_web_source/split.html

@@ -5,10 +5,10 @@
     <meta name="viewport" content="width=device-width, initial-scale=1.0" />
     <title>数据拆分 - 数富AI</title>
     <script type="module" crossorigin src="/assets/split.js"></script>
-    <link rel="modulepreload" crossorigin href="/assets/pywebview-3bFwh9Cg.js">
-    <link rel="modulepreload" crossorigin href="/assets/brand-D__LhhPM.js">
-    <link rel="stylesheet" crossorigin href="/assets/pywebview-B7UKcPVJ.css">
-    <link rel="stylesheet" crossorigin href="/assets/split-BoBVrLdC.css">
+    <link rel="modulepreload" crossorigin href="/assets/java-modules-B8c-YG5x.js">
+    <link rel="modulepreload" crossorigin href="/assets/brand-COze15GJ.js">
+    <link rel="stylesheet" crossorigin href="/assets/java-modules-l5anrOZ2.css">
+    <link rel="stylesheet" crossorigin href="/assets/split-CR1PtDQE.css">
   </head>
   <body>
     <div id="app"></div>

backend-java/src/main/java/com/nanri/aiimage/common/util/FailedStatusRowFilter.java

@@ -9,12 +9,12 @@ import java.util.function.Predicate;
 
 public final class FailedStatusRowFilter {
 
-    private static final String STATUS_HEADER_CN = "\u72b6\u6001";
-    private static final String FAILED_STATUS_ERROR = "\u9519\u8bef";
-    private static final String FAILED_STATUS_FAILED_CN = "\u5931\u8d25";
+    private static final String STATUS_HEADER_CN = "状态";
+    private static final String FAILED_STATUS_ERROR = "错误";
+    private static final String FAILED_STATUS_FAILED_CN = "失败";
     private static final String FAILED_STATUS_FAILED_EN = "failed";
     private static final String NO_MATCHED_ROWS_MESSAGE =
-            "\u72b6\u6001\u5217\u8fc7\u6ee4\u540e\u672a\u627e\u5230 \u9519\u8bef/\u5931\u8d25/FAILED \u6570\u636e";
+            "状态列过滤后未找到 错误/失败/FAILED 数据";
 
     private static final Set<String> STATUS_HEADER_ALIASES = Set.of(
             normalizeHeader(STATUS_HEADER_CN),
@@ -56,8 +56,8 @@ public final class FailedStatusRowFilter {
 
     /**
      * 暴露表头归一化逻辑给其它模块复用，避免每个模块各自维护一份规则。
-     * 处理：去 BOM、全角空格转半角、首尾 trim、连续空白合一、转小写，
-     *      并剥除常见分隔符（含中文全角括号、方括号、冒号）。
+     * 处理：去 BOM、全角空格转半角、首尾 trim、连续空白合并、转小写，
+     * 并剔除常见分隔符（含中文全角括号、方括号、冒号）。
      */
     public static String canonicalizeHeader(String value) {
         return normalizeHeader(value);
@@ -92,8 +92,8 @@ public final class FailedStatusRowFilter {
     }
 
     private static String normalizeHeader(String value) {
-        // 扩展：兼容中文全角括号 （）、中文方括号 【】、中文冒号 ：，避免
-        // "状态（结果）"、"标题维度（商标）" 等全角括号表头被识别失败。
+        // 兼容中文全角括号（）、中文方括号【】、中文冒号：，
+        // 避免“状态（结果）”“标题维度（商标）”等表头识别失败。
         return normalizeValue(value).replaceAll("[\\s_\\-()\\[\\]{}:：（）【】]+", "");
     }
 

backend-java/src/main/java/com/nanri/aiimage/config/SimilarAsinProperties.java

@@ -16,7 +16,14 @@ public class SimilarAsinProperties {
     private String cozeToken = "";
     private List<CozeCredential> cozeCredentials = new ArrayList<>();
     private int cozeCredentialStripeSize = 0;
-    private int cozeBatchSize = 10;
+    /**
+     * P0-1：单次提交 Coze 工作流的 row 数量。
+     * 历史值 10，在含 puzzle 多图行的场景下频繁触发 720712008
+     * "node executed out of limit: 1000"。降到 3 以避免节点上限被打爆。
+     * 出现持续 720712008 时还会被 P1-1 滑窗自适应再降到 1。
+     * 不影响 AppearancePatentProperties 的同名值。
+     */
+    private int cozeBatchSize = 3;
     private int cozeConnectTimeoutMillis = 10000;
     private int cozeReadTimeoutMillis = 60000;
     private int cozePollIntervalMillis = 30000;
@@ -24,6 +31,54 @@ public class SimilarAsinProperties {
     private int staleTimeoutMinutes = 30;
     private String staleFinalizeCron = "0 */2 * * * *";
 
+    /**
+     * 同一 credential 两次提交之间的最小间隔（毫秒）。
+     * 历史值硬编码 30000（持锁 sleep），导致单凭证仅 2 batch/分钟。
+     * 几千行任务场景下成为提交吞吐瓶颈，下调到 5000ms 并改为锁外冷却。
+     * 出现 Coze 限流加重时可通过 AIIMAGE_SIMILAR_ASIN_COZE_SUBMIT_MIN_INTERVAL_MILLIS 调高。
+     */
+    private long cozeSubmitMinIntervalMillis = 5000L;
+
+    /**
+     * 末尾零头 batch 的强制 flush 阈值（分钟）：当不足 cozeBatchSize 的零头 row
+     * 长时间挂着（Python 慢回传）时触发提交。
+     * 任务级实测：345 行 / 4h 总耗时中，约 2-3 小时是 batch 永远凑不满 batchSize 在等下一波回传，
+     * 把阈值从 15 调到 1：最多 60s 后 1-2 行也强制提交，让 Coze 提交侧持续进票，
+     * 总耗时降到与 Python 回传节奏接近。配合 cozeBatchSize=3、cozeSubmitMinIntervalMillis=5000，
+     * 实际不会触发 Coze 限流。出现限流加重再调回 5/10。
+     */
+    private int cozeFlushPendingMinutes = 1;
+
+    /**
+     * 同 batch retry + split retry 共享的最大重试次数。原硬编码 5。
+     */
+    private int cozeSubmitMaxRetryCount = 5;
+
+    /**
+     * 图片嵌入下载线程池大小。原 SimilarAsinImageEmbedder.DOWNLOAD_POOL_SIZE = 8。
+     * P2-10：1000+ 行 ×3 列图片场景下，pool=16 仍是 assemble 阶段瓶颈（实测下载 244s/918s），
+     * 提到 32 配合 retry=2、global deadline 显著拉低尾延迟；
+     * 受 2GB 堆约束，单图缩略图维持 300KB 以内，整体内存峰值 ≈ 32 * 300KB ≈ 10MB。
+     */
+    private int imageDownloadPoolSize = 32;
+
+    /**
+     * 单张图片下载超时（秒）。
+     * P2-10：放宽到 8 + retry=1 在快源（aiproxy/m.media-amazon）下没问题，
+     * 但慢源（cbu01.alicdn）会一直挂 8s 才进入 retry，整体串行时间放大。
+     * 调到 5s + retry=2，让慢源更早重试新连接，单图最坏耗时 ≈ 5s * (1+2) = 15s。
+     */
+    private int imageDownloadTimeoutSeconds = 5;
+
+    /**
+     * assemble 阶段 taskImageCache 的字节上限。
+     * 默认 256MB：5000 行 × 3 列 × 平均 100KB = 1.5GB 远超 2GB 堆，
+     * 用 BoundedImageCache 按字节累计 LRU 淘汰避免爆堆。
+     * 由于 embed() 写完即 remove()，活跃图片字节通常 ≤ 100MB，仅在极端 prefetch 领先场景才会触发淘汰。
+     * 出现淘汰过频影响命中率时可上调到 512MB；2GB 堆约束下不建议超过 768MB。
+     */
+    private long imageCacheMaxBytes = 256L * 1024L * 1024L;
+
     /**
      * 是否在 Coze 请求 parameters 中附带 api_key 字段。
      * 默认 true：线上 Coze 工作流将该字段视为必填，缺失会得到 4000
@@ -41,6 +96,31 @@ public class SimilarAsinProperties {
      */
     private boolean cozeUseLegacyItemFieldOrder = false;
 
+    /**
+     * 是否启用 P0-3 merge 增量缓冲：每个 batch DONE 时仅缓冲 cozeRows，
+     * 不立即合并到 chunk；finalize 前一次性按 chunkScopeHash 分组合并，
+     * 把 OSS chunk 读写从 1000+ 次降到 chunk 数量级。
+     * 仅作用于"正常 poll DONE"路径；失败 batch / 单 batch 任务 / 其他
+     * 11 个 mergeCozeRowsIntoChunk 调用点保留原立即 merge 行为。
+     * 出现问题时可通过 AIIMAGE_SIMILAR_ASIN_COZE_RESULT_BUFFER_ENABLED=false
+     * 一键回滚到老路径。
+     */
+    private boolean cozeResultBufferEnabled = true;
+
+    /**
+     * P0-4：单 credential 抢 Coze 提交锁的最长等待时间（毫秒）。
+     * 原硬编码 1000ms，在高并发 split retry 时大量抛 "Coze submit throttle lock timeout"
+     * 并把整批行 markFailed。应与 cozeSubmitMinIntervalMillis（5000ms）保持 1.5-2 倍关系，
+     * 默认 10000ms 给抢锁更多时间。
+     */
+    private long cozeSubmitLockWaitMillis = 10000L;
+
+    /**
+     * P0-4：抢 Coze 提交锁失败后下次重试间隔（毫秒）。
+     * 原硬编码 500ms，会在指数退避算法中作为基础值（500/1000/2000/4000ms 上限 4000）。
+     */
+    private long cozeSubmitLockRetryDelayMillis = 500L;
+
     @Data
     public static class CozeCredential {
         private String name;

backend-java/src/main/java/com/nanri/aiimage/modules/appearancepatent/client/AppearancePatentCozeClient.java

@@ -355,13 +355,14 @@ public class AppearancePatentCozeClient {
         List<String> rowIds = rows.stream().map(row -> nonBlank(row.getId(), "")).toList();
         List<String> asins = rows.stream().map(row -> nonBlank(row.getAsin(), "")).toList();
         List<String> countries = rows.stream().map(row -> nonBlank(row.getCountry(), "")).toList();
+        List<String> skus = rows.stream().map(row -> nonBlank(row.getSku(), "")).toList();
         List<String> titles = rows.stream().map(row -> nonBlank(row.getTitle(), row.getAsin())).toList();
         List<String> urls = rows.stream().map(row -> nonBlank(row.getUrl(), "")).toList();
 
         Map<String, Object> parameters = new LinkedHashMap<>();
         parameters.put("title_list", titles);
         parameters.put("url_list", urls);
-        parameters.put("items", buildItemObjects(rows, groupKeys, rowIds, asins, countries, titles, urls));
+        parameters.put("items", buildItemObjects(rows, groupKeys, rowIds, asins, countries, skus, titles, urls));
         // 前端没填 prompt 就一律不向 Coze 透传该字段，避免无关默认提示词污染工作流。
         if (prompt != null && !prompt.isBlank()) {
             parameters.put("prompt", prompt);
@@ -403,6 +404,7 @@ public class AppearancePatentCozeClient {
                                                        List<String> rowIds,
                                                        List<String> asins,
                                                        List<String> countries,
+                                                       List<String> skus,
                                                        List<String> titles,
                                                        List<String> urls) {
         List<Map<String, Object>> items = new ArrayList<>(rows.size());
@@ -412,6 +414,7 @@ public class AppearancePatentCozeClient {
             item.put("row_id", rowIds.get(i));
             item.put("asin", asins.get(i));
             item.put("country", countries.get(i));
+            item.put("sku", skus.get(i));
             item.put("title", titles.get(i));
             item.put("url", urls.get(i));
             items.add(item);
@@ -465,7 +468,9 @@ public class AppearancePatentCozeClient {
                         text(firstNonNull(node.get("patent_reason"),
                                 firstNonNull(firstNonNull(node.get("patentReason"), node.get("patent reason")),
                                         firstNonNull(itemNode.get("patent_reason"),
-                                                firstNonNull(itemNode.get("patentReason"), itemNode.get("patent reason"))))))
+                                                firstNonNull(itemNode.get("patentReason"), itemNode.get("patent reason")))))),
+                        text(firstNonNull(firstNonNull(node.get("score"), node.get("Score")),
+                                firstNonNull(itemNode.get("score"), itemNode.get("Score"))))
                 ));
             }
         }
@@ -570,6 +575,7 @@ public class AppearancePatentCozeClient {
         row.setTitleReason(result.titleReason());
         row.setAppearanceReason(result.appearanceReason());
         row.setPatentReason(result.patentReason());
+        row.setScore(result.score());
     }
 
     private RestClient restClient() {
@@ -588,6 +594,7 @@ public class AppearancePatentCozeClient {
         row.setId(source.getId());
         row.setAsin(source.getAsin());
         row.setCountry(source.getCountry());
+        row.setSku(source.getSku());
         row.setUrl(source.getUrl());
         row.setTitle(source.getTitle());
         row.setError(source.getError());
@@ -600,13 +607,14 @@ public class AppearancePatentCozeClient {
         row.setTitleReason(source.getTitleReason());
         row.setAppearanceReason(source.getAppearanceReason());
         row.setPatentReason(source.getPatentReason());
+        row.setScore(source.getScore());
         return row;
     }
 
     private AppearancePatentResultRowDto markFailed(AppearancePatentResultRowDto row, String failureMessage) {
         String reviewMessage = failureMessage == null || failureMessage.isBlank()
-                ? "Coze 检测失败，待人工复核"
-                : "Coze 检测失败，待人工复核：" + failureMessage;
+                ? "Coze 检测失败"
+                : "Coze 检测失败：" + failureMessage;
         if (row.getError() == null || row.getError().isBlank()) {
             row.setError(failureMessage);
         }
@@ -797,7 +805,9 @@ public class AppearancePatentCozeClient {
                 || item.has("appearanceReason")
                 || item.has("patent_reason")
                 || item.has("patentReason")
-                || item.has("patent reason"));
+                || item.has("patent reason")
+                || item.has("score")
+                || item.has("Score"));
     }
 
     private boolean isSuccessfulWorkflowStatus(String status) {
@@ -1054,7 +1064,8 @@ public class AppearancePatentCozeClient {
             String status,
             String titleReason,
             String appearanceReason,
-            String patentReason
+            String patentReason,
+            String score
     ) {
     }
 

backend-java/src/main/java/com/nanri/aiimage/modules/appearancepatent/model/dto/AppearancePatentResultRowDto.java

