feat: summary index

web/app/components/base/icons/assets/vender/knowledge/search-lines-sparkle.svg

@@ -0,0 +1,6 @@
+<svg width="16" height="16" viewBox="0 0 16 16" fill="none" xmlns="http://www.w3.org/2000/svg">
+<path d="M12 7.33337V2.66671H4.00002V13.3334H8.00002C8.36821 13.3334 8.66669 13.6319 8.66669 14C8.66669 14.3682 8.36821 14.6667 8.00002 14.6667H3.33335C2.96516 14.6667 2.66669 14.3682 2.66669 14V2.00004C2.66669 1.63185 2.96516 1.33337 3.33335 1.33337H12.6667C13.0349 1.33337 13.3334 1.63185 13.3334 2.00004V7.33337C13.3334 7.70156 13.0349 8.00004 12.6667 8.00004C12.2985 8.00004 12 7.70156 12 7.33337Z" fill="#354052"/>
+<path d="M10 4.00004C10.3682 4.00004 10.6667 4.29852 10.6667 4.66671C10.6667 5.0349 10.3682 5.33337 10 5.33337H6.00002C5.63183 5.33337 5.33335 5.0349 5.33335 4.66671C5.33335 4.29852 5.63183 4.00004 6.00002 4.00004H10Z" fill="#354052"/>
+<path d="M8.00002 6.66671C8.36821 6.66671 8.66669 6.96518 8.66669 7.33337C8.66669 7.70156 8.36821 8.00004 8.00002 8.00004H6.00002C5.63183 8.00004 5.33335 7.70156 5.33335 7.33337C5.33335 6.96518 5.63183 6.66671 6.00002 6.66671H8.00002Z" fill="#354052"/>
+<path d="M12.827 10.7902L12.3624 9.58224C12.3048 9.43231 12.1607 9.33337 12 9.33337C11.8394 9.33337 11.6953 9.43231 11.6376 9.58224L11.173 10.7902C11.1054 10.9662 10.9662 11.1054 10.7902 11.173L9.58222 11.6376C9.43229 11.6953 9.33335 11.8394 9.33335 12C9.33335 12.1607 9.43229 12.3048 9.58222 12.3624L10.7902 12.827C10.9662 12.8947 11.1054 13.0338 11.173 13.2099L11.6376 14.4178C11.6953 14.5678 11.8394 14.6667 12 14.6667C12.1607 14.6667 12.3048 14.5678 12.3624 14.4178L12.827 13.2099C12.8947 13.0338 13.0338 12.8947 13.2099 12.827L14.4178 12.3624C14.5678 12.3048 14.6667 12.1607 14.6667 12C14.6667 11.8394 14.5678 11.6953 14.4178 11.6376L13.2099 11.173C13.0338 11.1054 12.8947 10.9662 12.827 10.7902Z" fill="#354052"/>
+</svg>

