Refa: migrate chunk APIs to RESTful routes (#14291)

### What problem does this PR solve? migrate chunk APIs to RESTful routes ### Type of change - [x] Refactoring
2026-07-23 00:46:42 +08:00 · 2026-04-23 14:17:23 +08:00
parent 76b017ca32
commit 7817b0d779
30 changed files with 1594 additions and 2237 deletions
--- a/web/src/hooks/route-hook.ts
+++ b/web/src/hooks/route-hook.ts
@@ -2,6 +2,7 @@ import {
  KnowledgeRouteKey,
  KnowledgeSearchParams,
 } from '@/constants/knowledge';
+import { Routes } from '@/routes';
 import { useCallback } from 'react';
 import { useLocation, useNavigate, useSearchParams } from 'react-router';

@@ -27,13 +28,16 @@ export const useThirdPathName = () => {

 export const useGetKnowledgeSearchParams = () => {
  const [currentQueryParameters] = useSearchParams();
+  const { pathname } = useLocation();
+  const isDataflowResultPage = pathname === Routes.DataflowResult;

  return {
    type: currentQueryParameters.get(KnowledgeSearchParams.Type) || '',
    documentId:
      currentQueryParameters.get(KnowledgeSearchParams.DocumentId) || '',
-    knowledgeId:
-      currentQueryParameters.get(KnowledgeSearchParams.KnowledgeId) || '',
+    knowledgeId: isDataflowResultPage
+      ? currentQueryParameters.get('knowledgeId') || ''
+      : currentQueryParameters.get(KnowledgeSearchParams.KnowledgeId) || '',
  };
 };

--- a/web/src/hooks/use-chunk-request.ts
+++ b/web/src/hooks/use-chunk-request.ts
@@ -40,6 +40,7 @@ export const useSelectChunkList = () => {
 export const useDeleteChunk = () => {
  const queryClient = useQueryClient();
  const { setPaginationParams } = useSetPaginationParams();
+  const { knowledgeId } = useGetKnowledgeSearchParams();
  const {
    data,
    isPending: loading,
@@ -47,7 +48,10 @@ export const useDeleteChunk = () => {
  } = useMutation({
    mutationKey: ['deleteChunk'],
    mutationFn: async (params: { chunkIds: string[]; doc_id: string }) => {
-      const { data } = await kbService.rmChunk(params);
+      const { data } = await kbService.rmChunk({
+        ...params,
+        kb_id: knowledgeId,
+      });
      if (data.code === 0) {
        setPaginationParams(1);
        queryClient.invalidateQueries({ queryKey: ['fetchChunkList'] });
@@ -62,6 +66,7 @@ export const useDeleteChunk = () => {
 export const useCreateChunk = () => {
  const { t } = useTranslation();
  const queryClient = useQueryClient();
+  const { knowledgeId } = useGetKnowledgeSearchParams();

  const {
    data,
@@ -74,7 +79,10 @@ export const useCreateChunk = () => {
      if (payload.chunk_id) {
        service = kbService.setChunk;
      }
-      const { data } = await service(payload);
+      const { data } = await service({
+        ...payload,
+        kb_id: payload.kb_id || knowledgeId,
+      });
      if (data.code === 0) {
        message.success(t('message.created'));
        setTimeout(() => {
@@ -88,14 +96,20 @@ export const useCreateChunk = () => {
  return { data, loading, createChunk: mutateAsync };
 };

-export const useFetchChunk = (chunkId?: string): ResponseType<any> => {
+export const useFetchChunk = (
+  chunkId?: string,
+  documentId?: string,
+): ResponseType<any> => {
+  const { knowledgeId } = useGetKnowledgeSearchParams();
  const { data } = useQuery({
-    queryKey: ['fetchChunk'],
-    enabled: !!chunkId,
+    queryKey: ['fetchChunk', knowledgeId, documentId, chunkId],
+    enabled: !!chunkId && !!documentId && !!knowledgeId,
    initialData: {},
    gcTime: 0,
    queryFn: async () => {
      const data = await kbService.getChunk({
+        kb_id: knowledgeId,
+        doc_id: documentId,
        chunk_id: chunkId,
      });

