From c4fc9784e8be47aab99c82c4fbc31abd6a20e06b Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?=E9=BB=84=E8=87=AA=E8=BE=BE?= <980324341@qq.com>
Date: Wed, 28 Oct 2026 19:31:50 +0800
Subject: [PATCH] =?UTF-8?q?feat(collect-data):=20=E7=BB=93=E6=9E=9C=20Exce?=
=?UTF-8?q?l=20=E8=BF=BD=E5=8A=A0=E3=80=8C=E8=A2=AB=E8=BF=87=E6=BB=A4?=
=?UTF-8?q?=E6=98=8E=E7=BB=86=E3=80=8Dsheet=EF=BC=8C=E5=93=81=E7=89=8C?=
=?UTF-8?q?=E6=A3=80=E6=B5=8B=E5=A4=B1=E8=B4=A5=E5=8E=9F=E5=9B=A0=E5=8F=AF?=
=?UTF-8?q?=E8=A7=81?=
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit
task 32573 回传 600 行只留 1 行,被砍掉的 599 行在系统里只剩四个计数,
谁都答不出「是哪些 ASIN、死在哪一道」。四类被过滤行现在于提交时当场
逐行落库(V179 biz_collectdata_filtered_row),结果 Excel 追加第 3 个
sheet 供逐行对账。不在装配期回读 chunk 重判——去重库与无效品牌库读的
都是当前态,任务窗口期的写入会让判定漂移(错标/漏标/分类串台),
依据逐条写在 V179 头注释里。
同批:brandQueryFailedCount 与原因进历史 VO 和任务卡片;损耗通知不再对
「未配置代理」的用户承诺稍后重跑——补测队列两轮 21/21、28/28 全败、净增
0 行,那句是谎。
台账:XML_STATEMENT_TOTAL 141→143(新 mapper 两条写侧语句);静默 catch
棘轮 total 1028→1031、不重抛 579→582、只 log 181→182(三枚审计件容错,
逐处记名于 CatchSilenceRatchetTest 上方注释)。
---
.../mapper/CollectDataFilteredRowMapper.java | 21 ++
.../entity/CollectDataFilteredRowEntity.java | 51 ++++
.../model/vo/CollectDataHistoryItemVo.java | 6 +
.../CollectDataBrandRecheckInjector.java | 8 +
.../CollectDataExcelAssemblyService.java | 136 ++++++++++-
.../CollectDataItemCleanupSpiImpl.java | 13 +-
.../service/CollectDataService.java | 140 ++++++++++-
.../util/CollectDataBatchQuery.java | 24 +-
.../util/CollectDataFilteredRowStore.java | 225 ++++++++++++++++++
.../task/spi/CollectDataItemCleanupSpi.java | 14 +-
.../src/main/resources/application.yml | 5 +
.../db/V179__collectdata_filtered_row.sql | 48 ++++
.../CollectDataFilteredRowMapper.xml | 32 +++
.../aiimage/SqlLayeringConventionTest.java | 5 +-
.../config/CatchSilenceRatchetTest.java | 35 ++-
.../CollectDataBrandRecheckInjectorTest.java | 12 +-
.../service/CollectDataFilteredSheetTest.java | 182 ++++++++++++++
.../CollectDataPartialResultNoticeTest.java | 59 ++++-
...llectDataResultFileReuploadCompatTest.java | 114 +++++++++
.../CollectDataServiceTxBoundaryTest.java | 2 +-
.../CollectDataSubmitResultSnapshotTest.java | 2 +-
.../util/CollectDataBatchQueryTest.java | 33 +++
.../util/CollectDataFilteredRowStoreTest.java | 183 ++++++++++++++
.../CollectDataSelfConsistencyTest.java | 3 +
.../DuplicateSubmissionContractTest.java | 2 +-
.../contract/SuccessTimingContractTest.java | 4 +-
.../golden/submit-result-snapshot.txt | 16 +-
.../brand/components/BrandCollectDataTab.vue | 10 +
.../shared/api/types/modules/collect-data.ts | 4 +
...ollect-brand-query-failure-visible.test.ts | 47 ++++
30 files changed, 1384 insertions(+), 52 deletions(-)
create mode 100644 backend-java/src/main/java/com/nanri/aiimage/modules/collectdata/mapper/CollectDataFilteredRowMapper.java
create mode 100644 backend-java/src/main/java/com/nanri/aiimage/modules/collectdata/model/entity/CollectDataFilteredRowEntity.java
create mode 100644 backend-java/src/main/java/com/nanri/aiimage/modules/collectdata/util/CollectDataFilteredRowStore.java
create mode 100644 backend-java/src/main/resources/db/V179__collectdata_filtered_row.sql
create mode 100644 backend-java/src/main/resources/mapper/collectdata/CollectDataFilteredRowMapper.xml
create mode 100644 backend-java/src/test/java/com/nanri/aiimage/modules/collectdata/service/CollectDataFilteredSheetTest.java
create mode 100644 backend-java/src/test/java/com/nanri/aiimage/modules/collectdata/service/CollectDataResultFileReuploadCompatTest.java
create mode 100644 backend-java/src/test/java/com/nanri/aiimage/modules/collectdata/util/CollectDataFilteredRowStoreTest.java
create mode 100644 frontend-vue/tests/collect-brand-query-failure-visible.test.ts
diff --git a/backend-java/src/main/java/com/nanri/aiimage/modules/collectdata/mapper/CollectDataFilteredRowMapper.java b/backend-java/src/main/java/com/nanri/aiimage/modules/collectdata/mapper/CollectDataFilteredRowMapper.java
new file mode 100644
index 000000000..43ea4491e
--- /dev/null
+++ b/backend-java/src/main/java/com/nanri/aiimage/modules/collectdata/mapper/CollectDataFilteredRowMapper.java
@@ -0,0 +1,21 @@
+package com.nanri.aiimage.modules.collectdata.mapper;
+
+import com.baomidou.mybatisplus.core.mapper.BaseMapper;
+import com.nanri.aiimage.modules.collectdata.model.entity.CollectDataFilteredRowEntity;
+import org.apache.ibatis.annotations.Mapper;
+import org.apache.ibatis.annotations.Param;
+
+import java.util.Collection;
+import java.util.List;
+
+@Mapper
+public interface CollectDataFilteredRowMapper extends BaseMapper 必须在提交时当场记,不能装配时回读 chunk 重判——去重与品牌库读的都是当前态,
