From a14c1d87aacb3787dd31b09bcecdf270618697bc Mon Sep 17 00:00:00 2001
From: jw_mac <163wangjack@gmail.com>
Date: Thu, 25 Jun 2026 21:03:11 +0800
Subject: [PATCH] feat(knowledge): add batch reparse capability

This commit introduces the Batch Reparse feature for the Knowledge Base.
Users can now select multiple documents and trigger a re-parsing pipeline in batches, streamlining the recovery from parse failures.
---
 frontend/src/api/knowledge-base/index.ts      |  10 +-
 .../src/views/knowledge/KnowledgeBase.vue     |  36 ++++++-
 .../knowledge/components/DocumentBatchBar.vue |  24 ++++-
 internal/handler/knowledge.go                 | 102 ++++++++++++++++++
 internal/router/router.go                     |   1 +
 5 files changed, 168 insertions(+), 5 deletions(-)

diff --git a/frontend/src/api/knowledge-base/index.ts b/frontend/src/api/knowledge-base/index.ts
index f4e33459..e532d9b8 100644
--- a/frontend/src/api/knowledge-base/index.ts
+++ b/frontend/src/api/knowledge-base/index.ts
@@ -120,7 +120,7 @@ export function updateKnowledgeBase(id: string, data: {
     };
   }
 }) {
-  return put(`/api/v1/knowledge-bases/${id}` , data);
+  return put(`/api/v1/knowledge-bases/${id}`, data);
 }
 
 export function rebuildKBIndex(kbId: string) {
@@ -481,3 +481,11 @@ export function knowledgeSemanticSearch(data: {
 }) {
   return post('/api/v1/knowledge-search', data);
 }
+
+export function batchReparseKnowledge(kbId: string, ids: string[], processConfig?: any) {
+  return post(`/api/v1/knowledge/batch-reparse`, {
+    kb_id: kbId,
+    ids,
+    process_config: processConfig
+  });
+}
\ No newline at end of file
diff --git a/frontend/src/views/knowledge/KnowledgeBase.vue b/frontend/src/views/knowledge/KnowledgeBase.vue
index cdfe39a8..b1f6deb1 100644
--- a/frontend/src/views/knowledge/KnowledgeBase.vue
+++ b/frontend/src/views/knowledge/KnowledgeBase.vue
@@ -36,6 +36,7 @@ import {
   reparseKnowledge,
   cancelKnowledgeParse,
   batchDeleteKnowledge,
+  batchReparseKnowledge,
   getKnowledgeSpans,
   getKnowledgeDetails,
 } from "@/api/knowledge-base/index";
@@ -60,7 +61,7 @@ import { formatFileSize } from '@/utils/files';
 import { useMarqueeSelect } from '@/hooks/useMarqueeSelect';
 import type { ParserEngineInfo } from '@/api/system';
 const route = useRoute();
-const { t } = useI18n();
+const { t, te } = useI18n();
 const kbId = computed(() => (route.params as any).kbId as string || '');
 const kbInfo = ref<any>(null);
 const uploadSourceRef = ref<InstanceType<typeof KbUploadSourceDropdown> | null>(null);
@@ -415,6 +416,35 @@ watch(viewMode, (v) => {
 const selectedIds = ref<Set<string>>(new Set());
 let lastSelectedIndex = -1;
 const batchDeleting = ref(false);
+const batchReparsing = ref(false);
+
+const confirmBatchReparse = async () => {
+  if (batchReparsing.value || selectedIds.value.size === 0) return;
+  const ids = Array.from(selectedIds.value);
+  batchReparsing.value = true;
+  try {
+    const res: any = await batchReparseKnowledge(kbId.value, ids);
+    if (res?.success) {
+      if (res.partial && res.failed_ids && res.failed_ids.length > 0) {
+        MessagePlugin.warning(`部分重构任务提交成功，失败 ${res.failed_ids.length} 个`);
+      } else {
+        const successMsg = te('knowledgeBase.batchReparseSuccess') 
+          ? t('knowledgeBase.batchReparseSuccess', { count: ids.length }) 
+          : `已提交 ${ids.length} 个重构任务`;
+        MessagePlugin.success(successMsg);
+      }
+      clearSelection();
+      batchMode.value = false;
+      loadKnowledgeFiles(kbId.value);
+    } else {
+      MessagePlugin.error(res?.message || '批量重建失败');
+    }
+  } catch (e: any) {
+    MessagePlugin.error(e?.message || '批量重建失败');
+  } finally {
+    batchReparsing.value = false;
+  }
+};
 