@@ -115,7 +129,7 @@ export const useFetchNextChunkList = (
 }> &
  IChunkListResult => {
  const { pagination, setPagination } = useGetPaginationWithRouter();
-  const { documentId } = useGetKnowledgeSearchParams();
+  const { documentId, knowledgeId } = useGetKnowledgeSearchParams();
  const { searchString, handleInputChange } = useHandleSearchChange();
  const [available, setAvailable] = useState<number | undefined>();
  const debouncedSearchString = useDebounce(searchString, { wait: 500 });
@@ -127,6 +141,7 @@ export const useFetchNextChunkList = (
  } = useQuery({
    queryKey: [
      'fetchChunkList',
+      knowledgeId,
      documentId,
      pagination.current,
      pagination.pageSize,
@@ -136,9 +151,10 @@ export const useFetchNextChunkList = (
    placeholderData: (previousData: any) =>
      previousData ?? { data: [], total: 0, documentInfo: {} }, // https://github.com/TanStack/query/issues/8183
    gcTime: 0,
-    enabled,
+    enabled: enabled && !!knowledgeId && !!documentId,
    queryFn: async () => {
      const { data } = await kbService.chunkList({
+        kb_id: knowledgeId,
        doc_id: documentId,
        page: pagination.current,
        size: pagination.pageSize,
@@ -195,6 +211,7 @@ export const useFetchNextChunkList = (

 export const useSwitchChunk = () => {
  const { t } = useTranslation();
+  const { knowledgeId } = useGetKnowledgeSearchParams();
  const {
    data,
    isPending: loading,
@@ -206,7 +223,10 @@ export const useSwitchChunk = () => {
      available_int?: number;
      doc_id: string;
    }) => {
-      const { data } = await kbService.switchChunk(params);
+      const { data } = await kbService.switchChunk({
+        ...params,
+        kb_id: knowledgeId,
+      });
      if (data.code === 0) {
        message.success(t('message.modified'));
      }
--- a/web/src/pages/chunk/parsed-result/add-knowledge/components/knowledge-chunk/components/chunk-creating-modal/index.tsx
+++ b/web/src/pages/chunk/parsed-result/add-knowledge/components/knowledge-chunk/components/chunk-creating-modal/index.tsx
@@ -26,7 +26,6 @@ import type { ChunkDocType } from '@/interfaces/database/knowledge';
 import React, { useCallback, useEffect, useState } from 'react';
 import { FieldValues, FormProvider, useForm } from 'react-hook-form';
 import { useTranslation } from 'react-i18next';
-import { useDeleteChunkByIds } from '../../hooks';
 import {
  transformTagFeaturesArrayToObject,
  transformTagFeaturesObjectToArray,
@@ -75,8 +74,7 @@ const ChunkCreatingModal: React.FC<IModalProps<any> & kFProps> = ({
    },
  });
  const [checked, setChecked] = useState(false);
-  const { removeChunk } = useDeleteChunkByIds();
-  const { data } = useFetchChunk(chunkId);
+  const { data } = useFetchChunk(chunkId, doc_id);
  const { t } = useTranslation();
  const isEditMode = !!chunkId;

@@ -99,12 +97,6 @@ const ChunkCreatingModal: React.FC<IModalProps<any> & kFProps> = ({

  const handleOk = form.handleSubmit(onSubmit);

-  const handleRemove = useCallback(() => {
-    if (chunkId) {
-      return removeChunk([chunkId], doc_id);
-    }
-  }, [chunkId, doc_id, removeChunk]);
-
  const handleCheck = useCallback(() => {
    setChecked(!checked);
  }, [checked]);
--- a/web/src/pages/dataflow-result/components/chunk-creating-modal/index.tsx
+++ b/web/src/pages/dataflow-result/components/chunk-creating-modal/index.tsx
@@ -57,7 +57,7 @@ const ChunkCreatingModal: React.FC<IModalProps<any> & kFProps> = ({
  });
  const [checked, setChecked] = useState(false);
  const { removeChunk } = useDeleteChunkByIds();
-  const { data } = useFetchChunk(chunkId);
+  const { data } = useFetchChunk(chunkId, doc_id);
  const { t } = useTranslation();