+ * 任务窗口期的写入会让判定漂移(错标/漏标/分类串台),依据写在 V179 迁移头注释里。 这里不用 {@code status} 字符串列:TaskStatusVocabularyRatchetTest 把"模块里另抄一套
+ * {@code STATUS_X = "字面量"}"当成棘轮要消灭的东西,而这列的语义本来就只有"回没回来"。
+ * 换成时间戳还多得一个信息——回流发生在哪一轮,{@code RESOLVED} 把它丢了。 清理范围是「本模块拥有的行」,不只 biz_collect_data_item:V179 的被过滤行明细同样
+ * 归本模块,任务过期后必须一起删掉,否则它只会单向堆积。> rawItemsSupplier) {
+ writeWorkbookSegmented(outputXlsx, items, summaries, rawItemsSupplier, null);
}
/**
@@ -55,7 +71,8 @@ public class CollectDataExcelAssemblyService {
public void writeWorkbookSegmented(File outputXlsx,
List
> rawItemsSupplier) {
+ Supplier
> rawItemsSupplier,
+ FilteredRowSource filteredSource) {
// List 入参统一转成分批来源(单批即全部行),对外行为保持不变。
writeWorkbookStreaming(outputXlsx,
batchConsumer -> {
@@ -69,7 +86,16 @@ public class CollectDataExcelAssemblyService {
if (rawItems != null && !rawItems.isEmpty()) {
batchConsumer.accept(rawItems);
}
- });
+ },
+ filteredSource);
+ }
+
+ /** 不出「被过滤明细」sheet 的那一档(同 writeWorkbookSegmented 的 4 参版)。 */
+ public long writeWorkbookStreaming(File outputXlsx,
+ RowBatchSource itemsSource,
+ List
> batchConsumer);
}
+ /**
+ * 被过滤行(V179)的分批来源。除分批外还要自报「超出上限而未逐条记录」的缺口数——
+ * 少了这个数字,sheet 只会安静地少给行,没人知道少给了多少。
+ */
+ public interface FilteredRowSource {
+ void forEachBatch(Consumer
> batchConsumer);
+
+ default long notRecordedCount() {
+ return 0L;
+ }
+ }
+
private void deleteQuietly(File file) {
try {
Files.deleteIfExists(file.toPath());
@@ -241,6 +285,88 @@ public class CollectDataExcelAssemblyService {
}
}
+ /**
+ * 「被过滤明细」尾 sheet。两条硬规矩:
+ *
+ *
+ */
+ private void writeFilteredSheet(SxssfWorkbookWriter writer, FilteredRowSource source) {
+ if (source == null) {
+ return;
+ }
+ Sheet[] created = new Sheet[1];
+ int[] rowIndex = new int[]{1};
+ boolean incomplete = false;
+ try {
+ source.forEachBatch(batch -> {
+ if (batch == null || batch.isEmpty()) {
+ return;
+ }
+ for (CollectDataFilteredRowEntity item : batch) {
+ if (item == null) {
+ continue;
+ }
+ if (created[0] == null) {
+ created[0] = createFilteredSheet(writer);
+ }
+ Row row = created[0].createRow(rowIndex[0]++);
+ SxssfWorkbookWriter.setText(row, 0, BrandLabel.strip(item.getBrand()));
+ SxssfWorkbookWriter.setText(row, 1, item.getAsin());
+ SxssfWorkbookWriter.setText(row, 2, item.getKeyword());
+ SxssfWorkbookWriter.setText(row, 3, filteredReasonLabel(item.getReasonCode()));
+ SxssfWorkbookWriter.setText(row, 4, item.getReasonDetail());
+ }
+ });
+ long notRecorded = source.notRecordedCount();
+ if (notRecorded > 0) {
+ // 超出单任务追溯上限:如实说出缺口,而不是安静地少给行
+ if (created[0] == null) {
+ created[0] = createFilteredSheet(writer);
+ }
+ Row row = created[0].createRow(rowIndex[0]);
+ SxssfWorkbookWriter.setText(row, 3, "未逐条记录");
+ SxssfWorkbookWriter.setText(row, 4, "另有 " + notRecorded
+ + " 行被过滤但未逐条记录(超出单任务追溯上限),行数已计入各道闸的计数");
+ }
+ } catch (RuntimeException ex) {
+ incomplete = true;
+ log.warn("[collect-data] filtered sheet write failed, deliver without it: {}", ex.getMessage());
+ }
+ if (incomplete && created[0] != null) {
+ // 这句收尾不再套第二层 catch:它只是一次本地 POI 写,写不动说明整个 workbook
+ // 都写不动——那由外层 writeWorkbookStreaming 的既有出口处理(与前两张 sheet 同姿势),
+ // 而不是在这里再吞一层、把"交付没做成"伪装成"交付做成了但清单不完整"。
+ Row row = created[0].createRow(rowIndex[0]);
+ SxssfWorkbookWriter.setText(row, 3, "清单不完整");
+ SxssfWorkbookWriter.setText(row, 4, "本 sheet 生成过程中出错,以上仅部分被过滤行");
+ }
+ }
+
+ private Sheet createFilteredSheet(SxssfWorkbookWriter writer) {
+ Sheet sheet = writer.createSheet(SHEET_FILTERED_NAME, SHEET_FILTERED_HEADER, null);
+ for (int i = 0; i < SHEET_FILTERED_HEADER.length; i++) {
+ sheet.setColumnWidth(i, SxssfWorkbookWriter.width(i == 4 ? 46 : 18));
+ }
+ return sheet;
+ }
+
+ /** 库里存的是 reason_code(机器读),这一列给人读;未知码原样回显,不静默变空。 */
+ private static String filteredReasonLabel(String reasonCode) {
+ if (reasonCode == null || reasonCode.isBlank()) {
+ return "";
+ }
+ return switch (reasonCode) {
+ case CollectDataFilteredRowEntity.REASON_DEDUPE -> "已存在于去重总数据库";
+ case CollectDataFilteredRowEntity.REASON_INVALID_BRAND -> "命中无效品牌库";
+ case CollectDataFilteredRowEntity.REASON_BRAND_REJECTED -> "品牌检测判定侵权";
+ case CollectDataFilteredRowEntity.REASON_BRAND_QUERY_FAILED -> "品牌检测未返回结果";
+ default -> reasonCode;
+ };
+ }
+
private String safe(String value) {
return value == null ? "" : value;
}
diff --git a/backend-java/src/main/java/com/nanri/aiimage/modules/collectdata/service/CollectDataItemCleanupSpiImpl.java b/backend-java/src/main/java/com/nanri/aiimage/modules/collectdata/service/CollectDataItemCleanupSpiImpl.java
index 632251391..992218408 100644
