task-6: 分组数据改为索引/范围引用，避免 groups 嵌套复制完整行对象

buildParsedGroups 只写 [startIndex, endIndex) 半开区间引用（同 baseId 行在 items 中天然连续），行对象仅存在于 items 一次；resolveAllRows 优先按引用展开并安全跳过越界/非法区间；响应组仍内嵌预览行兼容前端。旧 payload 内嵌 items 格式继续可读。新增 8 个测试覆盖引用写入、批量、幂等、空输入、单组、越界 clamp、非法区间、依赖失败恢复。
2026-08-29 15:02:15 +08:00
parent 16bec1c0fa
commit db6e8f0ce8
3 changed files with 468 additions and 5 deletions
@@ -1,5 +1,6 @@
 package com.nanri.aiimage.modules.similarasin.model.vo;

+import com.fasterxml.jackson.annotation.JsonInclude;
 import io.swagger.v3.oas.annotations.media.Schema;
 import lombok.Data;

@@ -27,6 +28,21 @@ public class SimilarAsinParsedGroupVo {
    @Schema(description = "分组内行数")
    private Integer itemCount;

-    @Schema(description = "分组内全部行，顺序与原 Excel 保持一致")
+    /**
+     * 组内行在载荷 items 中的起始下标（含）。写入侧只输出引用，
+     * 行对象仅存在于 items 一次，避免 groups 嵌套复制完整行对象。
+     */
+    @Schema(description = "组内行在 items 中的起始下标（含）")
+    private Integer startIndex;
+
+    /**
+     * 组内行在载荷 items 中的结束下标（不含），半开区间 [startIndex, endIndex)。
+     */
+    @Schema(description = "组内行在 items 中的结束下标（不含），半开区间 [startIndex, endIndex)")
+    private Integer endIndex;
+
+    /** 兼容字段：旧 payload 内嵌的完整行。新写入不再输出，仅读取旧 payload 时使用。 */
+    @JsonInclude(JsonInclude.Include.NON_EMPTY)
+    @Schema(description = "兼容旧链路字段：旧 payload 内嵌的完整行，新写入不再输出", hidden = true)
    private List<SimilarAsinParsedRowVo> items = new ArrayList<>();
 }
@@ -511,7 +511,7 @@ public class SimilarAsinTaskService {
        vo.setImgSwitch(Boolean.TRUE.equals(request.getImgSwitch()));
        vo.setCategorySwitch(Boolean.TRUE.equals(request.getCategorySwitch()));
        vo.setItems(buildResponsePreviewRows(allRows));
-        vo.setGroups(buildResponsePreviewGroups(groups));
+        vo.setGroups(buildResponsePreviewGroups(groups, allRows));
        long finishedAt = System.nanoTime();
        log.info("[similar-asin] parse timing taskId={} files={} rows={} groups={} totalMs={} parseMs={} groupMs={} taskInsertMs={} payloadJsonMs={} payloadStoreMs={} persistMs={} responseMs={}",
                task.getId(),
@@ -963,6 +963,10 @@ public class SimilarAsinTaskService {
        if (payload == null) {
            return List.of();
        }
+        if (containsGroupRefs(payload)) {
+            // 新格式：分组携带索引引用，信任引用展开结果（越界/非法区间安全跳过）
+            return expandGroupRefs(payload);
+        }
        List<SimilarAsinParsedRowVo> rows = payload.getItems();
        if (rows == null || rows.isEmpty()) {
            rows = payload.getAllItems();
@@ -976,6 +980,48 @@ public class SimilarAsinTaskService {
        return rows == null ? List.of() : rows;
    }

