Feat/kb index (#20868)

Co-authored-by: twwu <twwu@dify.ai>

api/core/indexing_runner.py

@@ -534,7 +534,7 @@ class IndexingRunner:
         # chunk nodes by chunk size
         indexing_start_at = time.perf_counter()
         tokens = 0
-        if dataset_document.doc_form != IndexType.PARENT_CHILD_INDEX:
+        if dataset_document.doc_form != IndexType.PARENT_CHILD_INDEX and dataset.indexing_technique == "economy":
             # create keyword index
             create_keyword_thread = threading.Thread(
                 target=self._process_keyword_index,
@@ -572,7 +572,7 @@ class IndexingRunner:
 
                 for future in futures:
                     tokens += future.result()
-        if dataset_document.doc_form != IndexType.PARENT_CHILD_INDEX:
+        if dataset_document.doc_form != IndexType.PARENT_CHILD_INDEX and dataset.indexing_technique == "economy":
             create_keyword_thread.join()
         indexing_end_at = time.perf_counter()
 

api/core/rag/index_processor/processor/paragraph_index_processor.py

@@ -76,6 +76,7 @@ class ParagraphIndexProcessor(BaseIndexProcessor):
         if dataset.indexing_technique == "high_quality":
             vector = Vector(dataset)
             vector.create(documents)
+            with_keywords = False
         if with_keywords:
             keywords_list = kwargs.get("keywords_list")
             keyword = Keyword(dataset)
@@ -91,6 +92,7 @@ class ParagraphIndexProcessor(BaseIndexProcessor):
                 vector.delete_by_ids(node_ids)
             else:
                 vector.delete()
+            with_keywords = False
         if with_keywords:
             keyword = Keyword(dataset)
             if node_ids:

api/services/vector_service.py

@@ -97,16 +97,16 @@ class VectorService:
             vector = Vector(dataset=dataset)
             vector.delete_by_ids([segment.index_node_id])
             vector.add_texts([document], duplicate_check=True)
-
-        # update keyword index
-        keyword = Keyword(dataset)
-        keyword.delete_by_ids([segment.index_node_id])
-
-        # save keyword index
-        if keywords and len(keywords) > 0:
-            keyword.add_texts([document], keywords_list=[keywords])
         else:
-            keyword.add_texts([document])
+            # update keyword index
+            keyword = Keyword(dataset)
+            keyword.delete_by_ids([segment.index_node_id])
+
+            # save keyword index
+            if keywords and len(keywords) > 0:
+                keyword.add_texts([document], keywords_list=[keywords])
+            else:
+                keyword.add_texts([document])
 
     @classmethod
     def generate_child_chunks(

web/app/components/datasets/documents/detail/completed/segment-detail.tsx

@@ -1,4 +1,4 @@
-import React, { type FC, useMemo, useState } from 'react'
+import React, { type FC, useCallback, useMemo, useState } from 'react'
 import { useTranslation } from 'react-i18next'
 import {
   RiCloseLine,
@@ -16,8 +16,10 @@ import { useSegmentListContext } from './index'
 import { ChunkingMode, type SegmentDetailModel } from '@/models/datasets'
 import { useEventEmitterContextContext } from '@/context/event-emitter'
 import { formatNumber } from '@/utils/format'
-import classNames from '@/utils/classnames'
+import cn from '@/utils/classnames'
 import Divider from '@/app/components/base/divider'
+import { useDatasetDetailContextWithSelector } from '@/context/dataset-detail'
+import { IndexingType } from '../../../create/step-two'
 
 type ISegmentDetailProps = {
   segInfo?: Partial<SegmentDetailModel> & { id: string }
@@ -48,6 +50,7 @@ const SegmentDetail: FC<ISegmentDetailProps> = ({
   const toggleFullScreen = useSegmentListContext(s => s.toggleFullScreen)
   const mode = useDocumentContext(s => s.mode)
   const parentMode = useDocumentContext(s => s.parentMode)
+  const indexingTechnique = useDatasetDetailContextWithSelector(s => s.dataset?.indexing_technique)
 
   eventEmitter?.useSubscription((v) => {
     if (v === 'update-segment')
@@ -56,56 +59,41 @@ const SegmentDetail: FC<ISegmentDetailProps> = ({
       setLoading(false)
   })
 