--- a/backend-java/src/main/java/com/nanri/aiimage/modules/collectdata/service/CollectDataItemCleanupSpiImpl.java
+++ b/backend-java/src/main/java/com/nanri/aiimage/modules/collectdata/service/CollectDataItemCleanupSpiImpl.java
@@ -2,7 +2,9 @@ package com.nanri.aiimage.modules.collectdata.service;
import com.baomidou.mybatisplus.core.conditions.query.LambdaQueryWrapper;
import com.nanri.aiimage.modules.collectdata.mapper.CollectDataItemMapper;
+import com.nanri.aiimage.modules.collectdata.mapper.CollectDataFilteredRowMapper;
import com.nanri.aiimage.modules.collectdata.model.entity.CollectDataItemEntity;
+import com.nanri.aiimage.modules.collectdata.model.entity.CollectDataFilteredRowEntity;
import com.nanri.aiimage.modules.task.spi.CollectDataItemCleanupSpi;
import lombok.RequiredArgsConstructor;
import org.springframework.stereotype.Service;
@@ -10,20 +12,27 @@ import org.springframework.stereotype.Service;
import java.util.List;
/**
- * 采集明细行的历史清理实现(2026-09 全维度审查 G5:逻辑从 task 模块收回本模块)。
+ * 采集模块名下数据的清理实现(2026-09 全维度审查 G5:逻辑从 task 模块收回本模块)。
+ *
+ *
被砍行必须在这里当场交出去:本方法的两个判定读的都是当前态(去重库、无效品牌库),
+ * 事后回读 chunk 重判会漂移。计数与 {@code filteredRows} 由同一次循环产出,天然自洽。
+ */
+ public record FilterResult(List 写侧容错口径照 {@link CollectDataInvalidAsinBatchWriter}:按批 try/catch,失败只跳批告警,
+ * 绝不影响回传主流程。这份明细是审计件、不是判定依据——为它打掉用户已采回的数据,
+ * 是拿可追溯性去换交付。也因此它不是强一致的:对不上账时以 result_json 的四个计数为准。 空 ASIN 的行照记不误。空品牌被判 rejected 的行在黑名单表里根本不留痕
+ * ({@code CollectDataInvalidAsinBatchWriter.toEntity} 对空品牌返 null),装配期也无从反查,
+ * 这一格正是 V179 头注释里点名的「漏标」来源。 任务历史清理时除了 task 自己的表,还要连带删除 biz_collect_data_item 的明细行;
- * 该表属于采集模块,原先由 {@code ModuleHistoryCleanupService} 直接使用其 Mapper,
- * 构成 task → 业务的依赖。现由采集模块实现本接口。
+ * 任务历史清理时除了 task 自己的表,还要连带删除采集模块自己的行;这些表属于采集模块,
+ * 原先由 {@code ModuleHistoryCleanupService} 直接使用其 Mapper,构成 task → 业务的依赖。
+ * 现由采集模块实现本接口。
+ *
+ * 实现方的清理范围是「本模块拥有的全部行」,会随版本增加:目前是
+ * biz_collect_data_item 与 V179 的 biz_collectdata_filtered_row。调用方不需要知道有几张表,
+ * 也不该在 task 模块里为新增的采集表补删除——那等于把收回来的依赖又放回去。
*/
public interface CollectDataItemCleanupSpi {
/**
- * 删除这些任务名下的采集明细行。
+ * 删除这些任务名下的采集明细行(含本模块后续的随行数据)。
*
* @return 删除行数
*/
diff --git a/backend-java/src/main/resources/application.yml b/backend-java/src/main/resources/application.yml
index fe80a12f4..e34ce5c8b 100644
--- a/backend-java/src/main/resources/application.yml
+++ b/backend-java/src/main/resources/application.yml
@@ -488,6 +488,11 @@ aiimage:
brand-check-chunk-budget-millis: ${AIIMAGE_COLLECT_DATA_BRAND_CHECK_CHUNK_BUDGET_MILLIS:10000}
invalid-asin-batch-size: ${AIIMAGE_COLLECT_DATA_INVALID_ASIN_BATCH_SIZE:100}
result-item-batch-size: ${AIIMAGE_COLLECT_DATA_RESULT_ITEM_BATCH_SIZE:100}
+ filtered-row-batch-size: ${AIIMAGE_COLLECT_DATA_FILTERED_ROW_BATCH_SIZE:200}
+ # 单任务「被过滤行」逐条落库的上限(V179,供结果 Excel 的「被过滤明细」sheet 追溯)。
+ # 超限后只累加计数、不再记行,sheet 会如实写「另有 N 行未逐条记录」而不是静默少给。
+ # 参照量级:task 32573 被过滤 599 行、32058 被过滤 2616 行;设为 0 表示不限。
+ max-filtered-rows-per-task: ${AIIMAGE_COLLECT_DATA_MAX_FILTERED_ROWS_PER_TASK:50000}
# ASIN 采集结果跨任务缓存(详见 AsinResultCacheService)。新鲜度口径放在服务端,
# 采集端版本不齐时老客户端只会"问不到就照常开页",不会各按各的理解复用到过期数据。
asin-cache:
diff --git a/backend-java/src/main/resources/db/V179__collectdata_filtered_row.sql b/backend-java/src/main/resources/db/V179__collectdata_filtered_row.sql
new file mode 100644
index 000000000..2704a1ecf
--- /dev/null
+++ b/backend-java/src/main/resources/db/V179__collectdata_filtered_row.sql
@@ -0,0 +1,48 @@
+-- V179: 采集数据「被过滤行」明细表
+--
+-- 背景(2026-10-28 实证,task 32573 / uid=1210):客户端回传 600 行,结果只留 1 行。
+-- 被砍掉的 599 行在系统里**只剩四个计数**(去重 48 / 黑名单命中 377 / 品牌判侵权 1 /
+-- 品牌检测未出结论 173),行内容在 CollectDataBatchQuery.filter 与
+-- CollectDataBrandBatchFilter.filter 里当场丢弃。后果是任何人都答不出
+-- 「到底是哪些 ASIN、死在哪一道」:用户只看到"保留 1",运维只能逐 chunk 翻日志猜。
+-- 本表把这四类被过滤行**逐行带原因**记下来,供结果 Excel 的「被过滤明细」sheet 装配与对账。
+--
+-- 为什么必须在提交时记、不能在装配时回读 chunk 重判(三种错法,均已核到代码):
+-- 1. 错标:去重与无效品牌两支查的都是**当前态**。任务窗口期(32573 跑了近 4 小时)
+-- 任何后台导入,都会让当时 kept 的行在装配时"命中"。
+-- 2. 漏标:uk_data_value_brand + INSERT IGNORE 使已存在的 (asin,brand) 不落新行(零证据);
+-- 空品牌的 rejected 行 toEntity 直接返 null 从不写表,但 classify 把它们算进 rejected。
+-- 3. 分类串台:本任务自己的 rejected 行会先写进 biz_invalid_asin_data,而 filter() 只认
+-- 品牌表——装配期重判会把"当时被黑名单砍"和"本次刚被判侵权"统统归成同一类。
+--
+-- 口径:
+-- - 只记事实,不参与任何判定;本表任何一行写失败都**不得**影响回传主流程(writer 按批
+-- try/catch 跳批,与 CollectDataInvalidAsinBatchWriter 同口径),故它不是强一致审计件;
+-- - reason_code 四类:DEDUPE / INVALID_BRAND / BRAND_REJECTED / BRAND_QUERY_FAILED;
+-- - 补测回流不删行、也不另立 status 词表:resolved_at 从 NULL 变成时间戳就是唯一状态,
+-- 并在 reason_detail 追加「已由补测回流」。**不删行**——删了就把"当时确实没查动"这条
+-- 历史一起删掉了;
+-- - 体量闸门在应用侧(aiimage.collect-data.max-filtered-rows-per-task),不在库里做。
+--