 const selectedTagId = ref<string>('');
 const tagList = ref<any[]>([]);
@@ -2607,9 +2637,9 @@ async function createNewSession(value: string): Promise<void> {
                 </template>
               </div>
               <div class="doc-batch-bar-anchor" v-show="batchMode || selectedIds.size > 0">
-                <DocumentBatchBar :count="selectedIds.size" :loading="batchDeleting"
+                <DocumentBatchBar :count="selectedIds.size" :loading="batchDeleting || batchReparsing"
                   :visible="batchMode || selectedIds.size > 0" @cancel="handleBatchCancel"
-                  @delete="confirmBatchDelete" />
+                  @delete="confirmBatchDelete" @reparse="confirmBatchReparse" />
               </div>
             </div>
           </div>
diff --git a/frontend/src/views/knowledge/components/DocumentBatchBar.vue b/frontend/src/views/knowledge/components/DocumentBatchBar.vue
index 65096596..95dddf4f 100644
--- a/frontend/src/views/knowledge/components/DocumentBatchBar.vue
+++ b/frontend/src/views/knowledge/components/DocumentBatchBar.vue
@@ -12,9 +12,10 @@ defineProps<{
 const emit = defineEmits<{
   (e: 'cancel'): void;
   (e: 'delete'): void;
+  (e: 'reparse'): void;
 }>();
 