+    private static boolean containsGroupRefs(SimilarAsinParsedPayloadDto payload) {
+        if (payload.getGroups() == null) {
+            return false;
+        }
+        for (SimilarAsinParsedGroupVo group : payload.getGroups()) {
+            if (group != null && (group.getStartIndex() != null || group.getEndIndex() != null)) {
+                return true;
+            }
+        }
+        return false;
+    }
+
+    /**
+     * Task 6：按分组引用 [startIndex, endIndex) 从 items 展开组内行。
+     * 引用越界或区间非法时安全跳过，不抛异常；展开不修改 payload 内部状态。
+     */
+    static List<SimilarAsinParsedRowVo> expandGroupRefs(SimilarAsinParsedPayloadDto payload) {
+        if (payload == null || payload.getGroups() == null || payload.getGroups().isEmpty()) {
+            return List.of();
+        }
+        List<SimilarAsinParsedRowVo> rows = payload.getItems();
+        if (rows == null || rows.isEmpty()) {
+            rows = payload.getAllItems();
+        }
+        if (rows == null || rows.isEmpty()) {
+            return List.of();
+        }
+        List<SimilarAsinParsedRowVo> expanded = new ArrayList<>();
+        for (SimilarAsinParsedGroupVo group : payload.getGroups()) {
+            if (group == null) {
+                continue;
+            }
+            int start = group.getStartIndex() == null ? 0 : group.getStartIndex();
+            int end = group.getEndIndex() == null ? 0 : group.getEndIndex();
+            if (start < 0 || end <= start || end > rows.size()) {
+                continue;
+            }
+            expanded.addAll(rows.subList(start, end));
+        }
+        return expanded;
+    }
+
    private PersistSubmittedChunkResult persistSubmittedChunk(PreparedSubmittedChunk prepared) {
        Long taskId = prepared.taskId();
        FileTaskEntity task = fileTaskMapper.selectById(taskId);
@@ -1461,8 +1507,15 @@ public class SimilarAsinTaskService {
        return candidates;
    }

+    /**
+     * Task 6：分组数据改为索引/范围引用。组内行在原 items 中必然连续
+     * （parseWorkbook 按行遍历、同 baseId 块连续收集），因此只需要
+     * [startIndex, endIndex) 半开区间即可唯一定位组内行，
+     * 行对象仅存在于 items 一次，避免 groups 嵌套复制完整行对象。
+     */
    private List<SimilarAsinParsedGroupVo> buildParsedGroups(List<SimilarAsinParsedRowVo> rows) {
        List<SimilarAsinParsedGroupVo> groups = new ArrayList<>();
+        int cursor = 0;
        for (List<SimilarAsinParsedRowVo> siblings : groupRowsByBaseId(rows).values()) {
            if (siblings == null || siblings.isEmpty()) {
                continue;
@@ -1475,13 +1528,16 @@ public class SimilarAsinTaskService {
            group.setBaseId(baseId(first.getDisplayId()));
            group.setDisplayId(firstNonBlank(first.getDisplayId(), first.getSourceId()));
            group.setItemCount(siblings.size());
-            group.setItems(new ArrayList<>(siblings));
+            int end = cursor + siblings.size();
+            group.setStartIndex(cursor);
+            group.setEndIndex(end);
+            cursor = end;
            groups.add(group);
        }
        return groups;
    }

-    private List<SimilarAsinParsedGroupVo> buildResponsePreviewGroups(List<SimilarAsinParsedGroupVo> groups) {
+    private List<SimilarAsinParsedGroupVo> buildResponsePreviewGroups(List<SimilarAsinParsedGroupVo> groups, List<SimilarAsinParsedRowVo> allRows) {
        if (groups == null || groups.isEmpty()) {
            return List.of();
        }
@@ -1499,7 +1555,18 @@ public class SimilarAsinTaskService {
            copy.setBaseId(group.getBaseId());
            copy.setDisplayId(group.getDisplayId());
            copy.setItemCount(group.getItemCount());
-            copy.setItems(buildResponsePreviewRows(group.getItems()));
+            copy.setStartIndex(group.getStartIndex());
+            copy.setEndIndex(group.getEndIndex());
+            // 响应组携带内嵌预览行（切片自全量行），不暴露索引引用语义
+            int start = group.getStartIndex() == null ? 0 : group.getStartIndex();
+            int end = group.getEndIndex() == null ? 0 : group.getEndIndex();
+            if (start >= 0 && end > start && end <= allRows.size()) {
+                copy.setItems(buildResponsePreviewRows(allRows.subList(start, end)));
+            } else if (group.getItems() != null && !group.getItems().isEmpty()) {
+                copy.setItems(buildResponsePreviewRows(group.getItems()));
+            } else {
+                copy.setItems(List.of());
+            }
            preview.add(copy);
        }
        return preview;