@@ -29,6 +29,10 @@ public class AppearancePatentResultRowDto {
     @Schema(description = "站点或国家。来自 Excel 的国家列，例如英国、德国、法国。", example = "英国")
     private String country;
 
+    @Schema(description = "商品 SKU。Java 调用 Coze 时会放入 items[].sku。", example = "SKU-001")
+    @JsonAlias({"SKU", "sellerSku", "seller_sku", "merchantSku", "merchant_sku", "商品SKU", "商品 sku", "库存SKU"})
+    private String sku;
+
     @Schema(description = "商品主图或待检测图片 URL。Java 调用 Coze 时会放入 url_list。", example = "https://webstatic.aiproxy.vip/output/20260425/103322/demo.jpg")
     @JsonAlias({
             "imageUrl", "image_url", "imgUrl", "img_url", "pictureUrl", "picture_url",
@@ -78,6 +82,10 @@ public class AppearancePatentResultRowDto {
     @Schema(description = "Coze patent reason", accessMode = Schema.AccessMode.READ_ONLY)
     private String patentReason;
 
+    @JsonAlias({"score", "Score", "评分"})
+    @Schema(description = "Coze 回流的评分（外观维度），透传到最终 xlsx 第一张 sheet 的“评分”列。", accessMode = Schema.AccessMode.READ_ONLY)
+    private String score;
+
     /**
      * 仅内存生命周期标记，标识当前 status 是 markFailed 时合成出来的（而不是用户/Python 真实回传）。
      * 不入库、不参与 chunk 序列化（@JsonIgnore），用于导出 / 重新上传判定时区分"系统合成 FAILED"与"用户真正失败"，

backend-java/src/main/java/com/nanri/aiimage/modules/appearancepatent/model/vo/AppearancePatentParsedRowVo.java

@@ -36,6 +36,9 @@ public class AppearancePatentParsedRowVo {
     @Schema(description = "价格。", example = "12.99")
     private String price;
 
+    @Schema(description = "商品 SKU。Java 调用 Coze 时会放入 items[].sku。", example = "SKU-001")
+    private String sku;
+
     @Schema(description = "商品图片 URL 或商品 URL，供 Coze 检测使用。", example = "https://webstatic.aiproxy.vip/output/demo.jpg")
     private String url;
 

backend-java/src/main/java/com/nanri/aiimage/modules/appearancepatent/service/AppearancePatentTaskService.java

@@ -126,9 +126,9 @@ public class AppearancePatentTaskService {
             "价格",
             "标题维度（商标）",
             "外观维度（外观设计专利）",
-            "专利维度（发明/实用新型专利）",
             "结论",
-            "状态"
+            "状态",
+            "评分"
     );
 
     private final LocalFileStorageService localFileStorageService;
@@ -1254,6 +1254,7 @@ public class AppearancePatentTaskService {
         row.setId(sibling.getDisplayId());
         row.setAsin(sibling.getAsin());
         row.setCountry(sibling.getCountry());
+        row.setSku(firstNonBlank(representative.getSku(), sibling.getSku()));
         row.setUrl(firstNonBlank(representative.getUrl(), sibling.getUrl()));
         row.setTitle(firstNonBlank(representative.getTitle(), sibling.getTitle()));
         row.setError(representative.getError());
@@ -1265,6 +1266,7 @@ public class AppearancePatentTaskService {
         row.setTitleReason(representative.getTitleReason());
         row.setAppearanceReason(representative.getAppearanceReason());
         row.setPatentReason(representative.getPatentReason());
+        row.setScore(representative.getScore());
         row.setDone(representative.getDone());
         return row;
     }
@@ -2176,8 +2178,8 @@ public class AppearancePatentTaskService {
                 || normalized.contains("execution limit")
                 || normalized.contains("720712008")
                 || normalized.contains("720701002")
-                || normalized.contains("\u5de5\u4f5c\u6d41\u8282\u70b9\u6267\u884c\u8d85\u9650")
-                || normalized.contains("\u8c03\u7528\u8d85\u65f6");
+                || normalized.contains("工作流节点执行超限")
+                || normalized.contains("调用超时");
     }
 
     private boolean isRetryableCozeFailure(String failureMessage) {
@@ -2195,10 +2197,10 @@ public class AppearancePatentTaskService {
                 || normalized.contains("720712008")
                 || normalized.contains("720701002")
                 || normalized.contains("plugin limit")
-                || normalized.contains("\u9650\u6d41")
-                || normalized.contains("\u7a0d\u540e\u91cd\u8bd5")
-                || normalized.contains("\u5de5\u4f5c\u6d41\u8282\u70b9\u6267\u884c\u8d85\u9650")
-                || normalized.contains("\u8c03\u7528\u8d85\u65f6");
+                || normalized.contains("限流")
+                || normalized.contains("稍后重试")
+                || normalized.contains("工作流节点执行超限")
+                || normalized.contains("调用超时");
     }
 
     private void updateCozeStateRunning(TaskScopeStateEntity state, String error) {
@@ -3213,6 +3215,7 @@ public class AppearancePatentTaskService {
             List<String> resultHeaders = new ArrayList<>(RESULT_HEADERS);
             resultHeaders.add(4, "标题");
             resultHeaders.add(5, "图片链接");
+            resultHeaders.add(6, "sku");
             Row header = sheet.createRow(0);
             for (int i = 0; i < resultHeaders.size(); i++) {
                 Cell cell = header.createCell(i);
@@ -3236,11 +3239,12 @@ public class AppearancePatentTaskService {
                 row.createCell(col++).setCellValue(firstNonBlank(parsedRow.getPrice(), ""));
                 row.createCell(col++).setCellValue(resultRow == null ? firstNonBlank(parsedRow.getTitle(), "") : firstNonBlank(resultRow.getTitle(), parsedRow.getTitle()));
                 row.createCell(col++).setCellValue(resultRow == null ? firstNonBlank(parsedRow.getUrl(), "") : firstNonBlank(resultRow.getUrl(), parsedRow.getUrl()));
+                row.createCell(col++).setCellValue(resultRow == null ? firstNonBlank(parsedRow.getSku(), "") : firstNonBlank(resultRow.getSku(), parsedRow.getSku()));
                 row.createCell(col++).setCellValue(resultRow == null ? missingReason : userFacingCozeCellValue(resultRow, resultRow.getTitleRisk()));
                 row.createCell(col++).setCellValue(resultRow == null ? missingReason : userFacingCozeCellValue(resultRow, resultRow.getAppearanceRisk()));
-                row.createCell(col++).setCellValue(resultRow == null ? missingReason : userFacingCozeCellValue(resultRow, resultRow.getPatentRisk()));
                 row.createCell(col++).setCellValue(resultRow == null ? "未送检" : userFacingConclusion(resultRow));
-                row.createCell(col).setCellValue(resultRow == null ? "" : userFacingStatus(resultRow));
+                row.createCell(col++).setCellValue(resultRow == null ? "" : userFacingStatus(resultRow));
+                row.createCell(col).setCellValue(resultRow == null ? "" : firstNonBlank(resultRow.getScore(), ""));
             }
             writeReasonSheet(workbook, headerStyle, rowsToWrite, resultMap);
             workbook.write(fos);
@@ -3376,6 +3380,8 @@ public class AppearancePatentTaskService {
             int asinCol = findRequiredHeader(headerMap, "asin");
             int countryCol = findRequiredHeader(headerMap, "国家", "country");
             int priceCol = findOptionalHeaderExact(headerMap, "价格", "price");
+            int skuCol = findOptionalHeaderExact(headerMap,
+                    "sku", "seller sku", "seller_sku", "merchant sku", "merchant_sku", "商品sku", "商品 sku", "库存sku");
             int urlCol = findOptionalHeaderExact(headerMap,
                     "url", "rul", "link", "image", "img", "pic", "picture",
                     "链接", "商品链接", "图片", "商品图片", "主图", "商品主图", "图片链接", "主图链接");
@@ -3423,6 +3429,7 @@ public class AppearancePatentTaskService {
                 vo.setAsin(asin);
                 vo.setCountry(country);
                 vo.setPrice(priceCol >= 0 ? cell(row, priceCol, formatter) : "");
+                vo.setSku(skuCol >= 0 ? cell(row, skuCol, formatter) : "");
                 vo.setUrl(urlCol >= 0 ? cell(row, urlCol, formatter) : "");
                 vo.setTitle(titleCol >= 0 ? cell(row, titleCol, formatter) : "");
                 parsedRows.add(new ParsedAppearanceRow(vo, statusCol >= 0 ? cell(row, statusCol, formatter) : ""));
@@ -3467,9 +3474,11 @@ public class AppearancePatentTaskService {
         for (List<AppearancePatentParsedRowVo> siblings : rowsByBaseId.values()) {
             String title = "";
             String url = "";
+            String sku = "";
             for (AppearancePatentParsedRowVo sibling : siblings) {
                 title = firstNonBlank(title, sibling.getTitle());
                 url = firstNonBlank(url, sibling.getUrl());
+                sku = firstNonBlank(sku, sibling.getSku());
             }
             for (AppearancePatentParsedRowVo sibling : siblings) {
                 if (normalize(sibling.getTitle()).isBlank()) {
@@ -3478,6 +3487,9 @@ public class AppearancePatentTaskService {
                 if (normalize(sibling.getUrl()).isBlank()) {
                     sibling.setUrl(url);
                 }
+                if (normalize(sibling.getSku()).isBlank()) {
+                    sibling.setSku(sku);
+                }
             }
         }
     }
@@ -4041,11 +4053,12 @@ public class AppearancePatentTaskService {
         if (!normalizedValue.isBlank() && !isTechnicalCozeFailure(normalizedValue)) {
             return value;
         }
+        // coze 技术性失败：有错误信息则放入错误信息，没有则留空
         if (row != null && isTechnicalCozeFailure(row.getError())) {
-            return "待人工复核";
+            return firstNonBlank(row.getError(), "");
         }
         if (row != null && isFailedCozeStatusValue(row.getStatus())) {
-            return "待人工复核";
+            return firstNonBlank(row.getError(), "");
         }
         return firstNonBlank(value, "");
     }
@@ -4058,11 +4071,12 @@ public class AppearancePatentTaskService {
         if (!conclusion.isBlank() && !isTechnicalCozeFailure(conclusion)) {
             return row.getConclusion();
         }
+        // coze 技术性失败：有错误信息则放入错误信息，没有则留空
         if (isTechnicalCozeFailure(row.getError())) {
-            return "待人工复核";
+            return firstNonBlank(row.getError(), "");
         }
         if (isFailedCozeStatusValue(row.getStatus())) {
-            return "待人工复核";
+            return firstNonBlank(row.getError(), "");
         }
         return firstNonBlank(row.getConclusion(), "");
     }
@@ -4070,8 +4084,8 @@ public class AppearancePatentTaskService {
     /**
      * 导出最终 xlsx 时的状态展示：
      *  - 如果 status 是 markFailed 合成的（failureSyntheticStatus=true），
-     *    且 row 实际并没有任何风险维度结果，导出层降级为"待人工复核"，避免用户原样
-     *    重新上传时该行被反复识别为失败行重新触发 Coze。
+     *    且 row 实际并没有任何风险维度结果，导出层展示错误信息（有则放入，无则留空），
+     *    避免用户原样重新上传时该行被反复识别为失败行重新触发 Coze。
      *  - 真正业务侧失败 / Python 真实回传 FAILED 仍按原值显示。
      */
     private String userFacingStatus(AppearancePatentResultRowDto row) {
@@ -4082,7 +4096,8 @@ public class AppearancePatentTaskService {
         if (!row.isFailureSyntheticStatus()) {
             return original;
         }
-        return "待人工复核";
+        // 合成状态：有错误信息则放入错误信息，没有则留空
+        return firstNonBlank(row.getError(), "");
     }
 
     private boolean isTechnicalCozeFailure(String value) {

backend-java/src/main/java/com/nanri/aiimage/modules/collectdata/controller/CollectDataController.java