-const { t } = useI18n();
+const { t, te } = useI18n();
 </script>
 
 <template>
@@ -39,6 +40,27 @@ const { t } = useI18n();
           </t-button>
         </div>
         <div class="batch-bar-actions">
+          <t-popconfirm
+            theme="warning"
+            :content="te('knowledgeBase.overviewDescReparse') ? t('knowledgeBase.overviewDescReparse') : '确认后现有内容将被清除并重新解析该文档'"
+            :confirm-btn="{ content: te('knowledgeBase.confirmReparse') ? t('knowledgeBase.confirmReparse') : '确认并重新解析', theme: 'warning' }"
+            :cancel-btn="{ content: t('common.cancel') }"
+            placement="top"
+            @confirm="emit('reparse')"
+          >
+            <t-button
+              theme="default"
+              variant="outline"
+              size="small"
+              :disabled="count === 0"
+              :loading="loading"
+              @click.stop
+            >
+              <template #icon><t-icon name="refresh" size="14px" /></template>
+              {{ t('knowledgeBase.rebuildDocument') }}
+            </t-button>
+          </t-popconfirm>
+
           <t-popconfirm
             theme="warning"
             :content="t('knowledgeBase.confirmBatchDeleteDocument', { count })"
diff --git a/internal/handler/knowledge.go b/internal/handler/knowledge.go
index fac959b6..80ffa406 100644
--- a/internal/handler/knowledge.go
+++ b/internal/handler/knowledge.go
@@ -2228,3 +2228,105 @@ func sliceContains(ss []string, target string) bool {
 	}
 	return false
 }
+
+type batchReparseKnowledgeRequest struct {
+	KBID          string                           `json:"kb_id" binding:"required"`
+	IDs           []string                         `json:"ids" binding:"required"`
+	ProcessConfig *types.KnowledgeProcessOverrides `json:"process_config,omitempty"`
+}
+
+// BatchReparseKnowledge godoc
+// @Summary      批量重新解析知识
+// @Description  按 ID 列表批量重新解析单个知识库下的多个知识条目
+// @Tags         知识管理
+// @Accept       json
+// @Produce      json
+// @Param        request  body      batchReparseKnowledgeRequest  true  "批量重解析请求"
+// @Success      200      {object}  map[string]interface{}        "任务已提交"
+// @Failure      400      {object}  errors.AppError               "请求参数错误"
+// @Failure      403      {object}  errors.AppError               "权限不足"
+// @Security     Bearer
+// @Security     ApiKeyAuth
+// @Router       /knowledge/batch-reparse [post]
+func (h *KnowledgeHandler) BatchReparseKnowledge(c *gin.Context) {
+	ctx := c.Request.Context()
+	var req batchReparseKnowledgeRequest
+
+	if err := c.ShouldBindJSON(&req); err != nil {
+		logger.Errorf(ctx, "failed to parse batch reparse knowledge request: %v", err)
+		c.Error(errors.NewBadRequestError("invalid batch reparse knowledge request parameters"))
+		return
+	}
+
+	var ids []string
+	seen := make(map[string]struct{})
+	for _, id := range req.IDs {
+		if id != "" {
+			if _, ok := seen[id]; !ok {
+				ids = append(ids, id)
+				seen[id] = struct{}{}
+			}
+		}
+	}
+
+	if len(ids) == 0 {
+		c.Error(errors.NewBadRequestError("no knowledge IDs provided for batch reparse"))
+		return
+	}
+	if len(ids) > 200 {
+		c.Error(errors.NewBadRequestError("the number of IDs exceeds the maximum batch size of 200"))
+		return
+	}
+
+	_, _, effectiveTenantID, permission, err := h.validateKnowledgeBaseAccessWithKBID(c, req.KBID)
+	if err != nil {
+		c.Error(err)
+		return
+	}
+	if permission != types.OrgRoleAdmin && permission != types.OrgRoleEditor {
+		c.Error(errors.NewForbiddenError("no permission to reparse knowledge in this kb"))
+		return
+	}
+	ctx = context.WithValue(ctx, types.TenantIDContextKey, effectiveTenantID)
+	knows, err := h.kgService.GetKnowledgeBatch(ctx, effectiveTenantID, ids)
+	if err != nil {
+		logger.Errorf(ctx, "failed to get knowledge batch, kb_id: %s, size: %d, err: %v", req.KBID, len(ids), err)
+		c.Error(errors.NewInternalServerError("failed to get knowledge batch"))
+		return
+	}
+	if len(knows) != len(ids) {
+		c.Error(errors.NewBadRequestError("some knowledge entries were not found"))
+		return
+	}
+
+	for _, know := range knows {
+		if know.KnowledgeBaseID != req.KBID {
+			c.Error(errors.NewBadRequestError("some knowledge entries do not belong to the specified knowledge base"))
+			return
+		}
+	}
+	var failedIDs []string
+	for _, id := range ids {
+		_, err := h.kgService.ReparseKnowledge(ctx, id, req.ProcessConfig)
+		if err != nil {
+			logger.Errorf(ctx, "Failed to reparse knowledge %s: %v", id, err)
+			failedIDs = append(failedIDs, id)
+			continue
+		}
+	}
+
+	if len(failedIDs) > 0 {
+		c.JSON(200, gin.H{
+			"success":    len(failedIDs) < len(ids),
+			"partial":    len(failedIDs) < len(ids),
+			"message":    "Batch reparse completed with some failures",
+			"failed_ids": failedIDs,
+		})
+		return
+	}
+
+	c.JSON(200, gin.H{
+		"success": true,
+		"message": "Batch reparse task submitted successfully",
+	})
+}
diff --git a/internal/router/router.go b/internal/router/router.go
index 5dca8911..c604c368 100644
--- a/internal/router/router.go
+++ b/internal/router/router.go
@@ -323,6 +323,7 @@ func RegisterKnowledgeRoutes(r *gin.RouterGroup, handler *handler.KnowledgeHandl
 		// "must own every targeted KB" guard if the requirement
 		// surfaces.
 		k.PUT("/tags", g.Contributor(), handler.UpdateKnowledgeTagBatch)
+		k.POST("/batch-reparse", g.Contributor(), handler.BatchReparseKnowledge)
 		k.GET("/search", g.Viewer(), handler.SearchKnowledge)
 		k.POST("/batch-delete", g.Contributor(), handler.BatchDeleteKnowledge)
 		k.POST("/move", g.Contributor(), handler.MoveKnowledge)