+-- 影响表与数据量:新建空表。单任务量级参照 32573(599 行)与 32058(2616 行)。
+--
+-- 回滚(手工):
+-- DROP TABLE IF EXISTS biz_collectdata_filtered_row;
+
+CREATE TABLE IF NOT EXISTS biz_collectdata_filtered_row (
+ id BIGINT NOT NULL AUTO_INCREMENT COMMENT 'primary key',
+ task_id BIGINT NOT NULL COMMENT 'biz_file_task.id(COLLECT_DATA 模块)',
+ chunk_index INT NOT NULL DEFAULT 0 COMMENT '产生本行的回传分片序号(0=收尾/未知)',
+ asin VARCHAR(32) NOT NULL DEFAULT '' COMMENT '被过滤的 ASIN(归一化大写,与结果 sheet 同口径)',
+ keyword VARCHAR(255) NULL COMMENT '所属关键词',
+ brand VARCHAR(255) NULL COMMENT '品牌(BrandLabel.strip 展示口径,与明细 sheet 同形)',
+ reason_code VARCHAR(24) NOT NULL COMMENT 'DEDUPE / INVALID_BRAND / BRAND_REJECTED / BRAND_QUERY_FAILED',
+ reason_detail VARCHAR(500) NULL COMMENT '可展示的原因(如「用户未配置代理」)',
+ resolved_at DATETIME NULL COMMENT '补测回流时间;NULL=当时没查动且至今没回来(本表的唯一状态)',
+ created_at DATETIME NOT NULL DEFAULT CURRENT_TIMESTAMP COMMENT 'created time',
+ updated_at DATETIME NOT NULL DEFAULT CURRENT_TIMESTAMP ON UPDATE CURRENT_TIMESTAMP COMMENT 'updated time',
+ PRIMARY KEY (id),
+ KEY idx_filtered_task_reason (task_id, reason_code) COMMENT '装配 sheet 与四类计数对账',
+ KEY idx_filtered_task_asin (task_id, asin) COMMENT '补测回流按 ASIN 定位未回流行'
+) ENGINE = InnoDB DEFAULT CHARSET = utf8mb4 COLLATE = utf8mb4_general_ci ROW_FORMAT = DYNAMIC
+ COMMENT = '采集数据被过滤行明细(四类过滤逐行带原因,供结果 Excel 追溯与对账)';
diff --git a/backend-java/src/main/resources/mapper/collectdata/CollectDataFilteredRowMapper.xml b/backend-java/src/main/resources/mapper/collectdata/CollectDataFilteredRowMapper.xml
new file mode 100644
index 000000000..58f84a13c
--- /dev/null
+++ b/backend-java/src/main/resources/mapper/collectdata/CollectDataFilteredRowMapper.xml
@@ -0,0 +1,32 @@
+
+
+ 10-28 采集「被过滤行」追溯(V179,结果 Excel 第 3 个 sheet)抬 total 1028 -> 1031,逐处记名。
+ * 三枚都是审计件不得打掉交付那一形,块体各只有一句带 {@code [collect-data]} 锚点的 WARN:
+ * 背景:task 32573 回传 600 行、结果只剩 1 行,被砍的 599 行在系统里只有四个计数,
+ * 谁也无法回答"到底是哪些 ASIN、死在哪一道"。这张 sheet 就是为了让它可追溯。 三件事必须钉住:① 它排在最后(外部读取方要么只认 sheet 0、要么按 sheet 名认列);
+ * ② 它的行数**绝不能**累进 writeWorkbookStreaming 的返回值——调用方拿那个数写
+ * stats.finalRowCount 与 result.rowCount,并据此跑 guardDetailRowsWritten 校验
+ * 「下载 Excel 行数 == 保留行数」,一旦污染就等于把交付说谎重新引回来;
+ * ③ 一行被过滤都没有时不建这张表,只留表头会让人误以为"全被过滤了"。 锁定契约:文案只陈述可观测事实(缺失行数、不计入结果条数、未判无效、可重跑),
- * 不指认代码没有依据的具体故障原因;真实原因留在 BrandCheckClient 的 warn 日志里
- * (brand-check 的 lastErr 字段),由运维按日志排查。
+ * 锁定契约(2026-10-28 修订):文案只陈述可观测事实(缺失行数、不计入结果条数、未判无效),
+ * **并且只在这层真的观测得到原因时才指名原因**——品牌检测的失败原因现已由
+ * {@code BrandBatchOutcome#reason} 落进 result_json(brandQueryFailedReason),所以
+ * 「用户未配置代理」这类有据可查的原因必须写出来并指向「密钥设置→代理设置」。
+ *
+ * 原契约(10-02)之所以禁止指名原因,是因为当时这一层确实观测不到;不是"永远不许说"。
+ * 反过来,原因缺失时仍然不猜:不写「超时」「限流」「上游不可用」,只说系统会自动重试。
+ * 旧文案那句「可稍后单独重跑这些 ASIN」是必须拆掉的误导——未配置代理时重跑一万次也不会变好
+ * (现网 task 32573:补测队列两轮 21/21、28/28 全败、净增 0 行)。
*/
@ExtendWith(MockitoExtension.class)
class CollectDataPartialResultNoticeTest {
@@ -77,13 +83,19 @@ class CollectDataPartialResultNoticeTest {
}
private FileTaskEntity taskWithQueryFailedRows(int queryFailedRows) {
+ return taskWithQueryFailedRows(queryFailedRows, null);
+ }
+
+ private FileTaskEntity taskWithQueryFailedRows(int queryFailedRows, String reason) {
FileTaskEntity task = new FileTaskEntity();
task.setId(TASK_ID);
task.setModuleType("COLLECT_DATA");
task.setStatus("RUNNING");
task.setTaskNo("CD-29991");
task.setUserId(7L);
- task.setResultJson("{\"brandQueryFailedCount\":" + queryFailedRows + "}");
+ task.setResultJson("{\"brandQueryFailedCount\":" + queryFailedRows
+ + (reason == null ? "" : ",\"brandQueryFailedReason\":\"" + reason + "\"")
+ + "}");
return task;
}
@@ -103,9 +115,13 @@ class CollectDataPartialResultNoticeTest {
}
private String stubHappyPathAndCaptureDetail() {
- when(fileTaskMapper.selectById(TASK_ID)).thenReturn(taskWithQueryFailedRows(7));
+ return stubHappyPathAndCaptureDetail(null);
+ }
+
+ private String stubHappyPathAndCaptureDetail(String reason) {
+ when(fileTaskMapper.selectById(TASK_ID)).thenReturn(taskWithQueryFailedRows(7, reason));
when(fileResultMapper.selectById(5001L)).thenReturn(result());
- when(excelAssemblyService.writeWorkbookStreaming(any(), any(), any(), any())).thenReturn(12L);
+ when(excelAssemblyService.writeWorkbookStreaming(any(), any(), any(), any(), any())).thenReturn(12L);
// 空明细守卫按明细表行数校验:happy path 下表内行数与写出行数一致。
when(taskResultItemMapper.selectCount(any())).thenReturn(12L);
when(ossStorageService.uploadResultFile(any(), anyString())).thenReturn("collect-data/29991.xlsx");
@@ -136,14 +152,35 @@ class CollectDataPartialResultNoticeTest {
assertThat(detail).contains("不计入结果条数");
assertThat(detail).contains("未返回结果");