@@ -0,0 +1,128 @@
+package com.nanri.aiimage.modules.collectdata.controller;
+
+import com.nanri.aiimage.common.api.ApiResponse;
+import com.nanri.aiimage.modules.collectdata.model.dto.CollectDataCountryPreferenceSaveRequest;
+import com.nanri.aiimage.modules.collectdata.model.dto.CollectDataParseRequest;
+import com.nanri.aiimage.modules.collectdata.model.dto.CollectDataTaskBatchRequest;
+import com.nanri.aiimage.modules.collectdata.model.vo.CollectDataCountryPreferenceVo;
+import com.nanri.aiimage.modules.collectdata.model.vo.CollectDataDashboardVo;
+import com.nanri.aiimage.modules.collectdata.model.vo.CollectDataHistoryVo;
+import com.nanri.aiimage.modules.collectdata.model.vo.CollectDataItemsPageVo;
+import com.nanri.aiimage.modules.collectdata.model.vo.CollectDataParseVo;
+import com.nanri.aiimage.modules.collectdata.model.vo.CollectDataTaskBatchVo;
+import com.nanri.aiimage.modules.collectdata.service.CollectDataService;
+import io.swagger.v3.oas.annotations.Operation;
+import io.swagger.v3.oas.annotations.Parameter;
+import io.swagger.v3.oas.annotations.tags.Tag;
+import jakarta.validation.Valid;
+import lombok.RequiredArgsConstructor;
+import org.springframework.web.bind.annotation.DeleteMapping;
+import org.springframework.web.bind.annotation.GetMapping;
+import org.springframework.web.bind.annotation.PathVariable;
+import org.springframework.web.bind.annotation.PostMapping;
+import org.springframework.web.bind.annotation.PutMapping;
+import org.springframework.web.bind.annotation.RequestBody;
+import org.springframework.web.bind.annotation.RequestMapping;
+import org.springframework.web.bind.annotation.RequestParam;
+import org.springframework.web.bind.annotation.RestController;
+
+@RestController
+@RequiredArgsConstructor
+@RequestMapping("/api/collect-data")
+@Tag(name = "采集数据", description = "采集数据任务接口。前端上传 Excel 后由 Java 解析并落库；Python 端按页拉取明细数据进行采集。")
+public class CollectDataController {
+
+    private final CollectDataService service;
+
+    @PostMapping("/parse")
+    @Operation(summary = "解析 Excel 并创建任务", description = "解析上传后的 Excel 文件，按行入库到 biz_collect_data_item，并保存任务筛选条件。任务初始状态为 PENDING。")
+    public ApiResponse<CollectDataParseVo> parse(@Valid @RequestBody CollectDataParseRequest request) {
+        return ApiResponse.success(service.parseAndCreateTask(request));
+    }
+
+    @PostMapping("/tasks/{taskId}/activate")
+    @Operation(summary = "激活任务", description = "前端推送 Python 队列前调用，将任务状态从 PENDING 修改为 RUNNING，表示已交给 Python 处理。")
+    public ApiResponse<Void> activate(
+            @Parameter(description = "采集任务 ID", required = true, example = "9001")
+            @PathVariable Long taskId,
+            @Parameter(description = "当前用户 ID", required = true, example = "1")
+            @RequestParam("user_id") Long userId) {
+        service.activateTask(taskId, userId);
+        return ApiResponse.success(null);
+    }
+
+    @GetMapping("/tasks/{taskId}/items")
+    @Operation(summary = "分页获取任务明细数据", description = "供 Python 端拉取，默认每页 50 条；返回任务关联的筛选条件，便于 Python 端按筛选条件采集。")
+    public ApiResponse<CollectDataItemsPageVo> items(
+            @Parameter(description = "采集任务 ID", required = true, example = "9001")
+            @PathVariable Long taskId,
+            @Parameter(description = "用户 ID，可选；传入时会校验任务归属")
+            @RequestParam(value = "user_id", required = false) Long userId,
+            @Parameter(description = "页码，从 1 开始", example = "1")
+            @RequestParam(value = "page", required = false, defaultValue = "1") Integer page,
+            @Parameter(description = "每页条数，默认 50，最大 200", example = "50")
+            @RequestParam(value = "page_size", required = false, defaultValue = "50") Integer pageSize) {
+        return ApiResponse.success(service.getItemsPage(taskId, userId, page, pageSize));
+    }
+
+    @GetMapping("/dashboard")
+    @Operation(summary = "查询采集任务总览")
+    public ApiResponse<CollectDataDashboardVo> dashboard(
+            @Parameter(description = "当前用户 ID", required = true, example = "1")
+            @RequestParam("user_id") Long userId) {
+        return ApiResponse.success(service.dashboard(userId));
+    }
+
+    @GetMapping("/history")
+    @Operation(summary = "查询采集任务历史")
+    public ApiResponse<CollectDataHistoryVo> history(
+            @Parameter(description = "当前用户 ID", required = true, example = "1")
+            @RequestParam("user_id") Long userId,
+            @Parameter(description = "返回数量上限，默认 50，最大 100", example = "50")
+            @RequestParam(value = "limit", required = false, defaultValue = "50") Integer limit) {
+        return ApiResponse.success(service.history(userId, limit));
+    }
+
+    @PostMapping("/tasks/progress/batch")
+    @Operation(summary = "批量查询任务进度")
+    public ApiResponse<CollectDataTaskBatchVo> progressBatch(@Valid @RequestBody CollectDataTaskBatchRequest request) {
+        return ApiResponse.success(service.progressBatch(request.getTaskIds()));
+    }
+
+    @DeleteMapping("/tasks/{taskId}")
+    @Operation(summary = "删除任务", description = "删除任务及其明细行、关联结果记录。")
+    public ApiResponse<Void> deleteTask(
+            @Parameter(description = "采集任务 ID", required = true, example = "9001")
+            @PathVariable Long taskId,
+            @Parameter(description = "当前用户 ID", required = true, example = "1")
+            @RequestParam("user_id") Long userId) {
+        service.deleteTask(taskId, userId);
+        return ApiResponse.success(null);
+    }
+
+    @DeleteMapping("/history/{resultId}")
+    @Operation(summary = "删除历史记录")
+    public ApiResponse<Void> deleteHistory(
+            @Parameter(description = "结果记录 ID", required = true, example = "1001")
+            @PathVariable Long resultId,
+            @Parameter(description = "当前用户 ID", required = true, example = "1")
+            @RequestParam("user_id") Long userId) {
+        service.deleteHistory(resultId, userId);
+        return ApiResponse.success(null);
+    }
+
+    @GetMapping("/country-preference")
+    @Operation(summary = "查询国家处理顺序", description = "返回当前用户保存的国家代码列表，顺序即处理顺序。")
+    public ApiResponse<CollectDataCountryPreferenceVo> getCountryPreference(
+            @Parameter(description = "当前用户 ID", required = true, example = "1")
+            @RequestParam("user_id") Long userId) {
+        return ApiResponse.success(service.getCountryPreference(userId));
+    }
+
+    @PutMapping("/country-preference")
+    @Operation(summary = "保存国家处理顺序", description = "保存当前用户的国家勾选与处理顺序。")
+    public ApiResponse<CollectDataCountryPreferenceVo> saveCountryPreference(
+            @Valid @RequestBody CollectDataCountryPreferenceSaveRequest request) {
+        return ApiResponse.success(service.saveCountryPreference(request));
+    }
+}

backend-java/src/main/java/com/nanri/aiimage/modules/collectdata/mapper/CollectDataCountryPrefMapper.java

@@ -0,0 +1,9 @@
+package com.nanri.aiimage.modules.collectdata.mapper;
+
+import com.baomidou.mybatisplus.core.mapper.BaseMapper;
+import com.nanri.aiimage.modules.collectdata.model.entity.CollectDataCountryPrefEntity;
+import org.apache.ibatis.annotations.Mapper;
+
+@Mapper
+public interface CollectDataCountryPrefMapper extends BaseMapper<CollectDataCountryPrefEntity> {
+}

backend-java/src/main/java/com/nanri/aiimage/modules/collectdata/mapper/CollectDataItemMapper.java

@@ -0,0 +1,26 @@
+package com.nanri.aiimage.modules.collectdata.mapper;
+
+import com.baomidou.mybatisplus.core.mapper.BaseMapper;
+import com.nanri.aiimage.modules.collectdata.model.entity.CollectDataItemEntity;
+import org.apache.ibatis.annotations.Insert;
+import org.apache.ibatis.annotations.Mapper;
+import org.apache.ibatis.annotations.Param;
+
+import java.util.List;
+
+@Mapper
+public interface CollectDataItemMapper extends BaseMapper<CollectDataItemEntity> {
+
+    @Insert("""
+            <script>
+            INSERT INTO biz_collect_data_item
+                (task_id, row_index, source_file_key, source_filename, keyword, status_value, extra_json, created_at)
+            VALUES
+            <foreach collection="rows" item="row" separator=",">
+                (#{row.taskId}, #{row.rowIndex}, #{row.sourceFileKey}, #{row.sourceFilename},
+                 #{row.keyword}, #{row.statusValue}, #{row.extraJson}, #{row.createdAt})
+            </foreach>
+            </script>
+            """)
+    int insertBatch(@Param("rows") List<CollectDataItemEntity> rows);
+}

backend-java/src/main/java/com/nanri/aiimage/modules/collectdata/model/dto/CollectDataCountryPreferenceSaveRequest.java

@@ -0,0 +1,27 @@
+package com.nanri.aiimage.modules.collectdata.model.dto;
+
+import com.fasterxml.jackson.annotation.JsonProperty;
+import io.swagger.v3.oas.annotations.media.Schema;
+import jakarta.validation.constraints.NotEmpty;
+import jakarta.validation.constraints.NotNull;
+import jakarta.validation.constraints.Size;
+import lombok.Data;
+
+import java.util.ArrayList;
+import java.util.List;
+
+@Data
+@Schema(description = "保存采集数据五国处理顺序：至少 1 个、最多 5 个，代码须为 DE、UK、FR、IT、ES，且不能重复")
+public class CollectDataCountryPreferenceSaveRequest {
+
+    @NotNull(message = "user_id 不能为空")
+    @JsonProperty("user_id")
+    @Schema(description = "当前用户 ID", requiredMode = Schema.RequiredMode.REQUIRED, example = "1")
+    private Long userId;
+
+    @NotEmpty(message = "country_codes 不能为空")
+    @Size(max = 5, message = "country_codes 最多 5 项")
+    @JsonProperty("country_codes")
+    @Schema(description = "国家代码列表（DE、UK、FR、IT、ES），顺序即处理顺序", requiredMode = Schema.RequiredMode.REQUIRED)
+    private List<String> countryCodes = new ArrayList<>();
+}

backend-java/src/main/java/com/nanri/aiimage/modules/collectdata/model/dto/CollectDataFiltersDto.java

@@ -0,0 +1,41 @@
+package com.nanri.aiimage.modules.collectdata.model.dto;
+
+import io.swagger.v3.oas.annotations.media.Schema;
+import com.fasterxml.jackson.annotation.JsonAlias;
+import lombok.Data;
+
+import java.math.BigDecimal;
+import java.util.ArrayList;
+import java.util.List;
+
+@Data
+@Schema(description = "采集数据筛选条件")
+public class CollectDataFiltersDto {
+
+    @Schema(description = "金额（手动输入），可为空", example = "59.99")
+    private BigDecimal amount;
+
+    @JsonAlias({"min_amount"})
+    @Schema(description = "最小金额，可为空", example = "10.00")
+    private BigDecimal minAmount;
+
+    @JsonAlias({"max_amount"})
+    @Schema(description = "最大金额，可为空", example = "59.99")
+    private BigDecimal maxAmount;
+
+    @Schema(description = "是否启用排名条件")
+    private Boolean rank;
+
+    @Schema(description = "是否包含 FBA 商品")
+    private Boolean fba;
+
+    @Schema(description = "是否包含 FBM 商品")
+    private Boolean fbm;
+
+    @Schema(description = "国家代码列表，按用户拖拽顺序保存", example = "[\"DE\",\"UK\"]")
+    private List<String> countryCodes = new ArrayList<>();
+
+    @JsonAlias({"country_code"})
+    @Schema(description = "单选国家代码，采集数据页面当前只选择一个国家", example = "DE")
+    private String countryCode;
+}

backend-java/src/main/java/com/nanri/aiimage/modules/collectdata/model/dto/CollectDataParseRequest.java

@@ -0,0 +1,32 @@
+package com.nanri.aiimage.modules.collectdata.model.dto;
+
+import com.fasterxml.jackson.annotation.JsonProperty;
+import io.swagger.v3.oas.annotations.media.Schema;
+import jakarta.validation.Valid;
+import jakarta.validation.constraints.NotEmpty;
+import jakarta.validation.constraints.NotNull;
+import lombok.Data;
+
+import java.util.List;
+
+@Data
+@Schema(description = "采集数据解析请求：提交筛选条件和上传后的 Excel 文件，由后端解析并落库。")
+public class CollectDataParseRequest {
+
+    @JsonProperty("user_id")
+    @NotNull
+    @Schema(description = "当前用户 ID", example = "1", requiredMode = Schema.RequiredMode.REQUIRED)
+    private Long userId;
+
+    @NotEmpty
+    @Schema(description = "已上传的 Excel 文件列表", requiredMode = Schema.RequiredMode.REQUIRED)
+    private List<CollectDataSourceFileDto> files;
+
+    @JsonProperty("task_type")
+    @Schema(description = "任务类型，前端可自定义传入；为空时默认 collect-data", example = "collect-data")
+    private String taskType;
+
+    @Valid
+    @Schema(description = "筛选条件")
+    private CollectDataFiltersDto filters;
+}

backend-java/src/main/java/com/nanri/aiimage/modules/collectdata/model/dto/CollectDataSourceFileDto.java

@@ -0,0 +1,17 @@
+package com.nanri.aiimage.modules.collectdata.model.dto;
+
+import io.swagger.v3.oas.annotations.media.Schema;
+import lombok.Data;
+
+@Data
+@Schema(description = "采集数据源文件信息")
+public class CollectDataSourceFileDto {
+    @Schema(description = "上传接口返回的临时文件 key", requiredMode = Schema.RequiredMode.REQUIRED)
+    private String fileKey;
+
+    @Schema(description = "原始文件名")
+    private String originalFilename;
+
+    @Schema(description = "相对目录路径")
+    private String relativePath;
+}

backend-java/src/main/java/com/nanri/aiimage/modules/collectdata/model/dto/CollectDataTaskBatchRequest.java