-  const handleCancel = () => {
+  const handleCancel = useCallback(() => {
     onCancel()
-  }
+  }, [onCancel])
 
-  const handleSave = () => {
+  const handleSave = useCallback(() => {
     onUpdate(segInfo?.id || '', question, answer, keywords)
-  }
+  }, [onUpdate, segInfo?.id, question, answer, keywords])
 
-  const handleRegeneration = () => {
+  const handleRegeneration = useCallback(() => {
     setShowRegenerationModal(true)
-  }
+  }, [])
 
-  const onCancelRegeneration = () => {
+  const onCancelRegeneration = useCallback(() => {
     setShowRegenerationModal(false)
-  }
+  }, [])
 
-  const onConfirmRegeneration = () => {
+  const onConfirmRegeneration = useCallback(() => {
     onUpdate(segInfo?.id || '', question, answer, keywords, true)
-  }
-
-  const isParentChildMode = useMemo(() => {
-    return mode === 'hierarchical'
-  }, [mode])
-
-  const isFullDocMode = useMemo(() => {
-    return mode === 'hierarchical' && parentMode === 'full-doc'
-  }, [mode, parentMode])
-
-  const titleText = useMemo(() => {
-    return isEditMode ? t('datasetDocuments.segment.editChunk') : t('datasetDocuments.segment.chunkDetail')
-  }, [isEditMode, t])
-
-  const isQAModel = useMemo(() => {
-    return docForm === ChunkingMode.qa
-  }, [docForm])
+  }, [onUpdate, segInfo?.id, question, answer, keywords])
 
   const wordCountText = useMemo(() => {
-    const contentLength = isQAModel ? (question.length + answer.length) : question.length
+    const contentLength = docForm === ChunkingMode.qa ? (question.length + answer.length) : question.length
     const total = formatNumber(isEditMode ? contentLength : segInfo!.word_count as number)
     const count = isEditMode ? contentLength : segInfo!.word_count as number
     return `${total} ${t('datasetDocuments.segment.characters', { count })}`
-  }, [isEditMode, question.length, answer.length, isQAModel, segInfo, t])
+  }, [isEditMode, question.length, answer.length, docForm, segInfo, t])
 
-  const labelPrefix = useMemo(() => {
-    return isParentChildMode ? t('datasetDocuments.segment.parentChunk') : t('datasetDocuments.segment.chunk')
-  }, [isParentChildMode, t])
+  const isFullDocMode = mode === 'hierarchical' && parentMode === 'full-doc'
+  const titleText = isEditMode ? t('datasetDocuments.segment.editChunk') : t('datasetDocuments.segment.chunkDetail')
+  const labelPrefix = mode === 'hierarchical' ? t('datasetDocuments.segment.parentChunk') : t('datasetDocuments.segment.chunk')
+  const isECOIndexing = indexingTechnique === IndexingType.ECONOMICAL
 