- assertThat(detail).contains("重跑");
+ // 原因缺失时不再承诺「你重跑就好」,只说系统会自动重试
+ assertThat(detail).contains("自动重试");
+ }
+
+ @Test
+ void 原因缺失时不猜故障方() {
+ // classify(空串) 会算成「上游」,那是没观测到的事——这一层不能替它背书
+ String detail = stubHappyPathAndCaptureDetail();
+
+ assertThat(detail).doesNotContain("上游");
+ assertThat(detail).doesNotContain("未配置代理");
+ }
+
+ @Test
+ void 未配置代理必须指名原因并指向可行动入口() {
+ // 旧文案对配置类原因是谎:补测队列两轮 21/21、28/28 全败、净增 0 行,
+ // 让用户「稍后单独重跑」只会让他白等
+ String detail = stubHappyPathAndCaptureDetail("用户未配置代理");
+
+ assertThat(detail).contains("密钥设置");
+ assertThat(detail).contains("代理设置");
+ assertThat(detail).doesNotContain("可稍后单独重跑这些 ASIN");
}
@Test
void 任务已被判失败时不推部分结果通知() {
when(fileTaskMapper.selectById(TASK_ID)).thenReturn(taskWithQueryFailedRows(7));
when(fileResultMapper.selectById(5001L)).thenReturn(result());
- when(excelAssemblyService.writeWorkbookStreaming(any(), any(), any(), any())).thenReturn(12L);
+ when(excelAssemblyService.writeWorkbookStreaming(any(), any(), any(), any(), any())).thenReturn(12L);
// 空明细守卫按明细表行数校验:happy path 下表内行数与写出行数一致。
when(taskResultItemMapper.selectCount(any())).thenReturn(12L);
when(ossStorageService.uploadResultFile(any(), anyString())).thenReturn("collect-data/29991.xlsx");
@@ -160,7 +197,7 @@ class CollectDataPartialResultNoticeTest {
void 明细缺行时不得上传文件并把任务标成功() {
when(fileTaskMapper.selectById(TASK_ID)).thenReturn(taskWithQueryFailedRows(0));
when(fileResultMapper.selectById(5001L)).thenReturn(result());
- when(excelAssemblyService.writeWorkbookStreaming(any(), any(), any(), any())).thenReturn(12L);
+ when(excelAssemblyService.writeWorkbookStreaming(any(), any(), any(), any(), any())).thenReturn(12L);
when(taskResultItemMapper.selectCount(any())).thenReturn(20L);
assertThatThrownBy(() -> service.processResultFileJob(job()))
@@ -182,7 +219,7 @@ class CollectDataPartialResultNoticeTest {
task.setResultJson("{\"finalRowCount\":93}");
when(fileTaskMapper.selectById(TASK_ID)).thenReturn(task);
when(fileResultMapper.selectById(5001L)).thenReturn(result());
- when(excelAssemblyService.writeWorkbookStreaming(any(), any(), any(), any())).thenReturn(0L);
+ when(excelAssemblyService.writeWorkbookStreaming(any(), any(), any(), any(), any())).thenReturn(0L);
when(taskResultItemMapper.selectCount(any())).thenReturn(0L);
assertThatThrownBy(() -> service.processResultFileJob(job()))
@@ -199,7 +236,7 @@ class CollectDataPartialResultNoticeTest {
task.setResultJson("{}");
when(fileTaskMapper.selectById(TASK_ID)).thenReturn(task);
when(fileResultMapper.selectById(5001L)).thenReturn(result());
- when(excelAssemblyService.writeWorkbookStreaming(any(), any(), any(), any())).thenReturn(0L);
+ when(excelAssemblyService.writeWorkbookStreaming(any(), any(), any(), any(), any())).thenReturn(0L);
when(taskResultItemMapper.selectCount(any())).thenReturn(0L);
when(ossStorageService.uploadResultFile(any(), anyString())).thenReturn("collect-data/29991.xlsx");
when(fileTaskMapper.update(isNull(), any())).thenReturn(1);
diff --git a/backend-java/src/test/java/com/nanri/aiimage/modules/collectdata/service/CollectDataResultFileReuploadCompatTest.java b/backend-java/src/test/java/com/nanri/aiimage/modules/collectdata/service/CollectDataResultFileReuploadCompatTest.java
new file mode 100644
index 000000000..232b2bf7d
--- /dev/null
+++ b/backend-java/src/test/java/com/nanri/aiimage/modules/collectdata/service/CollectDataResultFileReuploadCompatTest.java
@@ -0,0 +1,114 @@
+package com.nanri.aiimage.modules.collectdata.service;
+
+import com.nanri.aiimage.common.exception.BusinessException;
+import com.nanri.aiimage.modules.collectdata.model.dto.CollectDataSummaryRowDto;
+import com.nanri.aiimage.modules.collectdata.model.entity.CollectDataFilteredRowEntity;
+import com.nanri.aiimage.modules.collectdata.model.vo.CollectDataResultRowVo;
+import com.nanri.aiimage.modules.publish.service.PublishWorkbookService;
+import org.junit.jupiter.api.Test;
+import org.junit.jupiter.api.io.TempDir;
+
+import java.io.File;
+import java.nio.file.Path;
+import java.util.List;
+
+import static org.assertj.core.api.Assertions.assertThat;
+
+/**
+ * 采集结果文件被用户**回传**给别的模块时的兼容性钉子(V179 新增尾 sheet 的主要外溢风险)。
+ *
+ * 结果 Excel 不是终点物:用户会把它再上传到上架/跟价。那两个模块用的是
+ * {@code ExcelStreamReader.readAllSheets}(读全部 sheet),所以多一张 sheet 就可能多算一批行。
+ * 本用例钉住「加与不加那张 sheet,回传行为逐字一致」。 跟价那一侧({@code PriceTrackTaskService.parseWorkbookAsinRows})的机制是
+ * {@code resolveCountryCode(sheetName)} 认不出「被过滤明细」即整表跳过,且首表兜底只在
+ * sheetCount==1 时生效——那条路径上的方法是私有的,本用例覆盖不到,按真机验证项处理。 它守的是两件相反的事,弄反任何一件都会把系统推回"数对不上又说不清":> consumer) {
+ if (taskId == null || consumer == null) {