@@ -0,0 +1,15 @@
+package com.nanri.aiimage.modules.collectdata.model.dto;
+
+import io.swagger.v3.oas.annotations.media.Schema;
+import jakarta.validation.constraints.NotEmpty;
+import lombok.Data;
+
+import java.util.List;
+
+@Data
+@Schema(description = "批量查询采集数据任务进度的请求")
+public class CollectDataTaskBatchRequest {
+    @NotEmpty
+    @Schema(description = "任务 ID 列表", requiredMode = Schema.RequiredMode.REQUIRED)
+    private List<Long> taskIds;
+}

backend-java/src/main/java/com/nanri/aiimage/modules/collectdata/model/entity/CollectDataCountryPrefEntity.java

@@ -0,0 +1,22 @@
+package com.nanri.aiimage.modules.collectdata.model.entity;
+
+import com.baomidou.mybatisplus.annotation.IdType;
+import com.baomidou.mybatisplus.annotation.TableField;
+import com.baomidou.mybatisplus.annotation.TableId;
+import com.baomidou.mybatisplus.annotation.TableName;
+import lombok.Data;
+
+import java.time.LocalDateTime;
+
+@Data
+@TableName("biz_collect_data_country_pref")
+public class CollectDataCountryPrefEntity {
+
+    @TableId(type = IdType.INPUT)
+    private Long userId;
+
+    @TableField("country_codes_json")
+    private String countryCodesJson;
+
+    private LocalDateTime updatedAt;
+}

backend-java/src/main/java/com/nanri/aiimage/modules/collectdata/model/entity/CollectDataItemEntity.java

@@ -0,0 +1,24 @@
+package com.nanri.aiimage.modules.collectdata.model.entity;
+
+import com.baomidou.mybatisplus.annotation.IdType;
+import com.baomidou.mybatisplus.annotation.TableId;
+import com.baomidou.mybatisplus.annotation.TableName;
+import lombok.Data;
+
+import java.time.LocalDateTime;
+
+@Data
+@TableName("biz_collect_data_item")
+public class CollectDataItemEntity {
+
+    @TableId(type = IdType.AUTO)
+    private Long id;
+    private Long taskId;
+    private Integer rowIndex;
+    private String sourceFileKey;
+    private String sourceFilename;
+    private String keyword;
+    private String statusValue;
+    private String extraJson;
+    private LocalDateTime createdAt;
+}

backend-java/src/main/java/com/nanri/aiimage/modules/collectdata/model/vo/CollectDataCountryPreferenceVo.java

@@ -0,0 +1,17 @@
+package com.nanri.aiimage.modules.collectdata.model.vo;
+
+import com.fasterxml.jackson.annotation.JsonProperty;
+import io.swagger.v3.oas.annotations.media.Schema;
+import lombok.Data;
+
+import java.util.ArrayList;
+import java.util.List;
+
+@Data
+@Schema(description = "采集数据用户五国处理偏好：按顺序排列的国家代码列表，如 DE、UK")
+public class CollectDataCountryPreferenceVo {
+
+    @JsonProperty("country_codes")
+    @Schema(description = "已选国家代码，顺序即处理顺序；未持久化时返回默认 DE -> UK -> FR -> IT -> ES")
+    private List<String> countryCodes = new ArrayList<>();
+}

backend-java/src/main/java/com/nanri/aiimage/modules/collectdata/model/vo/CollectDataDashboardVo.java

@@ -0,0 +1,17 @@
+package com.nanri.aiimage.modules.collectdata.model.vo;
+
+import io.swagger.v3.oas.annotations.media.Schema;
+import lombok.Data;
+
+@Data
+@Schema(description = "采集数据总览统计")
+public class CollectDataDashboardVo {
+    @Schema(description = "运行中任务数量")
+    private Long pendingTaskCount;
+    @Schema(description = "已结束任务数量")
+    private Long processedTaskCount;
+    @Schema(description = "成功任务数量")
+    private Long successTaskCount;
+    @Schema(description = "失败任务数量")
+    private Long failedTaskCount;
+}

backend-java/src/main/java/com/nanri/aiimage/modules/collectdata/model/vo/CollectDataHistoryItemVo.java

@@ -0,0 +1,54 @@
+package com.nanri.aiimage.modules.collectdata.model.vo;
+
+import com.nanri.aiimage.modules.collectdata.model.dto.CollectDataFiltersDto;
+import io.swagger.v3.oas.annotations.media.Schema;
+import lombok.Data;
+
+@Data
+@Schema(description = "采集数据历史/任务记录")
+public class CollectDataHistoryItemVo {
+    @Schema(description = "结果记录 ID")
+    private Long resultId;
+
+    @Schema(description = "任务 ID")
+    private Long taskId;
+
+    @Schema(description = "任务编号")
+    private String taskNo;
+
+    @Schema(description = "源文件聚合展示名")
+    private String sourceFilename;
+
+    @Schema(description = "结果文件名")
+    private String resultFilename;
+
+    @Schema(description = "下载地址")
+    private String downloadUrl;
+
+    @Schema(description = "任务状态：PENDING / RUNNING / SUCCESS / FAILED")
+    private String taskStatus;
+
+    @Schema(description = "结果是否成功")
+    private Boolean success;
+
+    @Schema(description = "错误信息")
+    private String error;
+
+    @Schema(description = "落库行数")
+    private Integer rowCount;
+
+    @Schema(description = "记录创建时间")
+    private String createdAt;
+
+    @Schema(description = "任务开始时间")
+    private String startedAt;
+
+    @Schema(description = "任务结束时间")
+    private String finishedAt;
+
+    @Schema(description = "任务类型")
+    private String taskType;
+
+    @Schema(description = "解析时保存的筛选条件")
+    private CollectDataFiltersDto filters;
+}

backend-java/src/main/java/com/nanri/aiimage/modules/collectdata/model/vo/CollectDataHistoryVo.java

@@ -0,0 +1,14 @@
+package com.nanri.aiimage.modules.collectdata.model.vo;
+
+import io.swagger.v3.oas.annotations.media.Schema;
+import lombok.Data;
+
+import java.util.ArrayList;
+import java.util.List;
+
+@Data
+@Schema(description = "采集数据历史列表")
+public class CollectDataHistoryVo {
+    @Schema(description = "历史记录列表")
+    private List<CollectDataHistoryItemVo> items = new ArrayList<>();
+}

backend-java/src/main/java/com/nanri/aiimage/modules/collectdata/model/vo/CollectDataItemVo.java

@@ -0,0 +1,32 @@
+package com.nanri.aiimage.modules.collectdata.model.vo;
+
+import io.swagger.v3.oas.annotations.media.Schema;
+import lombok.Data;
+
+import java.util.LinkedHashMap;
+import java.util.Map;
+
+@Data
+@Schema(description = "采集数据明细行")
+public class CollectDataItemVo {
+    @Schema(description = "数据库主键")
+    private Long id;
+
+    @Schema(description = "行号，从 1 开始")
+    private Integer rowIndex;
+
+    @Schema(description = "来源文件 key")
+    private String sourceFileKey;
+
+    @Schema(description = "来源文件名")
+    private String sourceFilename;
+
+    @Schema(description = "关键词列内容")
+    private String keyword;
+
+    @Schema(description = "状态列内容")
+    private String statusValue;
+
+    @Schema(description = "其它列内容（列名 -> 字符串值）")
+    private Map<String, String> extra = new LinkedHashMap<>();
+}

backend-java/src/main/java/com/nanri/aiimage/modules/collectdata/model/vo/CollectDataItemsPageVo.java

@@ -0,0 +1,45 @@
+package com.nanri.aiimage.modules.collectdata.model.vo;
+
+import com.nanri.aiimage.modules.collectdata.model.dto.CollectDataFiltersDto;
+import io.swagger.v3.oas.annotations.media.Schema;
+import lombok.Data;
+
+import java.util.ArrayList;
+import java.util.List;
+
+@Data
+@Schema(description = "采集数据明细分页结果，默认每页 50 条，供 Python 拉取")
+public class CollectDataItemsPageVo {
+    @Schema(description = "任务 ID")
+    private Long taskId;
+
+    @Schema(description = "任务编号")
+    private String taskNo;
+
+    @Schema(description = "任务类型")
+    private String taskType;
+
+    @Schema(description = "任务状态")
+    private String taskStatus;
+
+    @Schema(description = "页码，从 1 开始")
+    private Integer page;
+
+    @Schema(description = "每页条数")
+    private Integer pageSize;
+
+    @Schema(description = "本页条数")
+    private Integer count;
+
+    @Schema(description = "总条数")
+    private Long total;
+
+    @Schema(description = "总页数")
+    private Integer totalPages;
+
+    @Schema(description = "本任务关联的筛选条件")
+    private CollectDataFiltersDto filters;
+
+    @Schema(description = "明细列表")
+    private List<CollectDataItemVo> items = new ArrayList<>();
+}

backend-java/src/main/java/com/nanri/aiimage/modules/collectdata/model/vo/CollectDataParseVo.java

@@ -0,0 +1,33 @@
+package com.nanri.aiimage.modules.collectdata.model.vo;
+
+import io.swagger.v3.oas.annotations.media.Schema;
+import lombok.Data;
+
+@Data
+@Schema(description = "采集数据解析结果")
+public class CollectDataParseVo {
+
+    @Schema(description = "新创建的任务 ID", example = "9001")
+    private Long taskId;
+
+    @Schema(description = "任务编号")
+    private String taskNo;
+
+    @Schema(description = "聚合源文件展示名", example = "采集前数据.xlsx 等 1 个文件")
+    private String sourceFilename;
+
+    @Schema(description = "源文件数量")
+    private Integer sourceFileCount;
+
+    @Schema(description = "Excel 中检测到的非空数据总行数")
+    private Integer totalRows;
+
+    @Schema(description = "落库的有效行数")
+    private Integer acceptedRows;
+
+    @Schema(description = "因缺少必要字段被丢弃的行数")
+    private Integer droppedRows;
+
+    @Schema(description = "推荐分页大小，便于 Python 拉取分页数据", example = "50")
+    private Integer pageSize;
+}

backend-java/src/main/java/com/nanri/aiimage/modules/collectdata/model/vo/CollectDataTaskBatchVo.java

@@ -0,0 +1,17 @@
+package com.nanri.aiimage.modules.collectdata.model.vo;
+
+import io.swagger.v3.oas.annotations.media.Schema;
+import lombok.Data;
+
+import java.util.ArrayList;
+import java.util.List;
+
+@Data
+@Schema(description = "批量任务进度返回")
+public class CollectDataTaskBatchVo {
+    @Schema(description = "任务详情列表")
+    private List<CollectDataTaskDetailVo> items = new ArrayList<>();
+
+    @Schema(description = "未找到的任务 ID")
+    private List<Long> missingTaskIds = new ArrayList<>();
+}

backend-java/src/main/java/com/nanri/aiimage/modules/collectdata/model/vo/CollectDataTaskDetailVo.java

@@ -0,0 +1,17 @@
+package com.nanri.aiimage.modules.collectdata.model.vo;
+
+import io.swagger.v3.oas.annotations.media.Schema;
+import lombok.Data;
+
+import java.util.ArrayList;
+import java.util.List;
+
+@Data
+@Schema(description = "采集数据任务进度详情")
+public class CollectDataTaskDetailVo {
+    @Schema(description = "任务摘要")
+    private CollectDataTaskSummaryVo task;
+
+    @Schema(description = "结果/历史记录")
+    private List<CollectDataHistoryItemVo> items = new ArrayList<>();
+}

backend-java/src/main/java/com/nanri/aiimage/modules/collectdata/model/vo/CollectDataTaskSummaryVo.java

@@ -0,0 +1,29 @@
+package com.nanri.aiimage.modules.collectdata.model.vo;
+
+import io.swagger.v3.oas.annotations.media.Schema;
+import lombok.Data;
+
+@Data
+@Schema(description = "采集数据任务摘要")
+public class CollectDataTaskSummaryVo {
+    @Schema(description = "任务 ID")
+    private Long id;
+
+    @Schema(description = "任务编号")
+    private String taskNo;
+
+    @Schema(description = "任务状态")
+    private String status;
+
+    @Schema(description = "错误信息")
+    private String errorMessage;
+
+    @Schema(description = "创建时间")
+    private String createdAt;
+
+    @Schema(description = "更新时间")
+    private String updatedAt;
+
+    @Schema(description = "结束时间")
+    private String finishedAt;
+}

backend-java/src/main/java/com/nanri/aiimage/modules/collectdata/service/CollectDataService.java