  const isTagParser = parserId === 'tag';
--- a/web/src/pages/dataset/dataset-overview/overview-table.tsx
+++ b/web/src/pages/dataset/dataset-overview/overview-table.tsx
@@ -51,7 +51,7 @@ import { DocumentLog, FileLogsTableProps, IFileLogItem } from './interface';
 export const getFileLogsTableColumns = (
  t: TFunction<'translation', string>,
  showLog: (row: Row<IFileLogItem & DocumentLog>, active: LogTabs) => void,
-  kowledgeId: string,
+  knowledgeId: string,
  navigateToDataflowResult: (
    props: NavigateToDataflowResultProps,
  ) => () => void,
@@ -210,7 +210,8 @@ export const getFileLogsTableColumns = (
              size="icon-sm"
              onClick={navigateToDataflowResult({
                id: row.original.id,
-                [PipelineResultSearchParams.KnowledgeId]: kowledgeId,
+                [PipelineResultSearchParams.KnowledgeId]:
+                  row.original.kb_id || knowledgeId,
                [PipelineResultSearchParams.DocumentId]:
                  row.original.document_id,
                [PipelineResultSearchParams.IsReadOnly]: 'false',
@@ -358,7 +359,7 @@ const FileLogsTable: FC<FileLogsTableProps> = ({
  const [isModalVisible, setIsModalVisible] = useState(false);
  const { navigateToDataflowResult } = useNavigatePage();
  const [logInfo, setLogInfo] = useState<IFileLogItem>();
-  const kowledgeId = useParams().id;
+  const knowledgeId = useParams().id;
  const showLog = (row: Row<IFileLogItem & DocumentLog>) => {
    const logDetail = {
      taskId: row.original?.dsl?.task_id,
@@ -382,7 +383,7 @@ const FileLogsTable: FC<FileLogsTableProps> = ({
      ? getFileLogsTableColumns(
          t,
          showLog,
-          kowledgeId || '',
+          knowledgeId || '',
          navigateToDataflowResult,
          dataSourceInfo,
        )
--- a/web/src/services/knowledge-service.ts
+++ b/web/src/services/knowledge-service.ts
@@ -21,12 +21,6 @@ const {
  documentCreate,
  documentChangeParser,
  documentThumbnails,
-  chunkList,
-  createChunk,
-  setChunk,
-  getChunk,
-  switchChunk,
-  rmChunk,
  retrievalTest,
  documentRun,
  documentUpload,
@@ -97,31 +91,6 @@ const methods = {
    url: setMeta,
    method: 'post',
  },
-  // chunk管理
-  chunkList: {
-    url: chunkList,
-    method: 'post',
-  },
-  createChunk: {
-    url: createChunk,
-    method: 'post',
-  },
-  setChunk: {
-    url: setChunk,
-    method: 'post',
-  },
-  getChunk: {
-    url: getChunk,
-    method: 'get',
-  },
-  switchChunk: {
-    url: switchChunk,
-    method: 'post',
-  },
-  rmChunk: {
-    url: rmChunk,
-    method: 'post',
-  },
  retrievalTest: {
    url: retrievalTest,
    method: 'post',
@@ -178,7 +147,139 @@ const methods = {
  },
 };

