fix(撞款扫描): 保留期 90→7 天，并单独兜住最新 SUCCESS 行

该表每行含整份聚合 payload（线上实测约 2.3MB/行，24 行占 55MB），
而读取侧只认最新一行（selectLatestFullRow/LightRow 都是 ORDER BY id DESC LIMIT 1），
历史行纯占磁盘。保留期降到 7 天，行数下限仍由 PROTECTED_ROWS=10 兜底。

同时补一条保护线：连续失败多日时，最新 SUCCESS 行会滑出「最新 N 行」保护窗口，
再按时间线被删掉，界面就空白了——现在取「第 N 新行」与「最新 SUCCESS 行」
两者更靠前的那个 id 作为保护线，确保它永不被删。

backend-java/src/main/java/com/nanri/aiimage/modules/shopduplicatecheck/service/ShopDataDuplicateScanRetentionService.java

@@ -2,6 +2,7 @@ package com.nanri.aiimage.modules.shopduplicatecheck.service;
 
 import com.nanri.aiimage.common.service.DistributedJobLockService;
 import com.nanri.aiimage.modules.shopduplicatecheck.mapper.ShopDataDuplicateScanMapper;
+import com.nanri.aiimage.modules.shopduplicatecheck.model.dto.ScanLightRowDto;
 import lombok.RequiredArgsConstructor;
 import lombok.extern.slf4j.Slf4j;
 import org.springframework.beans.factory.annotation.Value;
@@ -35,8 +36,8 @@ public class ShopDataDuplicateScanRetentionService {
     private final ShopDataDuplicateScanMapper scanMapper;
     private final DistributedJobLockService distributedJobLockService;
 
-    @Value("${aiimage.shop-duplicate-check.scan-retention-days:90}")
-    private int retentionDays = 90;
+    @Value("${aiimage.shop-duplicate-check.scan-retention-days:7}")
+    private int retentionDays = 7;
 
     @Value("${aiimage.shop-duplicate-check.scan-retention-batch-size:200}")
     private int retentionBatchSize = 200;
@@ -54,9 +55,13 @@ public class ShopDataDuplicateScanRetentionService {
             return;
         }
         try (lockHandle) {
+            // 两条保护线取更靠前（更小）的那个 id——保护线之内的行一律不删：
+            // ① 最新的 PROTECTED_ROWS 行：万一扫描停摆很久，界面上仍能看到最后一份结果；
+            // ② 最新 SUCCESS 行：读侧只认它。连续失败多日时它可能已滑出保护窗口，
+            //    按时间线会被删掉、界面直接空白，所以单独兜一条。
             Long nthNewestId = scanMapper.selectNthNewestId(PROTECTED_ROWS - 1);
-            // 行数不足保护量时没什么可删，直接返回
-            long protectFromId = nthNewestId == null ? Long.MAX_VALUE : nthNewestId;
+            ScanLightRowDto latestSuccess = scanMapper.selectLatestLightRow();
+            long protectFromId = minId(nthNewestId, latestSuccess == null ? null : latestSuccess.getId());
             int totalDeleted = 0;
             int batches = 0;
             while (batches < MAX_BATCHES_PER_RUN) {
@@ -73,4 +78,12 @@ public class ShopDataDuplicateScanRetentionService {
             log.warn("[shop-duplicate-check] 扫描结果保留清理失败 cutoff={} msg={}", cutoff, ex.getMessage(), ex);
         }
     }
+
+    /** 取两条保护线里更靠前（更小）的 id；都为 null 表示无需保护（等价于不设限）。 */
+    private static long minId(Long first, Long second) {
+        if (first == null) {
+            return second == null ? Long.MAX_VALUE : second;
+        }
+        return second == null ? first : Math.min(first, second);
+    }
 }

backend-java/src/main/resources/application.yml

@@ -233,7 +233,9 @@ aiimage:
     loop-run-retention-batch-size: ${AIIMAGE_PRICE_TRACK_LOOP_RUN_RETENTION_BATCH_SIZE:500}
     loop-run-retention-cron: ${AIIMAGE_PRICE_TRACK_LOOP_RUN_RETENTION_CRON:0 20 4 * * *}
   shop-duplicate-check:
-    scan-retention-days: ${AIIMAGE_SHOP_DUPLICATE_SCAN_RETENTION_DAYS:90}
+    # 每行含整份聚合 payload（实测约 2.3MB/行），而读取侧只认最新一行，
+    # 保留期给 7 天足够排查；行数下限由服务里的 PROTECTED_ROWS 兜底
+    scan-retention-days: ${AIIMAGE_SHOP_DUPLICATE_SCAN_RETENTION_DAYS:7}
     scan-retention-batch-size: ${AIIMAGE_SHOP_DUPLICATE_SCAN_RETENTION_BATCH_SIZE:200}
     scan-retention-cron: ${AIIMAGE_SHOP_DUPLICATE_SCAN_RETENTION_CRON:0 50 3 * * *}
   permission-schema-init:

backend-java/src/test/java/com/nanri/aiimage/modules/shopduplicatecheck/service/ShopDataDuplicateScanRetentionServiceTest.java

@@ -2,10 +2,12 @@ package com.nanri.aiimage.modules.shopduplicatecheck.service;
 
 import com.nanri.aiimage.common.service.DistributedJobLockService;
 import com.nanri.aiimage.modules.shopduplicatecheck.mapper.ShopDataDuplicateScanMapper;
+import com.nanri.aiimage.modules.shopduplicatecheck.model.dto.ScanLightRowDto;
 import org.junit.jupiter.api.Test;
 
 import java.time.LocalDateTime;
 
+import static org.junit.jupiter.api.Assertions.assertEquals;
 import static org.junit.jupiter.api.Assertions.assertTrue;
 import static org.mockito.ArgumentMatchers.anyLong;
 import static org.mockito.ArgumentMatchers.anyInt;
@@ -94,4 +96,29 @@ class ShopDataDuplicateScanRetentionServiceTest {
         verify(scanMapper).deleteOlderThanBatch(any(LocalDateTime.class), protectLine.capture(), anyInt());
         assertTrue(protectLine.getValue() == Long.MAX_VALUE, "保护线为 Long.MAX_VALUE，等于不额外设限");
     }
+
+    /**
+     * 连续失败多日时，最新 SUCCESS 行可能已滑出「最新 N 行」保护窗口；
+     * 读侧只认它，被删掉界面就空白了，必须单独兜一条更靠前的保护线。
+     */
+    @Test
+    void latestSuccessRowIsAlwaysProtected() {
+        lockAvailable();
+        when(scanMapper.selectNthNewestId(anyInt())).thenReturn(500L);
+        when(scanMapper.selectLatestLightRow()).thenReturn(successRow(120L));
+        when(scanMapper.deleteOlderThanBatch(any(LocalDateTime.class), anyLong(), anyInt())).thenReturn(0);
+
+        service().purgeExpiredScans();
+
+        ArgumentCaptor<Long> protectLine = ArgumentCaptor.forClass(Long.class);
+        verify(scanMapper).deleteOlderThanBatch(any(LocalDateTime.class), protectLine.capture(), anyInt());
+        assertEquals(120L, protectLine.getValue(), "取更靠前的保护线，SUCCESS 行不会被删");
+    }
+
+    private static ScanLightRowDto successRow(Long id) {
+        ScanLightRowDto row = new ScanLightRowDto();
+        row.setId(id);
+        row.setCreatedAt(LocalDateTime.now().minusDays(30));
+        return row;
+    }
 }