web/app/components/base/icons/src/vender/knowledge/SearchLinesSparkle.json

@@ -0,0 +1,53 @@
+{
+  "icon": {
+    "type": "element",
+    "isRootNode": true,
+    "name": "svg",
+    "attributes": {
+      "width": "16",
+      "height": "16",
+      "viewBox": "0 0 16 16",
+      "fill": "none",
+      "xmlns": "http://www.w3.org/2000/svg"
+    },
+    "children": [
+      {
+        "type": "element",
+        "name": "path",
+        "attributes": {
+          "d": "M12 7.33337V2.66671H4.00002V13.3334H8.00002C8.36821 13.3334 8.66669 13.6319 8.66669 14C8.66669 14.3682 8.36821 14.6667 8.00002 14.6667H3.33335C2.96516 14.6667 2.66669 14.3682 2.66669 14V2.00004C2.66669 1.63185 2.96516 1.33337 3.33335 1.33337H12.6667C13.0349 1.33337 13.3334 1.63185 13.3334 2.00004V7.33337C13.3334 7.70156 13.0349 8.00004 12.6667 8.00004C12.2985 8.00004 12 7.70156 12 7.33337Z",
+          "fill": "currentColor"
+        },
+        "children": []
+      },
+      {
+        "type": "element",
+        "name": "path",
+        "attributes": {
+          "d": "M10 4.00004C10.3682 4.00004 10.6667 4.29852 10.6667 4.66671C10.6667 5.0349 10.3682 5.33337 10 5.33337H6.00002C5.63183 5.33337 5.33335 5.0349 5.33335 4.66671C5.33335 4.29852 5.63183 4.00004 6.00002 4.00004H10Z",
+          "fill": "currentColor"
+        },
+        "children": []
+      },
+      {
+        "type": "element",
+        "name": "path",
+        "attributes": {
+          "d": "M8.00002 6.66671C8.36821 6.66671 8.66669 6.96518 8.66669 7.33337C8.66669 7.70156 8.36821 8.00004 8.00002 8.00004H6.00002C5.63183 8.00004 5.33335 7.70156 5.33335 7.33337C5.33335 6.96518 5.63183 6.66671 6.00002 6.66671H8.00002Z",
+          "fill": "currentColor"
+        },
+        "children": []
+      },
+      {
+        "type": "element",
+        "name": "path",
+        "attributes": {
+          "d": "M12.827 10.7902L12.3624 9.58224C12.3048 9.43231 12.1607 9.33337 12 9.33337C11.8394 9.33337 11.6953 9.43231 11.6376 9.58224L11.173 10.7902C11.1054 10.9662 10.9662 11.1054 10.7902 11.173L9.58222 11.6376C9.43229 11.6953 9.33335 11.8394 9.33335 12C9.33335 12.1607 9.43229 12.3048 9.58222 12.3624L10.7902 12.827C10.9662 12.8947 11.1054 13.0338 11.173 13.2099L11.6376 14.4178C11.6953 14.5678 11.8394 14.6667 12 14.6667C12.1607 14.6667 12.3048 14.5678 12.3624 14.4178L12.827 13.2099C12.8947 13.0338 13.0338 12.8947 13.2099 12.827L14.4178 12.3624C14.5678 12.3048 14.6667 12.1607 14.6667 12C14.6667 11.8394 14.5678 11.6953 14.4178 11.6376L13.2099 11.173C13.0338 11.1054 12.8947 10.9662 12.827 10.7902Z",
+          "fill": "currentColor"
+        },
+        "children": []
+      }
+    ]
+  },
+  "name": "SearchLinesSparkle"
+}

web/app/components/base/icons/src/vender/knowledge/SearchLinesSparkle.tsx

@@ -0,0 +1,20 @@
+// GENERATE BY script
+// DON NOT EDIT IT MANUALLY
+
+import type { IconData } from '@/app/components/base/icons/IconBase'
+import * as React from 'react'
+import IconBase from '@/app/components/base/icons/IconBase'
+import data from './SearchLinesSparkle.json'
+
+const Icon = (
+  {
+    ref,
+    ...props
+  }: React.SVGProps<SVGSVGElement> & {
+    ref?: React.RefObject<React.RefObject<HTMLOrSVGElement>>
+  },
+) => <IconBase {...props} ref={ref} data={data as IconData} />
+
+Icon.displayName = 'SearchLinesSparkle'
+
+export default Icon

web/app/components/base/icons/src/vender/knowledge/index.ts

@@ -11,5 +11,6 @@ export { default as HighQuality } from './HighQuality'
 export { default as HybridSearch } from './HybridSearch'
 export { default as ParentChildChunk } from './ParentChildChunk'
 export { default as QuestionAndAnswer } from './QuestionAndAnswer'
+export { default as SearchLinesSparkle } from './SearchLinesSparkle'
 export { default as SearchMenu } from './SearchMenu'
 export { default as VectorSearch } from './VectorSearch'

web/app/components/datasets/create/step-two/components/general-chunking-options.tsx