@@ -0,0 +1,732 @@
+package com.nanri.aiimage.modules.collectdata.service;
+
+import cn.hutool.core.util.IdUtil;
+import com.baomidou.mybatisplus.core.conditions.query.LambdaQueryWrapper;
+import com.fasterxml.jackson.core.type.TypeReference;
+import com.fasterxml.jackson.databind.JsonNode;
+import com.fasterxml.jackson.databind.ObjectMapper;
+import com.nanri.aiimage.common.exception.BusinessException;
+import com.nanri.aiimage.modules.collectdata.mapper.CollectDataCountryPrefMapper;
+import com.nanri.aiimage.modules.collectdata.mapper.CollectDataItemMapper;
+import com.nanri.aiimage.modules.collectdata.model.dto.CollectDataCountryPreferenceSaveRequest;
+import com.nanri.aiimage.modules.collectdata.model.dto.CollectDataFiltersDto;
+import com.nanri.aiimage.modules.collectdata.model.dto.CollectDataParseRequest;
+import com.nanri.aiimage.modules.collectdata.model.dto.CollectDataSourceFileDto;
+import com.nanri.aiimage.modules.collectdata.model.entity.CollectDataCountryPrefEntity;
+import com.nanri.aiimage.modules.collectdata.model.entity.CollectDataItemEntity;
+import com.nanri.aiimage.modules.collectdata.model.vo.CollectDataCountryPreferenceVo;
+import com.nanri.aiimage.modules.collectdata.model.vo.CollectDataDashboardVo;
+import com.nanri.aiimage.modules.collectdata.model.vo.CollectDataHistoryItemVo;
+import com.nanri.aiimage.modules.collectdata.model.vo.CollectDataHistoryVo;
+import com.nanri.aiimage.modules.collectdata.model.vo.CollectDataItemVo;
+import com.nanri.aiimage.modules.collectdata.model.vo.CollectDataItemsPageVo;
+import com.nanri.aiimage.modules.collectdata.model.vo.CollectDataParseVo;
+import com.nanri.aiimage.modules.collectdata.model.vo.CollectDataTaskBatchVo;
+import com.nanri.aiimage.modules.collectdata.model.vo.CollectDataTaskDetailVo;
+import com.nanri.aiimage.modules.collectdata.model.vo.CollectDataTaskSummaryVo;
+import com.nanri.aiimage.modules.file.service.LocalFileStorageService;
+import com.nanri.aiimage.modules.task.mapper.FileResultMapper;
+import com.nanri.aiimage.modules.task.mapper.FileTaskMapper;
+import com.nanri.aiimage.modules.task.model.entity.FileResultEntity;
+import com.nanri.aiimage.modules.task.model.entity.FileTaskEntity;
+import lombok.RequiredArgsConstructor;
+import lombok.extern.slf4j.Slf4j;
+import org.apache.poi.ss.usermodel.Cell;
+import org.apache.poi.ss.usermodel.DataFormatter;
+import org.apache.poi.ss.usermodel.Row;
+import org.apache.poi.ss.usermodel.Sheet;
+import org.apache.poi.ss.usermodel.Workbook;
+import org.apache.poi.ss.usermodel.WorkbookFactory;
+import org.springframework.stereotype.Service;
+import org.springframework.transaction.annotation.Transactional;
+import org.springframework.transaction.support.TransactionTemplate;
+
+import java.io.File;
+import java.io.FileInputStream;
+import java.time.LocalDateTime;
+import java.time.format.DateTimeFormatter;
+import java.util.ArrayList;
+import java.util.LinkedHashMap;
+import java.util.LinkedHashSet;
+import java.util.List;
+import java.util.Locale;
+import java.util.Map;
+import java.util.Objects;
+import java.util.regex.Pattern;
+
+@Service
+@RequiredArgsConstructor
+@Slf4j
+public class CollectDataService {
+
+    public static final String MODULE_TYPE = "COLLECT_DATA";
+    public static final int DEFAULT_PAGE_SIZE = 50;
+
+    /**
+     * 默认国家偏好顺序：DE -> UK -> FR -> IT -> ES，与前端保持一致。
+     */
+    public static final List<String> DEFAULT_COUNTRY_PREFERENCE_ORDER = List.of("DE", "UK", "FR", "IT", "ES");
+
+    private static final String STATUS_PENDING = "PENDING";
+    private static final String STATUS_RUNNING = "RUNNING";
+    private static final String STATUS_SUCCESS = "SUCCESS";
+    private static final String STATUS_FAILED = "FAILED";
+
+    private static final String DEFAULT_TASK_TYPE = "collect-data";
+    private static final int ITEM_INSERT_BATCH_SIZE = 500;
+
+    private static final List<String> KEYWORD_HEADER_ALIASES = List.of("关键词", "keyword", "key word");
+    private static final List<String> STATUS_HEADER_ALIASES = List.of("状态", "status");
+
+    private static final DateTimeFormatter TIME_FORMATTER = DateTimeFormatter.ofPattern("yyyy-MM-dd HH:mm:ss");
+    private static final Pattern WHITESPACE_PATTERN = Pattern.compile("\\s+");
+
+    private final LocalFileStorageService localFileStorageService;
+    private final FileTaskMapper fileTaskMapper;
+    private final FileResultMapper fileResultMapper;
+    private final CollectDataItemMapper collectDataItemMapper;
+    private final CollectDataCountryPrefMapper collectDataCountryPrefMapper;
+    private final ObjectMapper objectMapper;
+    private final TransactionTemplate transactionTemplate;
+
+    public CollectDataParseVo parseAndCreateTask(CollectDataParseRequest request) {
+        long startedAt = System.currentTimeMillis();
+        if (request == null || request.getUserId() == null || request.getUserId() <= 0) {
+            throw new BusinessException("user_id 不合法");
+        }
+        if (request.getFiles() == null || request.getFiles().isEmpty()) {
+            throw new BusinessException("请先上传 Excel 文件");
+        }
+        List<CollectDataSourceFileDto> sources = request.getFiles().stream()
+                .filter(Objects::nonNull)
+                .filter(f -> f.getFileKey() != null && !f.getFileKey().isBlank())
+                .toList();
+        if (sources.isEmpty()) {
+            throw new BusinessException("请先上传 Excel 文件");
+        }
+        String requestedTaskType = normalize(request.getTaskType());
+        final String taskType = requestedTaskType.isBlank() ? DEFAULT_TASK_TYPE : requestedTaskType;
+        CollectDataFiltersDto filters = request.getFilters() == null ? new CollectDataFiltersDto() : request.getFilters();
+
+        List<ParsedRow> parsedRows = new ArrayList<>();
+        int totalRows = 0;
+        int droppedRows = 0;
+        for (CollectDataSourceFileDto source : sources) {
+            File input = localFileStorageService.findLocalSourceFile(source.getFileKey());
+            if (input == null || !input.exists()) {
+                throw new BusinessException("源文件不存在");
+            }
+            ParsedWorkbook parsed = parseWorkbook(input, source);
+            totalRows += parsed.totalRows();
+            droppedRows += parsed.droppedRows();
+            parsedRows.addAll(parsed.rows());
+        }
+        if (parsedRows.isEmpty()) {
+            throw new BusinessException("未解析到有效数据行");
+        }
+
+        long parsedAt = System.currentTimeMillis();
+        String aggregateFilename = buildAggregateSourceFilenameLabel(sources);
+        PersistedParse persisted = transactionTemplate.execute(status ->
+                persistParsedTask(request, sources, taskType, filters, parsedRows, aggregateFilename));
+        if (persisted == null) {
+            throw new BusinessException("创建采集任务失败");
+        }
+        long persistedAt = System.currentTimeMillis();
+        log.info("[collect-data] parseAndCreateTask userId={} files={} totalRows={} acceptedRows={} droppedRows={} parseMs={} persistMs={} totalMs={}",
+                request.getUserId(), sources.size(), totalRows, parsedRows.size(), droppedRows,
+                parsedAt - startedAt, persistedAt - parsedAt, persistedAt - startedAt);
+
+        CollectDataParseVo vo = new CollectDataParseVo();
+        vo.setTaskId(persisted.task().getId());
+        vo.setTaskNo(persisted.task().getTaskNo());
+        vo.setSourceFilename(aggregateFilename);
+        vo.setSourceFileCount(sources.size());
+        vo.setTotalRows(totalRows);
+        vo.setAcceptedRows(parsedRows.size());
+        vo.setDroppedRows(droppedRows);
+        vo.setPageSize(DEFAULT_PAGE_SIZE);
+        return vo;
+    }
+
+    private PersistedParse persistParsedTask(CollectDataParseRequest request,
+                                            List<CollectDataSourceFileDto> sources,
+                                            String taskType,
+                                            CollectDataFiltersDto filters,
+                                            List<ParsedRow> parsedRows,
+                                            String aggregateFilename) {
+        FileTaskEntity task = new FileTaskEntity();
+        task.setTaskNo(MODULE_TYPE + "-" + IdUtil.getSnowflakeNextIdStr());
+        task.setModuleType(MODULE_TYPE);
+        task.setTaskMode("PYTHON_QUEUE");
+        task.setStatus(STATUS_PENDING);
+        task.setSourceFileCount(sources.size());
+        task.setSuccessFileCount(0);
+        task.setFailedFileCount(0);
+        task.setCreatedBy("user:" + request.getUserId());
+        task.setUserId(request.getUserId());
+        task.setCreatedAt(LocalDateTime.now());
+        task.setUpdatedAt(LocalDateTime.now());
+        try {
+            Map<String, Object> requestPayload = new LinkedHashMap<>();
+            requestPayload.put("userId", request.getUserId());
+            requestPayload.put("taskType", taskType);
+            requestPayload.put("filters", filters);
+            requestPayload.put("files", sources);
+            task.setRequestJson(objectMapper.writeValueAsString(requestPayload));
+            task.setResultJson("{}");
+        } catch (Exception ex) {
+            throw new BusinessException("序列化任务信息失败");
+        }
+        fileTaskMapper.insert(task);
+
+        int rowIndex = 0;
+        LocalDateTime itemCreatedAt = LocalDateTime.now();
+        List<CollectDataItemEntity> itemBatch = new ArrayList<>(Math.min(parsedRows.size(), ITEM_INSERT_BATCH_SIZE));
+        for (ParsedRow parsedRow : parsedRows) {
+            rowIndex++;
+            CollectDataItemEntity entity = new CollectDataItemEntity();
+            entity.setTaskId(task.getId());
+            entity.setRowIndex(rowIndex);
+            entity.setSourceFileKey(parsedRow.sourceFileKey());
+            entity.setSourceFilename(parsedRow.sourceFilename());
+            entity.setKeyword(parsedRow.keyword());
+            entity.setStatusValue(parsedRow.statusValue());
+            try {
+                entity.setExtraJson(objectMapper.writeValueAsString(parsedRow.extra()));
+            } catch (Exception ex) {
+                entity.setExtraJson("{}");
+            }
+            entity.setCreatedAt(itemCreatedAt);
+            itemBatch.add(entity);
+            if (itemBatch.size() >= ITEM_INSERT_BATCH_SIZE) {
+                collectDataItemMapper.insertBatch(itemBatch);
+                itemBatch.clear();
+            }
+        }
+        if (!itemBatch.isEmpty()) {
+            collectDataItemMapper.insertBatch(itemBatch);
+        }
+
+        FileResultEntity result = new FileResultEntity();
+        result.setTaskId(task.getId());
+        result.setModuleType(MODULE_TYPE);
+        result.setSourceFilename(aggregateFilename);
+        result.setSourceFileUrl(buildAggregateSourceFileUrl(sources));
+        result.setRowCount(parsedRows.size());
+        result.setUserId(request.getUserId());
+        result.setCreatedAt(LocalDateTime.now());
+        fileResultMapper.insert(result);
+
+        return new PersistedParse(task, result);
+    }
+
+    @Transactional
+    public void activateTask(Long taskId, Long userId) {
+        FileTaskEntity task = requireTask(taskId, userId);
+        if (STATUS_SUCCESS.equals(task.getStatus()) || STATUS_FAILED.equals(task.getStatus())) {
+            throw new BusinessException("任务已结束");
+        }
+        task.setStatus(STATUS_RUNNING);
+        task.setUpdatedAt(LocalDateTime.now());
+        fileTaskMapper.updateById(task);
+    }
+
+    public CollectDataDashboardVo dashboard(Long userId) {
+        CollectDataDashboardVo vo = new CollectDataDashboardVo();
+        vo.setPendingTaskCount(countTask(userId, STATUS_RUNNING));
+        vo.setSuccessTaskCount(countTask(userId, STATUS_SUCCESS));
+        vo.setFailedTaskCount(countTask(userId, STATUS_FAILED));
+        long processed = (vo.getSuccessTaskCount() == null ? 0 : vo.getSuccessTaskCount())
+                + (vo.getFailedTaskCount() == null ? 0 : vo.getFailedTaskCount());
+        vo.setProcessedTaskCount(processed);
+        return vo;
+    }
+
+    public CollectDataHistoryVo history(Long userId, Integer limit) {
+        CollectDataHistoryVo vo = new CollectDataHistoryVo();
+        if (userId == null || userId <= 0) {
+            return vo;
+        }
+        int safeLimit = Math.max(1, Math.min(limit == null ? 50 : limit, 100));
+        List<FileResultEntity> rows = fileResultMapper.selectList(new LambdaQueryWrapper<FileResultEntity>()
+                .eq(FileResultEntity::getModuleType, MODULE_TYPE)
+                .eq(FileResultEntity::getUserId, userId)
+                .orderByDesc(FileResultEntity::getCreatedAt)
+                .last("limit " + safeLimit));
+        if (rows == null || rows.isEmpty()) {
+            return vo;
+        }
+        List<Long> taskIds = rows.stream().map(FileResultEntity::getTaskId).filter(Objects::nonNull).distinct().toList();
+        Map<Long, FileTaskEntity> taskMap = loadTaskMap(taskIds);
+        for (FileResultEntity row : rows) {
+            FileTaskEntity task = row.getTaskId() == null ? null : taskMap.get(row.getTaskId());
+            vo.getItems().add(toHistoryItem(row, task));
+        }
+        return vo;
+    }
+
+    public CollectDataTaskBatchVo progressBatch(List<Long> taskIds) {
+        CollectDataTaskBatchVo vo = new CollectDataTaskBatchVo();
+        List<Long> normalizedIds = taskIds == null ? List.of() : taskIds.stream()
+                .filter(id -> id != null && id > 0)
+                .distinct()
+                .toList();
+        if (normalizedIds.isEmpty()) {
+            return vo;
+        }
+        Map<Long, FileTaskEntity> taskMap = loadTaskMap(normalizedIds);
+        List<FileResultEntity> resultRows = fileResultMapper.selectList(new LambdaQueryWrapper<FileResultEntity>()
+                .eq(FileResultEntity::getModuleType, MODULE_TYPE)
+                .in(FileResultEntity::getTaskId, normalizedIds)
+                .orderByDesc(FileResultEntity::getCreatedAt));
+        Map<Long, FileResultEntity> resultByTaskId = new LinkedHashMap<>();
+        if (resultRows != null) {
+            for (FileResultEntity row : resultRows) {
+                if (row.getTaskId() != null) {
+                    resultByTaskId.putIfAbsent(row.getTaskId(), row);
+                }
+            }
+        }
+        for (Long taskId : normalizedIds) {
+            FileTaskEntity task = taskMap.get(taskId);
+            if (task == null) {
+                vo.getMissingTaskIds().add(taskId);
+                continue;
+            }
+            CollectDataTaskDetailVo detail = new CollectDataTaskDetailVo();
+            detail.setTask(toTaskSummary(task));
+            FileResultEntity row = resultByTaskId.get(taskId);
+            if (row != null) {
+                detail.getItems().add(toHistoryItem(row, task));
+            }
+            vo.getItems().add(detail);
+        }
+        return vo;
+    }
+
+    public CollectDataItemsPageVo getItemsPage(Long taskId, Long userId, Integer page, Integer pageSize) {
+        FileTaskEntity task = requireTask(taskId, userId);
+        int safePage = page == null || page <= 0 ? 1 : page;
+        int safePageSize = pageSize == null || pageSize <= 0 ? DEFAULT_PAGE_SIZE : Math.min(pageSize, 200);
+
+        Long total = collectDataItemMapper.selectCount(new LambdaQueryWrapper<CollectDataItemEntity>()
+                .eq(CollectDataItemEntity::getTaskId, taskId));
+        long totalRows = total == null ? 0L : total;
+        int totalPages = totalRows == 0 ? 0 : (int) ((totalRows + safePageSize - 1) / safePageSize);
+
+        int offset = (safePage - 1) * safePageSize;
+        List<CollectDataItemEntity> rows = collectDataItemMapper.selectList(new LambdaQueryWrapper<CollectDataItemEntity>()
+                .eq(CollectDataItemEntity::getTaskId, taskId)
+                .orderByAsc(CollectDataItemEntity::getRowIndex)
+                .last("limit " + safePageSize + " offset " + Math.max(offset, 0)));
+
+        CollectDataItemsPageVo vo = new CollectDataItemsPageVo();
+        vo.setTaskId(task.getId());
+        vo.setTaskNo(task.getTaskNo());
+        vo.setTaskStatus(task.getStatus());
+        vo.setPage(safePage);
+        vo.setPageSize(safePageSize);
+        vo.setTotal(totalRows);
+        vo.setTotalPages(totalPages);
+        vo.setFilters(extractFiltersFromRequest(task));
+        vo.setTaskType(extractTaskTypeFromRequest(task));
+
+        List<CollectDataItemVo> items = new ArrayList<>();