+ return;
+ }
+ Long cursorId = null;
+ while (true) {
+ LambdaQueryWrapper
+ *
*/
- static final int STRICT_TOTAL = 1028;
+ static final int STRICT_TOTAL = 1031;
/**
* 重抛型 catch 的**下限**(不是上限):本轮把两处吞异常改成重抛之后实测 388。
* 这一格只能涨不能跌——它是"异常会往外走"的总量,把它钉成"只许减"就等于拦着别人修吞异常。
@@ -282,8 +300,13 @@ class CatchSilenceRatchetTest {
* 10-27 同日第二批再抬 577 -> 579(2 枚逐批/逐条目容错:{@code QueryAsinService.flushInsert}
* 与 {@code DedupeRunService.packageFolderDedupeResultsAsZip},块体都是 log.warn+计数,
* 坏项点名、其余照常交付,逐处记名见 STRICT_TOTAL 上方注释)。
+ * 10-28 再抬 579 -> 582(V179 采集「被过滤行」三枚审计件容错:{@code CollectDataFilteredRowStore.writeBatch}
+ * 逐批插入、{@code CollectDataFilteredRowStore.markResolved} 回流簿记、
+ * {@code CollectDataExcelAssemblyService.writeFilteredSheet} 尾 sheet 降级;三处都是带
+ * {@code [collect-data]} 锚点的 log.warn+善后,块体除第一枚外各带一句状态置位/返回值,
+ * 逐处记名见 STRICT_TOTAL 上方注释)。
*/
- static final int STRICT_NOT_RETHROW = 579;
+ static final int STRICT_NOT_RETHROW = 582;
/**
* 10-17 抬 172 -> 173,出处就是上面点名过的 {@code TaskHeartbeatLivenessService.stamp()}:
* 块体只有「取到 key 名 + log.warn」一句,写心跳证据失败时既不重试也不抛出——
@@ -303,8 +326,12 @@ class CatchSilenceRatchetTest {
* 与 {@code getAllFileAggregates} 的逐条跳过出口——块体各只有一句带 {@code [brand-storage]}
* 锚点的 WARN。两处此前是「一个文件坏 ⇒ 整批抛出」,改后其余文件照常返回,坏的那条点名可查;
* 同批 {@code rebuildAggregate} 与外观专利候选收集因块内另有计数/continue 只进 NOT_RETHROW。
+ * 10-28 抬 181 -> 182(V179 采集「被过滤行」的 {@code CollectDataFilteredRowStore.writeBatch}):
+ * 块体只有一句带 {@code [collect-data]} 锚点的 WARN(taskId+chunk+err),坏批点名后其余批照常
+ * 落库。同批另两处({@code markResolved} 追加 return、{@code writeFilteredSheet} 置 incomplete 位)
+ * 因块内多一句只进 NOT_RETHROW。
*/
- static final int STRICT_LOG_ONLY = 181; // 严格口径:块内每条顶层语句都是 log 调用才算
+ static final int STRICT_LOG_ONLY = 182; // 严格口径:块内每条顶层语句都是 log 调用才算
static final int STRICT_EMPTY = 56;
/**
diff --git a/backend-java/src/test/java/com/nanri/aiimage/modules/collectdata/service/CollectDataBrandRecheckInjectorTest.java b/backend-java/src/test/java/com/nanri/aiimage/modules/collectdata/service/CollectDataBrandRecheckInjectorTest.java
index 19c8ee677..0222c56c5 100644
--- a/backend-java/src/test/java/com/nanri/aiimage/modules/collectdata/service/CollectDataBrandRecheckInjectorTest.java
+++ b/backend-java/src/test/java/com/nanri/aiimage/modules/collectdata/service/CollectDataBrandRecheckInjectorTest.java
@@ -60,6 +60,8 @@ class CollectDataBrandRecheckInjectorTest {
private AsinResultCacheService asinResultCacheService;
@Mock
private CollectDataService collectDataService;
+ @Mock
+ private com.nanri.aiimage.modules.collectdata.util.CollectDataFilteredRowStore filteredRowStore;
private final CollectDataResultDetailCodec codec = new CollectDataResultDetailCodec(new ObjectMapper());
private CollectDataBrandRecheckInjector injector;
@@ -72,7 +74,8 @@ class CollectDataBrandRecheckInjectorTest {
TaskChunkEntity.class);
injector = new CollectDataBrandRecheckInjector(taskChunkMapper, fileResultMapper,
collectDataBatchQuery, codec, resultItemBatchWriter, invalidAsinBatchWriter,
- transientPayloadStorageService, asinResultCacheService, new ObjectMapper(), collectDataService);
+ transientPayloadStorageService, asinResultCacheService, new ObjectMapper(), collectDataService,
+ filteredRowStore);
}
private FileTaskEntity task(long id) {
@@ -120,7 +123,7 @@ class CollectDataBrandRecheckInjectorTest {
when(transientPayloadStorageService.resolvePayload(eq("ptr1"), anyString())).thenReturn(payload);
CollectDataResultRowVo kept = row("B01", "Nike");
when(collectDataBatchQuery.filter(anyList(), any())).thenReturn(
- new CollectDataBatchQuery.FilterResult(List.of(kept), 1, 0));
+ new CollectDataBatchQuery.FilterResult(List.of(kept), 1, 0, List.of()));
when(transientPayloadStorageService.storeResultPayload(eq("COLLECT_DATA"), eq(100L), eq("hash1"),
eq("recheck-chunk-1"), anyString())).thenReturn("ptr2");
when(resultItemBatchWriter.upsertAccepted(eq(100L), eq(9L), eq("s1"), eq(1),
@@ -133,6 +136,9 @@ class CollectDataBrandRecheckInjectorTest {
assertThat(outcome.acceptedRows()).isEqualTo(1);
assertThat(outcome.rejectedRows()).isZero();
verify(asinResultCacheService).writeCollectDataRows(eq("US"), eq("100"), eq(List.of(kept)));
+ // 补回结果的行要在 V179 明细里标 RESOLVED——但只标不删:删掉就把「当时确实没查动」
+ // 这条历史一起删了,而它是用户日后追问"我那批货当时为什么少"时唯一说得清的东西。
+ verify(filteredRowStore).markResolved(eq(100L), eq(List.of("B01")));
}
@Test
@@ -180,7 +186,7 @@ class CollectDataBrandRecheckInjectorTest {
"[{\"asin\":\"B01\",\"brand\":\"Nike\"}]");
// 补测窗口内该 ASIN 已被其他任务采走:复查全滤掉
when(collectDataBatchQuery.filter(anyList(), any())).thenReturn(
- new CollectDataBatchQuery.FilterResult(List.of(), 1, 0));