   return (
     <div className={'flex h-full flex-col'}>
-      <div className={classNames('flex items-center justify-between', fullScreen ? 'py-3 pr-4 pl-6 border border-divider-subtle' : 'pt-3 pr-3 pl-4')}>
+      <div className={cn('flex items-center justify-between', fullScreen ? 'border border-divider-subtle py-3 pl-6 pr-4' : 'pl-4 pr-3 pt-3')}>
         <div className='flex flex-col'>
           <div className='system-xl-semibold text-text-primary'>{titleText}</div>
           <div className='flex items-center gap-x-2'>
@@ -134,12 +122,12 @@ const SegmentDetail: FC<ISegmentDetailProps> = ({
           </div>
         </div>
       </div>
-      <div className={classNames(
+      <div className={cn(
         'flex grow',
-        fullScreen ? 'w-full flex-row justify-center px-6 pt-6 gap-x-8' : 'flex-col gap-y-1 py-3 px-4',
-        !isEditMode && 'pb-0 overflow-hidden',
+        fullScreen ? 'w-full flex-row justify-center gap-x-8 px-6 pt-6' : 'flex-col gap-y-1 px-4 py-3',
+        !isEditMode && 'overflow-hidden pb-0',
       )}>
-        <div className={classNames(isEditMode ? 'break-all whitespace-pre-line overflow-hidden' : 'overflow-y-auto', fullScreen ? 'w-1/2' : 'grow')}>
+        <div className={cn(isEditMode ? 'overflow-hidden whitespace-pre-line break-all' : 'overflow-y-auto', fullScreen ? 'w-1/2' : 'grow')}>
           <ChunkContent
             docForm={docForm}
             question={question}
@@ -149,7 +137,7 @@ const SegmentDetail: FC<ISegmentDetailProps> = ({
             isEditMode={isEditMode}
           />
         </div>
-        {mode === 'custom' && <Keywords
+        {isECOIndexing && <Keywords
           className={fullScreen ? 'w-1/5' : ''}
           actionType={isEditMode ? 'edit' : 'view'}
           segInfo={segInfo}

web/app/components/datasets/documents/detail/new-segment.tsx

@@ -1,4 +1,4 @@
-import { memo, useMemo, useRef, useState } from 'react'
+import { memo, useCallback, useMemo, useRef, useState } from 'react'
 import type { FC } from 'react'
 import { useTranslation } from 'react-i18next'
 import { useContext } from 'use-context-selector'
@@ -12,7 +12,6 @@ import Keywords from './completed/common/keywords'
 import ChunkContent from './completed/common/chunk-content'
 import AddAnother from './completed/common/add-another'
 import Dot from './completed/common/dot'
-import { useDocumentContext } from './index'
 import { useStore as useAppStore } from '@/app/components/app/store'
 import { ToastContext } from '@/app/components/base/toast'
 import { ChunkingMode, type SegmentUpdater } from '@/models/datasets'
@@ -20,6 +19,8 @@ import classNames from '@/utils/classnames'
 import { formatNumber } from '@/utils/format'
 import Divider from '@/app/components/base/divider'
 import { useAddSegment } from '@/service/knowledge/use-segment'
+import { useDatasetDetailContextWithSelector } from '@/context/dataset-detail'
+import { IndexingType } from '../../create/step-two'
 
 type NewSegmentModalProps = {
   onCancel: () => void
@@ -44,39 +45,37 @@ const NewSegmentModal: FC<NewSegmentModalProps> = ({
   const [addAnother, setAddAnother] = useState(true)
   const fullScreen = useSegmentListContext(s => s.fullScreen)
   const toggleFullScreen = useSegmentListContext(s => s.toggleFullScreen)
-  const mode = useDocumentContext(s => s.mode)
+  const indexingTechnique = useDatasetDetailContextWithSelector(s => s.dataset?.indexing_technique)
   const { appSidebarExpand } = useAppStore(useShallow(state => ({
     appSidebarExpand: state.appSidebarExpand,
   })))
   const refreshTimer = useRef<any>(null)
 
-  const CustomButton = <>
-    <Divider type='vertical' className='mx-1 h-3 bg-divider-regular' />
-    <button
-      type='button'
-      className='system-xs-semibold text-text-accent'
-      onClick={() => {
-        clearTimeout(refreshTimer.current)
-        viewNewlyAddedChunk()
-      }}>
-      {t('common.operation.view')}
-    </button>
-  </>
+  const CustomButton = useMemo(() => (
+    <>
+      <Divider type='vertical' className='mx-1 h-3 bg-divider-regular' />
+      <button
+        type='button'
+        className='system-xs-semibold text-text-accent'
+        onClick={() => {
+          clearTimeout(refreshTimer.current)
+          viewNewlyAddedChunk()
+        }}>
+        {t('common.operation.view')}
+      </button>
+    </>
+  ), [viewNewlyAddedChunk, t])
 