+        if (rows != null) {
+            for (CollectDataItemEntity row : rows) {
+                items.add(toItemVo(row));
+            }
+        }
+        vo.setItems(items);
+        vo.setCount(items.size());
+        return vo;
+    }
+
+    @Transactional
+    public void deleteTask(Long taskId, Long userId) {
+        FileTaskEntity task = requireTask(taskId, userId);
+        collectDataItemMapper.delete(new LambdaQueryWrapper<CollectDataItemEntity>()
+                .eq(CollectDataItemEntity::getTaskId, task.getId()));
+        fileResultMapper.delete(new LambdaQueryWrapper<FileResultEntity>()
+                .eq(FileResultEntity::getTaskId, task.getId())
+                .eq(FileResultEntity::getModuleType, MODULE_TYPE));
+        fileTaskMapper.deleteById(task.getId());
+    }
+
+    public void deleteHistory(Long resultId, Long userId) {
+        FileResultEntity row = fileResultMapper.selectById(resultId);
+        if (row == null || !MODULE_TYPE.equals(row.getModuleType()) || !Objects.equals(userId, row.getUserId())) {
+            throw new BusinessException("记录不存在");
+        }
+        fileResultMapper.deleteById(resultId);
+    }
+
+    public CollectDataCountryPreferenceVo getCountryPreference(Long userId) {
+        if (userId == null || userId <= 0) {
+            throw new BusinessException("user_id 不合法");
+        }
+        CollectDataCountryPrefEntity row = collectDataCountryPrefMapper.selectById(userId);
+        CollectDataCountryPreferenceVo vo = new CollectDataCountryPreferenceVo();
+        if (row == null || row.getCountryCodesJson() == null || row.getCountryCodesJson().isBlank()) {
+            vo.getCountryCodes().addAll(DEFAULT_COUNTRY_PREFERENCE_ORDER);
+            return vo;
+        }
+        try {
+            List<String> parsed = objectMapper.readValue(row.getCountryCodesJson(), new TypeReference<List<String>>() {
+            });
+            vo.getCountryCodes().addAll(sanitizeStoredCountryCodes(parsed));
+        } catch (Exception ex) {
+            vo.getCountryCodes().addAll(DEFAULT_COUNTRY_PREFERENCE_ORDER);
+        }
+        return vo;
+    }
+
+    @Transactional
+    public CollectDataCountryPreferenceVo saveCountryPreference(CollectDataCountryPreferenceSaveRequest request) {
+        if (request == null || request.getUserId() == null || request.getUserId() <= 0) {
+            throw new BusinessException("user_id 不合法");
+        }
+        List<String> normalized = validateCountryCodesForSave(request.getCountryCodes());
+        String json;
+        try {
+            json = objectMapper.writeValueAsString(normalized);
+        } catch (Exception ex) {
+            throw new BusinessException("保存国家偏好失败");
+        }
+        CollectDataCountryPrefEntity row = collectDataCountryPrefMapper.selectById(request.getUserId());
+        if (row == null) {
+            row = new CollectDataCountryPrefEntity();
+            row.setUserId(request.getUserId());
+            row.setCountryCodesJson(json);
+            collectDataCountryPrefMapper.insert(row);
+        } else {
+            row.setCountryCodesJson(json);
+            collectDataCountryPrefMapper.updateById(row);
+        }
+        CollectDataCountryPreferenceVo vo = new CollectDataCountryPreferenceVo();
+        vo.getCountryCodes().addAll(normalized);
+        return vo;
+    }
+
+    private static List<String> sanitizeStoredCountryCodes(List<String> raw) {
+        if (raw == null || raw.isEmpty()) {
+            return new ArrayList<>(DEFAULT_COUNTRY_PREFERENCE_ORDER);
+        }
+        LinkedHashSet<String> seen = new LinkedHashSet<>();
+        List<String> out = new ArrayList<>();
+        for (String code : raw) {
+            if (code == null || code.isBlank()) {
+                continue;
+            }
+            String upper = code.trim().toUpperCase(Locale.ROOT);
+            if (!DEFAULT_COUNTRY_PREFERENCE_ORDER.contains(upper)) {
+                continue;
+            }
+            if (seen.add(upper)) {
+                out.add(upper);
+            }
+        }
+        return out.isEmpty() ? new ArrayList<>(DEFAULT_COUNTRY_PREFERENCE_ORDER) : out;
+    }
+
+    private static List<String> validateCountryCodesForSave(List<String> raw) {
+        if (raw == null || raw.isEmpty()) {
+            throw new BusinessException("country_codes 至少选择 1 个国家");
+        }
+        LinkedHashSet<String> seen = new LinkedHashSet<>();
+        List<String> out = new ArrayList<>();
+        for (String code : raw) {
+            if (code == null || code.isBlank()) {
+                throw new BusinessException("country_codes 含空项");
+            }
+            String upper = code.trim().toUpperCase(Locale.ROOT);
+            if (!DEFAULT_COUNTRY_PREFERENCE_ORDER.contains(upper)) {
+                throw new BusinessException("非法国家代码: " + code);
+            }
+            if (!seen.add(upper)) {
+                throw new BusinessException("country_codes 存在重复: " + upper);
+            }
+            out.add(upper);
+        }
+        if (out.size() > DEFAULT_COUNTRY_PREFERENCE_ORDER.size()) {
+            throw new BusinessException("country_codes 最多 5 项");
+        }
+        return out;
+    }
+
+    private FileTaskEntity requireTask(Long taskId, Long userId) {
+        if (taskId == null || taskId <= 0) {
+            throw new BusinessException("任务不存在");
+        }
+        FileTaskEntity task = fileTaskMapper.selectById(taskId);
+        if (task == null || !MODULE_TYPE.equals(task.getModuleType())) {
+            throw new BusinessException("任务不存在");
+        }
+        if (userId != null && userId > 0 && !Objects.equals(userId, task.getUserId())) {
+            throw new BusinessException("任务不存在");
+        }
+        return task;
+    }
+
+    private long countTask(Long userId, String status) {
+        if (userId == null || userId <= 0) {
+            return 0L;
+        }
+        Long count = fileTaskMapper.selectCount(new LambdaQueryWrapper<FileTaskEntity>()
+                .eq(FileTaskEntity::getModuleType, MODULE_TYPE)
+                .eq(FileTaskEntity::getUserId, userId)
+                .eq(FileTaskEntity::getStatus, status));
+        return count == null ? 0L : count;
+    }
+
+    private Map<Long, FileTaskEntity> loadTaskMap(List<Long> taskIds) {
+        Map<Long, FileTaskEntity> taskMap = new LinkedHashMap<>();
+        if (taskIds == null || taskIds.isEmpty()) {
+            return taskMap;
+        }
+        List<FileTaskEntity> tasks = fileTaskMapper.selectList(new LambdaQueryWrapper<FileTaskEntity>()
+                .in(FileTaskEntity::getId, taskIds));
+        if (tasks != null) {
+            for (FileTaskEntity task : tasks) {
+                if (task != null && MODULE_TYPE.equals(task.getModuleType())) {
+                    taskMap.put(task.getId(), task);
+                }
+            }
+        }
+        return taskMap;
+    }
+
+    private CollectDataHistoryItemVo toHistoryItem(FileResultEntity row, FileTaskEntity task) {
+        CollectDataHistoryItemVo vo = new CollectDataHistoryItemVo();
+        vo.setResultId(row.getId());
+        vo.setTaskId(row.getTaskId());
+        vo.setSourceFilename(row.getSourceFilename());
+        vo.setResultFilename(row.getResultFilename());
+        vo.setRowCount(row.getRowCount());
+        vo.setSuccess(row.getSuccess() != null && row.getSuccess() == 1);
+        vo.setError(row.getErrorMessage());
+        vo.setCreatedAt(formatTime(row.getCreatedAt()));
+        if (task != null) {
+            vo.setTaskNo(task.getTaskNo());
+            vo.setTaskStatus(task.getStatus());
+            vo.setStartedAt(formatTime(task.getCreatedAt()));
+            vo.setFinishedAt(formatTime(task.getFinishedAt()));
+            vo.setTaskType(extractTaskTypeFromRequest(task));
+            vo.setFilters(extractFiltersFromRequest(task));
+        }
+        return vo;
+    }
+
+    private CollectDataTaskSummaryVo toTaskSummary(FileTaskEntity task) {
+        CollectDataTaskSummaryVo vo = new CollectDataTaskSummaryVo();
+        vo.setId(task.getId());
+        vo.setTaskNo(task.getTaskNo());
+        vo.setStatus(task.getStatus());
+        vo.setErrorMessage(task.getErrorMessage());
+        vo.setCreatedAt(formatTime(task.getCreatedAt()));
+        vo.setUpdatedAt(formatTime(task.getUpdatedAt()));
+        vo.setFinishedAt(formatTime(task.getFinishedAt()));
+        return vo;
+    }
+
+    private CollectDataItemVo toItemVo(CollectDataItemEntity row) {
+        CollectDataItemVo vo = new CollectDataItemVo();
+        vo.setId(row.getId());
+        vo.setRowIndex(row.getRowIndex());
+        vo.setSourceFileKey(row.getSourceFileKey());
+        vo.setSourceFilename(row.getSourceFilename());
+        vo.setKeyword(row.getKeyword());
+        vo.setStatusValue(row.getStatusValue());
+        if (row.getExtraJson() != null && !row.getExtraJson().isBlank()) {
+            try {
+                Map<String, String> extra = objectMapper.readValue(row.getExtraJson(), new TypeReference<>() {
+                });
+                if (extra != null) {
+                    vo.getExtra().putAll(extra);
+                }
+            } catch (Exception ex) {
+                log.warn("[collect-data] parse extra json failed rowId={} err={}", row.getId(), ex.getMessage());
+            }
+        }
+        return vo;
+    }
+
+    private CollectDataFiltersDto extractFiltersFromRequest(FileTaskEntity task) {
+        if (task == null || task.getRequestJson() == null || task.getRequestJson().isBlank()) {
+            return new CollectDataFiltersDto();
+        }
+        try {
+            JsonNode root = objectMapper.readTree(task.getRequestJson());
+            JsonNode filtersNode = root.get("filters");
+            if (filtersNode == null || filtersNode.isNull()) {
+                return new CollectDataFiltersDto();
+            }
+            return objectMapper.treeToValue(filtersNode, CollectDataFiltersDto.class);
+        } catch (Exception ex) {
+            log.warn("[collect-data] parse request filters failed taskId={} err={}", task.getId(), ex.getMessage());
+            return new CollectDataFiltersDto();
+        }
+    }
+
+    private String extractTaskTypeFromRequest(FileTaskEntity task) {
+        if (task == null || task.getRequestJson() == null || task.getRequestJson().isBlank()) {
+            return DEFAULT_TASK_TYPE;
+        }
+        try {
+            JsonNode root = objectMapper.readTree(task.getRequestJson());
+            JsonNode taskTypeNode = root.get("taskType");
+            String value = taskTypeNode == null || taskTypeNode.isNull() ? null : taskTypeNode.asText(null);
+            return value == null || value.isBlank() ? DEFAULT_TASK_TYPE : value;
+        } catch (Exception ex) {
+            return DEFAULT_TASK_TYPE;
+        }
+    }
+
+    private String buildAggregateSourceFilenameLabel(List<CollectDataSourceFileDto> sources) {
+        if (sources == null || sources.isEmpty()) {
+            return "";
+        }
+        String first = firstNonBlank(sources.get(0).getOriginalFilename(), sources.get(0).getFileKey());
+        if (sources.size() == 1) {
+            return first;
+        }
+        return first + " 等 " + sources.size() + " 个文件";
+    }
+
+    private String buildAggregateSourceFileUrl(List<CollectDataSourceFileDto> sources) {
+        if (sources == null || sources.isEmpty()) {
+            return "";
+        }
+        StringBuilder sb = new StringBuilder();
+        for (int i = 0; i < sources.size(); i++) {
+            if (i > 0) {
+                sb.append('|');
+            }
+            sb.append(sources.get(i).getFileKey());
+        }
+        return sb.toString();
+    }
+
+    private ParsedWorkbook parseWorkbook(File input, CollectDataSourceFileDto source) {
+        DataFormatter formatter = new DataFormatter();
+        try (FileInputStream fis = new FileInputStream(input); Workbook workbook = WorkbookFactory.create(fis)) {
+            Sheet sheet = workbook.getSheetAt(0);
+            Row header = sheet.getRow(0);
+            if (header == null) {
+                throw new BusinessException("Excel 表头为空");
+            }
+
+            List<String> headers = new ArrayList<>();
+            for (int i = 0; i < header.getLastCellNum(); i++) {
+                String value = normalize(formatter.formatCellValue(header.getCell(i)));
+                headers.add(value.isBlank() ? "列" + (i + 1) : value);
+            }
+
+            int keywordCol = findHeaderIndex(headers, KEYWORD_HEADER_ALIASES);
+            int statusCol = findHeaderIndex(headers, STATUS_HEADER_ALIASES);
+
+            List<ParsedRow> rows = new ArrayList<>();
+            int totalRows = 0;
+            int droppedRows = 0;
+            String filename = firstNonBlank(source.getOriginalFilename(), input.getName());
+            for (int i = 1; i <= sheet.getLastRowNum(); i++) {
+                Row row = sheet.getRow(i);
+                if (row == null) {
+                    continue;
+                }
+                Map<String, String> extra = new LinkedHashMap<>();
+                String keyword = "";
+                String statusValue = "";
+                boolean nonEmpty = false;
+                for (int c = 0; c < headers.size(); c++) {
+                    Cell cell = row.getCell(c);
+                    String value = normalize(cell == null ? "" : formatter.formatCellValue(cell));
+                    if (!value.isBlank()) {
+                        nonEmpty = true;
+                    }
+                    if (c == keywordCol) {
+                        keyword = value;
+                    } else if (c == statusCol) {
+                        statusValue = value;
+                    } else {
+                        extra.put(headers.get(c), value);
+                    }
+                }
+                if (!nonEmpty) {
+                    continue;
+                }
+                totalRows++;
+                if (keywordCol >= 0 && keyword.isBlank()) {
+                    droppedRows++;
+                    continue;
+                }
+                rows.add(new ParsedRow(source.getFileKey(), filename, keyword, statusValue, extra));
+            }
+            return new ParsedWorkbook(totalRows, droppedRows, rows);
+        } catch (BusinessException ex) {
+            throw ex;
+        } catch (Exception ex) {
+            log.warn("[collect-data] parse workbook failed file={} err={}", input, ex.getMessage());
+            throw new BusinessException("解析 Excel 失败");
+        }
+    }
+
+    private int findHeaderIndex(List<String> headers, List<String> aliases) {
+        if (headers == null || headers.isEmpty() || aliases == null || aliases.isEmpty()) {
+            return -1;
+        }
+        for (int i = 0; i < headers.size(); i++) {
+            String lower = headers.get(i).toLowerCase(Locale.ROOT);
+            for (String alias : aliases) {
+                if (lower.contains(alias.toLowerCase(Locale.ROOT))) {
+                    return i;
+                }
+            }
+        }
+        return -1;
+    }
+
+    private String formatTime(LocalDateTime time) {
+        return time == null ? null : time.format(TIME_FORMATTER);
+    }
+
+    private String firstNonBlank(String... values) {
+        if (values == null) {
+            return "";
+        }
+        for (String value : values) {
+            if (value != null && !value.isBlank()) {
+                return value;
+            }
+        }
+        return "";
+    }
+
+    private String normalize(String value) {
+        if (value == null) {
+            return "";
+        }
+        String normalized = value.replace(String.valueOf((char) 0xFEFF), "")
+                .replace((char) 0x3000, ' ')
+                .replace("\r\n", " ")
+                .replace("\r", " ")
+                .replace("\n", " ")
+                .replace("\t", " ")
+                .trim();
+        return WHITESPACE_PATTERN.matcher(normalized).replaceAll(" ");
+    }
+
+    private record ParsedRow(String sourceFileKey,
+                             String sourceFilename,
+                             String keyword,
+                             String statusValue,
+                             Map<String, String> extra) {
+    }
+
+    private record ParsedWorkbook(int totalRows, int droppedRows, List<ParsedRow> rows) {
+    }
+
+    private record PersistedParse(FileTaskEntity task, FileResultEntity result) {
+    }
+}