+ new CollectDataBatchQuery.FilterResult(List.of(), 1, 0, List.of()));
GroupOutcome outcome = injector.apply(task(100L), Map.of("nike", RecheckVerdict.CLEAN));
diff --git a/backend-java/src/test/java/com/nanri/aiimage/modules/collectdata/service/CollectDataFilteredSheetTest.java b/backend-java/src/test/java/com/nanri/aiimage/modules/collectdata/service/CollectDataFilteredSheetTest.java
new file mode 100644
index 000000000..1a58b2624
--- /dev/null
+++ b/backend-java/src/test/java/com/nanri/aiimage/modules/collectdata/service/CollectDataFilteredSheetTest.java
@@ -0,0 +1,182 @@
+package com.nanri.aiimage.modules.collectdata.service;
+
+import com.nanri.aiimage.modules.collectdata.model.dto.CollectDataSummaryRowDto;
+import com.nanri.aiimage.modules.collectdata.model.entity.CollectDataFilteredRowEntity;
+import com.nanri.aiimage.modules.collectdata.model.vo.CollectDataResultRowVo;
+import org.apache.poi.ss.usermodel.Sheet;
+import org.apache.poi.ss.usermodel.Workbook;
+import org.apache.poi.ss.usermodel.WorkbookFactory;
+import org.junit.jupiter.api.Test;
+import org.junit.jupiter.api.io.TempDir;
+
+import java.io.File;
+import java.nio.file.Path;
+import java.util.List;
+import java.util.function.Consumer;
+
+import static org.assertj.core.api.Assertions.assertThat;
+
+/**
+ * 「被过滤明细」尾 sheet(V179)的形状钉子。
+ *
+ * > batchConsumer) {
+ if (!rows.isEmpty()) {
+ batchConsumer.accept(rows);
+ }
+ }
+
+ @Override
+ public long notRecordedCount() {
+ return notRecorded;
+ }
+ };
+ }
+
+ @Test
+ void 尾sheet排在最后且明细行数不被污染() throws Exception {
+ File output = tempDir.resolve("filtered.xlsx").toFile();
+
+ long returned = assembly.writeWorkbookStreaming(
+ output,
+ consumer -> consumer.accept(List.of(detailRow("B000000001"), detailRow("B000000002"))),
+ List.of(summary()),
+ consumer -> { /* 本例不走 fallback 聚合 */ },
+ sourceOf(List.of(
+ filtered("B000000003", CollectDataFilteredRowEntity.REASON_DEDUPE, null),
+ filtered("B000000004", CollectDataFilteredRowEntity.REASON_BRAND_QUERY_FAILED, "用户未配置代理")),
+ 0L));
+
+ // 红线:返回值只算「采集数据结果」那 2 行,被过滤的 2 行一个都不能进来
+ assertThat(returned).isEqualTo(2L);
+
+ try (Workbook workbook = WorkbookFactory.create(output)) {
+ assertThat(workbook.getNumberOfSheets()).isEqualTo(3);
+ assertThat(workbook.getSheetName(0)).isEqualTo("采集数据结果");
+ assertThat(workbook.getSheetName(1)).isEqualTo("结果文件");
+ assertThat(workbook.getSheetName(2)).isEqualTo("被过滤明细");
+ Sheet sheet = workbook.getSheet("被过滤明细");
+ // 表头 + 2 行
+ assertThat(sheet.getLastRowNum()).isEqualTo(2);
+ assertThat(sheet.getRow(1).getCell(1).getStringCellValue()).isEqualTo("B000000003");
+ assertThat(sheet.getRow(1).getCell(3).getStringCellValue()).isEqualTo("已存在于去重总数据库");
+ assertThat(sheet.getRow(2).getCell(3).getStringCellValue()).isEqualTo("品牌检测未返回结果");
+ assertThat(sheet.getRow(2).getCell(4).getStringCellValue()).isEqualTo("用户未配置代理");
+ }
+ }
+
+ @Test
+ void 没有被过滤行时不建这张表() throws Exception {
+ File output = tempDir.resolve("empty.xlsx").toFile();
+
+ assembly.writeWorkbookStreaming(
+ output,
+ consumer -> consumer.accept(List.of(detailRow("B000000001"))),
+ List.of(summary()),
+ consumer -> { },
+ sourceOf(List.of(), 0L));
+
+ try (Workbook workbook = WorkbookFactory.create(output)) {
+ assertThat(workbook.getNumberOfSheets()).isEqualTo(2);
+ assertThat(workbook.getSheet("被过滤明细")).isNull();
+ }
+ }
+
+ @Test
+ void 超出上限未逐条记录时如实写出缺口() throws Exception {
+ File output = tempDir.resolve("gap.xlsx").toFile();
+
+ assembly.writeWorkbookStreaming(
+ output,
+ consumer -> consumer.accept(List.of(detailRow("B000000001"))),
+ List.of(summary()),
+ consumer -> { },
+ sourceOf(List.of(filtered("B000000009",
+ CollectDataFilteredRowEntity.REASON_INVALID_BRAND, null)), 317L));
+
+ try (Workbook workbook = WorkbookFactory.create(output)) {
+ Sheet sheet = workbook.getSheet("被过滤明细");
+ assertThat(sheet).isNotNull();
+ String note = sheet.getRow(2).getCell(4).getStringCellValue();
+ assertThat(note).contains("317");
+ // 缺口必须说清"行数已计入各道闸的计数",否则读的人会以为又丢了 317 行
+ assertThat(note).contains("已计入");
+ }
+ }
+
+ @Test
+ void 数据源抛错时降级掉这个sheet而不是打掉整份交付() throws Exception {
+ File output = tempDir.resolve("boom.xlsx").toFile();
+ CollectDataExcelAssemblyService.FilteredRowSource broken = new CollectDataExcelAssemblyService.FilteredRowSource() {
+ @Override
+ public void forEachBatch(Consumer
> batchConsumer) {
+ throw new IllegalStateException("明细表读不动");
+ }
+ };
+
+ long returned = assembly.writeWorkbookStreaming(
+ output,
+ consumer -> consumer.accept(List.of(detailRow("B000000001"))),
+ List.of(summary()),
+ consumer -> { },