-  const isQAModel = useMemo(() => {
-    return docForm === ChunkingMode.qa
-  }, [docForm])
-
-  const handleCancel = (actionType: 'esc' | 'add' = 'esc') => {
+  const handleCancel = useCallback((actionType: 'esc' | 'add' = 'esc') => {
     if (actionType === 'esc' || !addAnother)
       onCancel()
-  }
+  }, [onCancel, addAnother])
 
   const { mutateAsync: addSegment } = useAddSegment()
 
-  const handleSave = async () => {
+  const handleSave = useCallback(async () => {
     const params: SegmentUpdater = { content: '' }
-    if (isQAModel) {
+    if (docForm === ChunkingMode.qa) {
       if (!question.trim()) {
         return notify({
           type: 'error',
@@ -129,21 +128,27 @@ const NewSegmentModal: FC<NewSegmentModalProps> = ({
         setLoading(false)
       },
     })
-  }
+  }, [docForm, keywords, addSegment, datasetId, documentId, question, answer, notify, t, appSidebarExpand, CustomButton, handleCancel, onSave])
 
   const wordCountText = useMemo(() => {
-    const count = isQAModel ? (question.length + answer.length) : question.length
+    const count = docForm === ChunkingMode.qa ? (question.length + answer.length) : question.length
     return `${formatNumber(count)} ${t('datasetDocuments.segment.characters', { count })}`
-    // eslint-disable-next-line react-hooks/exhaustive-deps
-  }, [question.length, answer.length, isQAModel])
+  }, [question.length, answer.length, docForm, t])
+
+  const isECOIndexing = indexingTechnique === IndexingType.ECONOMICAL
 
   return (
     <div className={'flex h-full flex-col'}>
-      <div className={classNames('flex items-center justify-between', fullScreen ? 'py-3 pr-4 pl-6 border border-divider-subtle' : 'pt-3 pr-3 pl-4')}>
+      <div
+        className={classNames(
+          'flex items-center justify-between',
+          fullScreen ? 'border border-divider-subtle py-3 pl-6 pr-4' : 'pl-4 pr-3 pt-3',
+        )}
+      >
         <div className='flex flex-col'>
-          <div className='system-xl-semibold text-text-primary'>{
-            t('datasetDocuments.segment.addChunk')
-          }</div>
+          <div className='system-xl-semibold text-text-primary'>
+            {t('datasetDocuments.segment.addChunk')}
+          </div>
           <div className='flex items-center gap-x-2'>
             <SegmentIndexTag label={t('datasetDocuments.segment.newChunk')!} />
             <Dot />
@@ -171,8 +176,8 @@ const NewSegmentModal: FC<NewSegmentModalProps> = ({
           </div>
         </div>
       </div>
-      <div className={classNames('flex grow', fullScreen ? 'w-full flex-row justify-center px-6 pt-6 gap-x-8' : 'flex-col gap-y-1 py-3 px-4')}>
-        <div className={classNames('break-all overflow-hidden whitespace-pre-line', fullScreen ? 'w-1/2' : 'grow')}>
+      <div className={classNames('flex grow', fullScreen ? 'w-full flex-row justify-center gap-x-8 px-6 pt-6' : 'flex-col gap-y-1 px-4 py-3')}>
+        <div className={classNames('overflow-hidden whitespace-pre-line break-all', fullScreen ? 'w-1/2' : 'grow')}>
           <ChunkContent
             docForm={docForm}
             question={question}
@@ -182,7 +187,7 @@ const NewSegmentModal: FC<NewSegmentModalProps> = ({
             isEditMode={true}
           />
         </div>
-        {mode === 'custom' && <Keywords
+        {isECOIndexing && <Keywords
           className={fullScreen ? 'w-1/5' : ''}
           actionType='add'
           keywords={keywords}

web/eslint.config.mjs

@@ -213,7 +213,7 @@ export default combine(
     settings: {
       tailwindcss: {
         // These are the default values but feel free to customize
-        callees: ['classnames', 'clsx', 'ctl', 'cn'],
+        callees: ['classnames', 'clsx', 'ctl', 'cn', 'classNames'],
         config: 'tailwind.config.js', // returned from `loadConfig()` utility if not provided
         cssFiles: [
           '**/*.css',