backend-java/src/main/java/com/nanri/aiimage/modules/deletebrand/controller/DeleteBrandRunController.java

@@ -152,13 +152,13 @@ public class DeleteBrandRunController {
     @Operation(summary = "Download one delete-brand result")
     public void downloadResult(
             @PathVariable Long resultId,
-            @Parameter(name = "user_id", description = "褰撳墠鐧诲綍鐢ㄦ埛 ID", required = true, in = ParameterIn.QUERY)
+            @Parameter(name = "user_id", description = "当前登录用户 ID", required = true, in = ParameterIn.QUERY)
             @RequestParam("user_id") Long userId,
             jakarta.servlet.http.HttpServletResponse response) {
         String url = deleteBrandRunService.resolveResultDownloadUrl(resultId, userId);
         String filename = deleteBrandRunService.resolveResultDownloadFilename(resultId, userId);
         if (url == null || url.isBlank()) {
-            throw new ResponseStatusException(HttpStatus.NOT_FOUND, "缁撴灉鏃犲彲涓嬭浇");
+            throw new ResponseStatusException(HttpStatus.NOT_FOUND, "结果无可下载");
         }
         if (filename == null || filename.isBlank()) {
             String rawPath = URI.create(url).getPath();
@@ -177,7 +177,7 @@ public class DeleteBrandRunController {
                 response.getOutputStream().flush();
             }
         } catch (Exception ex) {
-            throw new ResponseStatusException(HttpStatus.INTERNAL_SERVER_ERROR, "涓嬭浇澶辫触");
+            throw new ResponseStatusException(HttpStatus.INTERNAL_SERVER_ERROR, "下载失败");
         }
     }
 

backend-java/src/main/java/com/nanri/aiimage/modules/deletebrand/service/DeleteBrandStaleTaskService.java

@@ -628,7 +628,7 @@ public class DeleteBrandStaleTaskService {
                         .eq(FileTaskEntity::getModuleType, MODULE_TYPE_QUERY_ASIN)
                         .eq(FileTaskEntity::getStatus, "RUNNING")
                         .set(FileTaskEntity::getStatus, "FAILED")
-                        .set(FileTaskEntity::getErrorMessage, "闀挎椂闂存湭鏀跺埌 Python 缁撴灉鍥炰紶锛屼换鍔″凡鑷姩澶辫触")
+                        .set(FileTaskEntity::getErrorMessage, "长时间未收到 Python 结果回传，任务已自动失败")
                         .set(FileTaskEntity::getUpdatedAt, LocalDateTime.now())
                         .set(FileTaskEntity::getFinishedAt, LocalDateTime.now()));
                 if (updated > 0) {

backend-java/src/main/java/com/nanri/aiimage/modules/invalidasin/controller/InvalidAsinDataController.java

@@ -0,0 +1,80 @@
+package com.nanri.aiimage.modules.invalidasin.controller;
+
+import com.nanri.aiimage.common.api.ApiResponse;
+import com.nanri.aiimage.modules.invalidasin.model.dto.InvalidAsinDataCreateRequest;
+import com.nanri.aiimage.modules.invalidasin.model.dto.InvalidAsinDataUpdateRequest;
+import com.nanri.aiimage.modules.invalidasin.model.vo.InvalidAsinDataItemVo;
+import com.nanri.aiimage.modules.invalidasin.model.vo.InvalidAsinDataPageVo;
+import com.nanri.aiimage.modules.invalidasin.service.InvalidAsinDataService;
+import io.swagger.v3.oas.annotations.Operation;
+import io.swagger.v3.oas.annotations.Parameter;
+import io.swagger.v3.oas.annotations.media.Content;
+import io.swagger.v3.oas.annotations.media.Schema;
+import io.swagger.v3.oas.annotations.responses.ApiResponses;
+import io.swagger.v3.oas.annotations.tags.Tag;
+import jakarta.validation.Valid;
+import lombok.RequiredArgsConstructor;
+import org.springframework.web.bind.annotation.DeleteMapping;
+import org.springframework.web.bind.annotation.GetMapping;
+import org.springframework.web.bind.annotation.PathVariable;
+import org.springframework.web.bind.annotation.PostMapping;
+import org.springframework.web.bind.annotation.PutMapping;
+import org.springframework.web.bind.annotation.RequestBody;
+import org.springframework.web.bind.annotation.RequestMapping;
+import org.springframework.web.bind.annotation.RequestParam;
+import org.springframework.web.bind.annotation.RestController;
+
+@RestController
+@RequiredArgsConstructor
+@RequestMapping("/api/admin/invalid-asin-data")
+@Tag(name = "不符合ASIN数据", description = "维护不符合ASIN数据列表，支持增删改查。")
+public class InvalidAsinDataController {
+
+    private final InvalidAsinDataService invalidAsinDataService;
+
+    @GetMapping
+    @Operation(summary = "分页查询不符合ASIN数据", description = "分页查询不符合ASIN数据，支持按 ASIN/品牌 模糊搜索。")
+    @ApiResponses({
+            @io.swagger.v3.oas.annotations.responses.ApiResponse(responseCode = "200", description = "查询成功", content = @Content(schema = @Schema(implementation = InvalidAsinDataPageVo.class)))
+    })
+    public ApiResponse<InvalidAsinDataPageVo> page(
+            @Parameter(description = "页码") @RequestParam(defaultValue = "1") Long page,
+            @Parameter(description = "每页数量") @RequestParam(defaultValue = "15") Long pageSize,
+            @Parameter(description = "模糊搜索关键字") @RequestParam(required = false) String keyword) {
+        return ApiResponse.success(invalidAsinDataService.page(page, pageSize, keyword));
+    }
+
+    @PostMapping
+    @Operation(summary = "新增不符合ASIN数据", description = "新增一条不符合ASIN数据。")
+    @ApiResponses({
+            @io.swagger.v3.oas.annotations.responses.ApiResponse(responseCode = "200", description = "创建成功", content = @Content(schema = @Schema(implementation = InvalidAsinDataItemVo.class))),
+            @io.swagger.v3.oas.annotations.responses.ApiResponse(responseCode = "400", description = "参数不合法或数据重复")
+    })
+    public ApiResponse<InvalidAsinDataItemVo> create(@Valid @RequestBody InvalidAsinDataCreateRequest request) {
+        return ApiResponse.success("创建成功", invalidAsinDataService.create(request));
+    }
+
+    @PutMapping("/{id}")
+    @Operation(summary = "更新不符合ASIN数据", description = "按 ID 更新一条不符合ASIN数据。")
+    @ApiResponses({
+            @io.swagger.v3.oas.annotations.responses.ApiResponse(responseCode = "200", description = "更新成功", content = @Content(schema = @Schema(implementation = InvalidAsinDataItemVo.class))),
+            @io.swagger.v3.oas.annotations.responses.ApiResponse(responseCode = "400", description = "参数不合法或数据重复"),
+            @io.swagger.v3.oas.annotations.responses.ApiResponse(responseCode = "404", description = "数据不存在")
+    })
+    public ApiResponse<InvalidAsinDataItemVo> update(
+            @Parameter(description = "主键ID", required = true) @PathVariable Long id,
+            @Valid @RequestBody InvalidAsinDataUpdateRequest request) {
+        return ApiResponse.success("更新成功", invalidAsinDataService.update(id, request));
+    }
+
+    @DeleteMapping("/{id}")
+    @Operation(summary = "删除不符合ASIN数据", description = "按 ID 删除一条不符合ASIN数据。")
+    @ApiResponses({
+            @io.swagger.v3.oas.annotations.responses.ApiResponse(responseCode = "200", description = "删除成功"),
+            @io.swagger.v3.oas.annotations.responses.ApiResponse(responseCode = "404", description = "数据不存在")
+    })
+    public ApiResponse<Void> delete(@Parameter(description = "主键ID", required = true) @PathVariable Long id) {
+        invalidAsinDataService.delete(id);
+        return ApiResponse.success("删除成功", null);
+    }
+}

backend-java/src/main/java/com/nanri/aiimage/modules/invalidasin/mapper/InvalidAsinDataMapper.java