+ broken);
+
+ assertThat(returned).isEqualTo(1L);
+ try (Workbook workbook = WorkbookFactory.create(output)) {
+ // 追溯 sheet 缺了就缺了,前两表照常交付——它是审计件,不是交付物。
+ // 若这里改成外抛,一次审计读失败就会把用户已采回的数据一起打掉。
+ assertThat(workbook.getNumberOfSheets()).isEqualTo(2);
+ assertThat(workbook.getSheetName(0)).isEqualTo("采集数据结果");
+ assertThat(workbook.getSheetName(1)).isEqualTo("结果文件");
+ }
+ }
+}
diff --git a/backend-java/src/test/java/com/nanri/aiimage/modules/collectdata/service/CollectDataPartialResultNoticeTest.java b/backend-java/src/test/java/com/nanri/aiimage/modules/collectdata/service/CollectDataPartialResultNoticeTest.java
index 7fdcd7170..f1760e4bf 100644
--- a/backend-java/src/test/java/com/nanri/aiimage/modules/collectdata/service/CollectDataPartialResultNoticeTest.java
+++ b/backend-java/src/test/java/com/nanri/aiimage/modules/collectdata/service/CollectDataPartialResultNoticeTest.java
@@ -43,9 +43,15 @@ import static org.mockito.Mockito.when;
* 上游的采集代理取不到 IP(余额不足)与 503 —— 代码在这个位置只观测到「品牌检测没返回结果」,
* 观测不到原因。写死一个猜测的原因会把排查方向从代理/上游带到自身服务上。
*
- *
> consumer) {
+ consumer.accept(List.of(entity));
+ }
+
+ @Override
+ public long notRecordedCount() {
+ return 0L;
+ }
+ };
+ }
+
+ /** 回传解析的落点:拒绝就说清第几列不匹配,接受就说 "accepted"。 */
+ private String outcomeOf(File file) {
+ try {
+ PublishWorkbookService.ParsedWorkbook parsed = publishParser.parse(file);
+ return "accepted rows=" + (parsed == null ? "-" : parsed.rows().size());
+ } catch (BusinessException ex) {
+ return "rejected: " + ex.getMessage();
+ }
+ }
+
+ @Test
+ void 多出被过滤sheet后上架解析器的回传行为逐字一致() throws Exception {
+ String twoSheets = outcomeOf(assemble(false));
+ String threeSheets = outcomeOf(assemble(true));
+
+ assertThat(threeSheets).isEqualTo(twoSheets);
+ // 并且必须是在**第一张表**就被表头校验挡下——它挡在前面,才轮不到新 sheet 出问题
+ assertThat(twoSheets).startsWith("rejected: ").contains("表头不匹配").contains("第 1 列");
+ }
+}
diff --git a/backend-java/src/test/java/com/nanri/aiimage/modules/collectdata/service/CollectDataServiceTxBoundaryTest.java b/backend-java/src/test/java/com/nanri/aiimage/modules/collectdata/service/CollectDataServiceTxBoundaryTest.java
index a0edafabe..d2918057f 100644
--- a/backend-java/src/test/java/com/nanri/aiimage/modules/collectdata/service/CollectDataServiceTxBoundaryTest.java
+++ b/backend-java/src/test/java/com/nanri/aiimage/modules/collectdata/service/CollectDataServiceTxBoundaryTest.java
@@ -142,7 +142,7 @@ class CollectDataServiceTxBoundaryTest {
CollectDataResultRowVo row = new CollectDataResultRowVo();
row.setAsin("B0COLLECT1");
lenient().when(collectDataBatchQuery.filter(any(), any())).thenReturn(
- new CollectDataBatchQuery.FilterResult(List.of(row), 0, 0));
+ new CollectDataBatchQuery.FilterResult(List.of(row), 0, 0, List.of()));
lenient().when(brandBatchFilter.filter(any(), any())).thenReturn(
new CollectDataBrandBatchFilter.BrandBatchOutcome(List.of(), List.of(), List.of(row), ""));
lenient().when(resultItemBatchWriter.upsertAccepted(anyLong(), anyLong(), anyString(), anyInt(), any(), anyString()))
diff --git a/backend-java/src/test/java/com/nanri/aiimage/modules/collectdata/service/support/CollectDataSubmitResultSnapshotTest.java b/backend-java/src/test/java/com/nanri/aiimage/modules/collectdata/service/support/CollectDataSubmitResultSnapshotTest.java
index 4c815ed05..3a0009174 100644
--- a/backend-java/src/test/java/com/nanri/aiimage/modules/collectdata/service/support/CollectDataSubmitResultSnapshotTest.java
+++ b/backend-java/src/test/java/com/nanri/aiimage/modules/collectdata/service/support/CollectDataSubmitResultSnapshotTest.java
@@ -335,7 +335,7 @@ class CollectDataSubmitResultSnapshotTest {
List
+ *
+ */
+@ExtendWith(MockitoExtension.class)
+@MockitoSettings(strictness = Strictness.LENIENT)
+class CollectDataFilteredRowStoreTest {
+
+ @Mock
+ private CollectDataFilteredRowMapper mapper;
+
+ private CollectDataFilteredRowStore store(int batchSize) {
+ return new CollectDataFilteredRowStore(mapper, batchSize);
+ }
+
+ private static CollectDataResultRowVo row(String asin, String brand, String keyword) {
+ CollectDataResultRowVo row = new CollectDataResultRowVo();
+ row.setAsin(asin);
+ row.setBrand(brand);
+ row.setKeyword(keyword);
+ return row;
+ }
+
+ @SuppressWarnings("unchecked")
+ private List> captor = ArgumentCaptor.forClass(List.class);
+ verify(mapper, times(expectedCalls)).insertBatch(captor.capture());
+ return captor.getAllValues().stream().flatMap(List::stream).toList();
+ }
+
+ @Test
+ void 三道闸的行各归各的reason_code() {
+ when(mapper.insertBatch(any())).thenAnswer(inv -> ((Collection>) inv.getArgument(0)).size());
+ List