-const kbService = registerServer<keyof typeof methods>(methods, request);
+const baseKbService = registerServer<keyof typeof methods>(methods, request);
+
+const getDatasetId = (params: Record<string, any>) =>
+  params.dataset_id || params.kb_id || params.knowledge_id;
+
+const getDocumentId = (params: Record<string, any>) =>
+  params.document_id || params.doc_id;
+
+const mapChunkToLegacy = (chunk: Record<string, any>) => ({
+  ...chunk,
+  chunk_id: chunk.chunk_id || chunk.id,
+  content_with_weight: chunk.content_with_weight || chunk.content,
+  doc_id: chunk.doc_id || chunk.document_id,
+  doc_name: chunk.doc_name || chunk.docnm_kwd,
+  image_id: chunk.image_id || chunk.img_id,
+  important_kwd: chunk.important_kwd || chunk.important_keywords || [],
+  question_kwd: chunk.question_kwd || chunk.questions || [],
+  available_int: chunk.available_int ?? (chunk.available === false ? 0 : 1),
+  positions: chunk.positions || chunk.position_int || [],
+});
+
+const mapDocumentToLegacy = (doc: Record<string, any>) => ({
+  ...doc,
+  chunk_num: doc.chunk_num ?? doc.chunk_count,
+  kb_id: doc.kb_id || doc.dataset_id,
+});
+
+const mapChunkPayloadToRest = (payload: Record<string, any>) => ({
+  content: payload.content ?? payload.content_with_weight,
+  important_keywords: payload.important_keywords ?? payload.important_kwd,
+  questions: payload.questions ?? payload.question_kwd,
+  tag_kwd: payload.tag_kwd,
+  tag_feas: payload.tag_feas,
+  positions: payload.positions,
+  available:
+    payload.available ??
+    (payload.available_int === undefined
+      ? undefined
+      : payload.available_int === 1),
+  image_base64: payload.image_base64,
+});
+
+const getAvailableParam = (available?: number) => {
+  if (available === undefined) {
+    return undefined;
+  }
+  return available === 1 ? 'true' : 'false';
+};
+
+const chunkService = {
+  chunkList: async (params: Record<string, any>) => {
+    const datasetId = getDatasetId(params);
+    const documentId = getDocumentId(params);
+    const response = await request.get(api.chunkList(datasetId, documentId), {
+      params: {
+        page: params.page,
+        page_size: params.page_size || params.size,
+        keywords: params.keywords,
+        available: getAvailableParam(params.available_int),
+      },
+    });
+
+    if (response.data?.code === 0) {
+      response.data.data = {
+        ...response.data.data,
+        chunks: (response.data.data?.chunks || []).map(mapChunkToLegacy),
+        doc: mapDocumentToLegacy(response.data.data?.doc || {}),
+      };
+    }
+
+    return response;
+  },
+  createChunk: async (payload: Record<string, any>) => {
+    const datasetId = getDatasetId(payload);
+    const documentId = getDocumentId(payload);
+    const response = await request.post(api.chunkList(datasetId, documentId), {
+      data: mapChunkPayloadToRest(payload),
+    });
+
+    if (response.data?.code === 0 && response.data.data?.chunk) {
+      response.data.data.chunk = mapChunkToLegacy(response.data.data.chunk);
+    }
+
+    return response;
+  },
+  setChunk: (payload: Record<string, any>) => {
+    const datasetId = getDatasetId(payload);
+    const documentId = getDocumentId(payload);
+    const chunkId = payload.chunk_id || payload.id;
+    return request.patch(api.chunkDetail(datasetId, documentId, chunkId), {
+      data: mapChunkPayloadToRest(payload),
+    });
+  },
+  getChunk: async (params: Record<string, any>) => {
+    const datasetId = getDatasetId(params);
+    const documentId = getDocumentId(params);
+    const chunkId = params.chunk_id || params.id;
+    const response = await request.get(
+      api.chunkDetail(datasetId, documentId, chunkId),
+    );
+
+    if (response.data?.code === 0) {
+      response.data.data = mapChunkToLegacy(response.data.data || {});
+    }
+
+    return response;
+  },
+  switchChunk: (params: Record<string, any>) => {
+    const datasetId = getDatasetId(params);
+    const documentId = getDocumentId(params);
+    return request.patch(api.chunkList(datasetId, documentId), {
+      data: {
+        chunk_ids: params.chunk_ids || params.chunkIds,
+        available_int: params.available_int,
+      },
+    });
+  },
+  rmChunk: (params: Record<string, any>) => {
+    const datasetId = getDatasetId(params);
+    const documentId = getDocumentId(params);
+    return request.delete(api.chunkList(datasetId, documentId), {
+      data: {
+        chunk_ids: params.chunk_ids || params.chunkIds,
+        delete_all: params.delete_all,
+      },
+    });
+  },
+};
+
+const kbService = {
+  ...baseKbService,
+  ...chunkService,
+};

 export const listTag = (knowledgeId: string) =>
  request.get(api.listTag(knowledgeId));
--- a/web/src/utils/api.ts
+++ b/web/src/utils/api.ts
@@ -99,12 +99,10 @@ export default {
  renameTag: (knowledgeId: string) => `${webAPI}/kb/${knowledgeId}/rename_tag`,

  // chunk
-  chunkList: `${webAPI}/chunk/list`,
-  createChunk: `${webAPI}/chunk/create`,
-  setChunk: `${webAPI}/chunk/set`,
-  getChunk: `${webAPI}/chunk/get`,
-  switchChunk: `${webAPI}/chunk/switch`,
-  rmChunk: `${webAPI}/chunk/rm`,
+  chunkList: (datasetId: string, documentId: string) =>
+    `${restAPIv1}/datasets/${datasetId}/documents/${documentId}/chunks`,
+  chunkDetail: (datasetId: string, documentId: string, chunkId: string) =>
+    `${restAPIv1}/datasets/${datasetId}/documents/${documentId}/chunks/${chunkId}`,
  retrievalTest: `${webAPI}/chunk/retrieval_test`,
  knowledgeGraph: `${webAPI}/chunk/knowledge_graph`,