@@ -0,0 +1,9 @@
+package com.nanri.aiimage.modules.invalidasin.mapper;
+
+import com.baomidou.mybatisplus.core.mapper.BaseMapper;
+import com.nanri.aiimage.modules.invalidasin.model.entity.InvalidAsinDataEntity;
+import org.apache.ibatis.annotations.Mapper;
+
+@Mapper
+public interface InvalidAsinDataMapper extends BaseMapper<InvalidAsinDataEntity> {
+}

backend-java/src/main/java/com/nanri/aiimage/modules/invalidasin/model/dto/InvalidAsinDataCreateRequest.java

@@ -0,0 +1,20 @@
+package com.nanri.aiimage.modules.invalidasin.model.dto;
+
+import io.swagger.v3.oas.annotations.media.Schema;
+import jakarta.validation.constraints.NotBlank;
+import jakarta.validation.constraints.Size;
+import lombok.Data;
+
+@Data
+@Schema(description = "不符合ASIN数据新增请求")
+public class InvalidAsinDataCreateRequest {
+
+    @NotBlank(message = "ASIN 不能为空")
+    @Size(max = 128, message = "ASIN 长度不能超过128个字符")
+    @Schema(description = "ASIN 值", requiredMode = Schema.RequiredMode.REQUIRED)
+    private String dataValue;
+
+    @Size(max = 128, message = "品牌长度不能超过128个字符")
+    @Schema(description = "品牌名称")
+    private String brand;
+}

backend-java/src/main/java/com/nanri/aiimage/modules/invalidasin/model/dto/InvalidAsinDataUpdateRequest.java

@@ -0,0 +1,20 @@
+package com.nanri.aiimage.modules.invalidasin.model.dto;
+
+import io.swagger.v3.oas.annotations.media.Schema;
+import jakarta.validation.constraints.NotBlank;
+import jakarta.validation.constraints.Size;
+import lombok.Data;
+
+@Data
+@Schema(description = "不符合ASIN数据更新请求")
+public class InvalidAsinDataUpdateRequest {
+
+    @NotBlank(message = "ASIN 不能为空")
+    @Size(max = 128, message = "ASIN 长度不能超过128个字符")
+    @Schema(description = "ASIN 值", requiredMode = Schema.RequiredMode.REQUIRED)
+    private String dataValue;
+
+    @Size(max = 128, message = "品牌长度不能超过128个字符")
+    @Schema(description = "品牌名称")
+    private String brand;
+}

backend-java/src/main/java/com/nanri/aiimage/modules/invalidasin/model/entity/InvalidAsinDataEntity.java

@@ -0,0 +1,20 @@
+package com.nanri.aiimage.modules.invalidasin.model.entity;
+
+import com.baomidou.mybatisplus.annotation.IdType;
+import com.baomidou.mybatisplus.annotation.TableId;
+import com.baomidou.mybatisplus.annotation.TableName;
+import lombok.Data;
+
+import java.time.LocalDateTime;
+
+@Data
+@TableName("biz_invalid_asin_data")
+public class InvalidAsinDataEntity {
+
+    @TableId(type = IdType.AUTO)
+    private Long id;
+    private String dataValue;
+    private String brand;
+    private LocalDateTime createdAt;
+    private LocalDateTime updatedAt;
+}

backend-java/src/main/java/com/nanri/aiimage/modules/invalidasin/model/vo/InvalidAsinDataItemVo.java

@@ -0,0 +1,23 @@
+package com.nanri.aiimage.modules.invalidasin.model.vo;
+
+import io.swagger.v3.oas.annotations.media.Schema;
+import lombok.Data;
+
+import java.time.LocalDateTime;
+
+@Data
+@Schema(description = "不符合ASIN数据项")
+public class InvalidAsinDataItemVo {
+
+    @Schema(description = "主键ID")
+    private Long id;
+
+    @Schema(description = "ASIN 值")
+    private String dataValue;
+
+    @Schema(description = "品牌名称")
+    private String brand;
+
+    @Schema(description = "创建时间")
+    private LocalDateTime createdAt;
+}

backend-java/src/main/java/com/nanri/aiimage/modules/invalidasin/model/vo/InvalidAsinDataPageVo.java

@@ -0,0 +1,23 @@
+package com.nanri.aiimage.modules.invalidasin.model.vo;
+
+import io.swagger.v3.oas.annotations.media.Schema;
+import lombok.Data;
+
+import java.util.List;
+
+@Data
+@Schema(description = "不符合ASIN数据分页结果")
+public class InvalidAsinDataPageVo {
+
+    @Schema(description = "列表")
+    private List<InvalidAsinDataItemVo> items;
+
+    @Schema(description = "总数")
+    private Long total;
+
+    @Schema(description = "页码")
+    private Long page;
+
+    @Schema(description = "每页数量")
+    private Long pageSize;
+}

backend-java/src/main/java/com/nanri/aiimage/modules/invalidasin/service/InvalidAsinDataService.java

@@ -0,0 +1,130 @@
+package com.nanri.aiimage.modules.invalidasin.service;
+
+import com.baomidou.mybatisplus.core.conditions.query.LambdaQueryWrapper;
+import com.nanri.aiimage.common.exception.BusinessException;
+import com.nanri.aiimage.modules.invalidasin.mapper.InvalidAsinDataMapper;
+import com.nanri.aiimage.modules.invalidasin.model.dto.InvalidAsinDataCreateRequest;
+import com.nanri.aiimage.modules.invalidasin.model.dto.InvalidAsinDataUpdateRequest;
+import com.nanri.aiimage.modules.invalidasin.model.entity.InvalidAsinDataEntity;
+import com.nanri.aiimage.modules.invalidasin.model.vo.InvalidAsinDataItemVo;
+import com.nanri.aiimage.modules.invalidasin.model.vo.InvalidAsinDataPageVo;
+import lombok.RequiredArgsConstructor;
+import org.springframework.stereotype.Service;
+import org.springframework.transaction.annotation.Transactional;
+
+import java.util.List;
+
+@Service
+@RequiredArgsConstructor
+public class InvalidAsinDataService {
+
+    private final InvalidAsinDataMapper invalidAsinDataMapper;
+
+    public InvalidAsinDataPageVo page(long page, long pageSize, String keyword) {
+        long safePage = Math.max(page, 1);
+        long safePageSize = Math.min(Math.max(pageSize, 1), 100);
+        String safeKeyword = keyword == null ? "" : keyword.trim();
+        LambdaQueryWrapper<InvalidAsinDataEntity> query = new LambdaQueryWrapper<InvalidAsinDataEntity>()
+                .and(!safeKeyword.isEmpty(), wrapper -> wrapper
+                        .like(InvalidAsinDataEntity::getDataValue, safeKeyword)
+                        .or()
+                        .like(InvalidAsinDataEntity::getBrand, safeKeyword))
+                .orderByDesc(InvalidAsinDataEntity::getId);
+        Long total = invalidAsinDataMapper.selectCount(query);
+        List<InvalidAsinDataItemVo> items = invalidAsinDataMapper
+                .selectList(query.last("LIMIT " + ((safePage - 1) * safePageSize) + ", " + safePageSize))
+                .stream()
+                .map(this::toItemVo)
+                .toList();
+        InvalidAsinDataPageVo vo = new InvalidAsinDataPageVo();
+        vo.setItems(items);
+        vo.setTotal(total);
+        vo.setPage(safePage);
+        vo.setPageSize(safePageSize);
+        return vo;
+    }
+
+    @Transactional
+    public InvalidAsinDataItemVo create(InvalidAsinDataCreateRequest request) {
+        String dataValue = normalizeRequired(request.getDataValue(), "ASIN 不能为空");
+        String brand = normalizeOptional(request.getBrand());
+        ensureUnique(dataValue, null);
+        InvalidAsinDataEntity entity = new InvalidAsinDataEntity();
+        entity.setDataValue(dataValue);
+        entity.setBrand(brand);
+        invalidAsinDataMapper.insert(entity);
+        return toItemVo(getById(entity.getId()));
+    }
+
+    @Transactional
+    public InvalidAsinDataItemVo update(Long id, InvalidAsinDataUpdateRequest request) {
+        InvalidAsinDataEntity entity = getById(id);
+        String dataValue = normalizeRequired(request.getDataValue(), "ASIN 不能为空");
+        String brand = normalizeOptional(request.getBrand());
+        ensureUnique(dataValue, id);
+        entity.setDataValue(dataValue);
+        entity.setBrand(brand);
+        invalidAsinDataMapper.updateById(entity);
+        return toItemVo(getById(id));
+    }
+
+    @Transactional
+    public void delete(Long id) {
+        InvalidAsinDataEntity entity = getById(id);
+        invalidAsinDataMapper.deleteById(entity.getId());
+    }
+
+    private InvalidAsinDataEntity getById(Long id) {
+        InvalidAsinDataEntity entity = invalidAsinDataMapper.selectById(id);
+        if (entity == null) {
+            throw new BusinessException("数据不存在");
+        }
+        return entity;
+    }
+
+    private void ensureUnique(String dataValue, Long excludeId) {
+        LambdaQueryWrapper<InvalidAsinDataEntity> query = new LambdaQueryWrapper<InvalidAsinDataEntity>()
+                .eq(InvalidAsinDataEntity::getDataValue, dataValue)
+                .last("LIMIT 1");
+        InvalidAsinDataEntity exists = invalidAsinDataMapper.selectOne(query);
+        if (exists != null && (excludeId == null || !exists.getId().equals(excludeId))) {
+            throw new BusinessException("ASIN 已存在");
+        }
+    }
+
+    private String normalizeRequired(String value, String message) {
+        String normalized = normalizeText(value);
+        if (normalized.isEmpty()) {
+            throw new BusinessException(message);
+        }
+        return normalized;
+    }
+
+    private String normalizeOptional(String value) {
+        String normalized = normalizeText(value);
+        return normalized.isEmpty() ? null : normalized;
+    }
+
+    private String normalizeText(String value) {
+        if (value == null) {
+            return "";
+        }
+        return value.replace("", "")
+                .replace("　", " ")
+                .replace("\r\n", " ")
+                .replace("\r", " ")
+                .replace("\n", " ")
+                .replace("\t", " ")
+                .trim()
+                .replaceAll("\\s+", " ");
+    }
+
+    private InvalidAsinDataItemVo toItemVo(InvalidAsinDataEntity entity) {
+        InvalidAsinDataItemVo vo = new InvalidAsinDataItemVo();
+        vo.setId(entity.getId());
+        vo.setDataValue(entity.getDataValue());
+        vo.setBrand(entity.getBrand());
+        vo.setCreatedAt(entity.getCreatedAt());
+        return vo;
+    }
+}

backend-java/src/main/java/com/nanri/aiimage/modules/patroldelete/service/PatrolDeleteTaskService.java

@@ -1044,7 +1044,7 @@ public class PatrolDeleteTaskService {
     private TaskDistributedLockService.LockHandle acquireTaskLockOrThrow(Long taskId) {
         TaskDistributedLockService.LockHandle lockHandle = acquireTaskLock(taskId);
         if (lockHandle == null) {
-            throw new BusinessException(40901, "浠诲姟姝ｅ湪澶勭悊涓紝璇风◢鍚庡啀璇?");
+            throw new BusinessException(40901, "任务正在处理中，请稍后再试");
         }
         return lockHandle;
     }

backend-java/src/main/java/com/nanri/aiimage/modules/permission/service/PermissionMenuSchemaInitializer.java

@@ -29,6 +29,7 @@ public class PermissionMenuSchemaInitializer {
             new DefaultAdminMenu("用户管理", "admin_users", "users", 10),
             new DefaultAdminMenu("栏目权限配置", "admin_columns", "columns", 20),
             new DefaultAdminMenu("数据去重总数据", "admin_dedupe_total_data", "dedupe-total-data", 30),
+            new DefaultAdminMenu("不符合ASIN数据", "admin_invalid_asin_data", "invalid-asin-data", 35),
             new DefaultAdminMenu("店铺密钥管理", "admin_shop_keys", "shop-keys", 40),
             new DefaultAdminMenu("店铺管理", "admin_shop_manage", "shop-manage", 50),
             new DefaultAdminMenu("跳过跟价ASIN", "admin_skip_price_asin", "skip-price-asin", 60),

backend-java/src/main/java/com/nanri/aiimage/modules/pricetrack/service/PriceTrackTaskService.java

@@ -1391,7 +1391,7 @@ public class PriceTrackTaskService {
     private TaskDistributedLockService.LockHandle acquireTaskLockOrThrow(Long taskId) {
         TaskDistributedLockService.LockHandle lockHandle = acquireTaskLock(taskId);
         if (lockHandle == null) {
-            throw new BusinessException(40901, "浠诲姟姝ｅ湪澶勭悊涓紝璇风◢鍚庡啀璇?");
+            throw new BusinessException(40901, "任务正在处理中，请稍后再试");
         }
         return lockHandle;
     }

backend-java/src/main/java/com/nanri/aiimage/modules/productrisk/service/ProductRiskTaskService.java

@@ -970,7 +970,7 @@ public class ProductRiskTaskService {
     private TaskDistributedLockService.LockHandle acquireTaskLockOrThrow(Long taskId) {
         TaskDistributedLockService.LockHandle lockHandle = acquireTaskLock(taskId);
         if (lockHandle == null) {
-            throw new BusinessException(40901, "浠诲姟姝ｅ湪澶勭悊涓紝璇风◢鍚庡啀璇?");
+            throw new BusinessException(40901, "任务正在处理中，请稍后再试");
         }
         return lockHandle;
     }