@@ -1,7 +1,7 @@
 'use client'
 
 import type { FC } from 'react'
-import type { PreProcessingRule } from '@/models/datasets'
+import type { PreProcessingRule, SummaryIndexSetting as SummaryIndexSettingType } from '@/models/datasets'
 import {
   RiAlertFill,
   RiSearchEyeLine,
@@ -12,6 +12,7 @@ import Button from '@/app/components/base/button'
 import Checkbox from '@/app/components/base/checkbox'
 import Divider from '@/app/components/base/divider'
 import Tooltip from '@/app/components/base/tooltip'
+import SummaryIndexSetting from '@/app/components/datasets/settings/summary-index-setting'
 import { IS_CE_EDITION } from '@/config'
 import { ChunkingMode } from '@/models/datasets'
 import SettingCog from '../../assets/setting-gear-mod.svg'
@@ -52,6 +53,8 @@ type GeneralChunkingOptionsProps = {
   onReset: () => void
   // Locale
   locale: string
+  summaryIndexSetting?: SummaryIndexSettingType
+  onSummaryIndexSettingChange?: (payload: SummaryIndexSettingType) => void
 }
 
 export const GeneralChunkingOptions: FC<GeneralChunkingOptionsProps> = ({
@@ -74,6 +77,8 @@ export const GeneralChunkingOptions: FC<GeneralChunkingOptionsProps> = ({
   onPreview,
   onReset,
   locale,
+  summaryIndexSetting,
+  onSummaryIndexSettingChange,
 }) => {
   const { t } = useTranslation()
 
@@ -146,6 +151,13 @@ export const GeneralChunkingOptions: FC<GeneralChunkingOptionsProps> = ({
                 </label>
               </div>
             ))}
+            <div className="mt-3">
+              <SummaryIndexSetting
+                entry="create-document"
+                summaryIndexSetting={summaryIndexSetting}
+                onSummaryIndexSettingChange={onSummaryIndexSettingChange}
+              />
+            </div>
             {IS_CE_EDITION && (
               <>
                 <Divider type="horizontal" className="my-4 bg-divider-subtle" />

web/app/components/datasets/create/step-two/components/parent-child-options.tsx

@@ -2,7 +2,7 @@
 
 import type { FC } from 'react'
 import type { ParentChildConfig } from '../hooks'
-import type { ParentMode, PreProcessingRule } from '@/models/datasets'
+import type { ParentMode, PreProcessingRule, SummaryIndexSetting as SummaryIndexSettingType } from '@/models/datasets'
 import { RiSearchEyeLine } from '@remixicon/react'
 import Image from 'next/image'
 import { useTranslation } from 'react-i18next'
@@ -11,6 +11,7 @@ import Checkbox from '@/app/components/base/checkbox'
 import Divider from '@/app/components/base/divider'
 import { ParentChildChunk } from '@/app/components/base/icons/src/vender/knowledge'
 import RadioCard from '@/app/components/base/radio-card'
+import SummaryIndexSetting from '@/app/components/datasets/settings/summary-index-setting'
 import { ChunkingMode } from '@/models/datasets'
 import FileList from '../../assets/file-list-3-fill.svg'
 import Note from '../../assets/note-mod.svg'
@@ -31,6 +32,8 @@ type ParentChildOptionsProps = {
   // State
   parentChildConfig: ParentChildConfig
   rules: PreProcessingRule[]
+  summaryIndexSetting?: SummaryIndexSettingType
+  onSummaryIndexSettingChange?: (payload: SummaryIndexSettingType) => void
   currentDocForm: ChunkingMode
   // Flags
   isActive: boolean
@@ -51,6 +54,7 @@ type ParentChildOptionsProps = {
 export const ParentChildOptions: FC<ParentChildOptionsProps> = ({
   parentChildConfig,
   rules,
+  summaryIndexSetting,
   currentDocForm: _currentDocForm,
   isActive,
   isInUpload,
@@ -62,6 +66,7 @@ export const ParentChildOptions: FC<ParentChildOptionsProps> = ({
   onChildDelimiterChange,
   onChildMaxLengthChange,
   onRuleToggle,
+  onSummaryIndexSettingChange,
   onPreview,
   onReset,
 }) => {
@@ -183,6 +188,13 @@ export const ParentChildOptions: FC<ParentChildOptionsProps> = ({
                 </label>
               </div>
             ))}
+            <div className="mt-3">
+              <SummaryIndexSetting
+                entry="create-document"
+                summaryIndexSetting={summaryIndexSetting}
+                onSummaryIndexSettingChange={onSummaryIndexSettingChange}
+              />
+            </div>
           </div>
         </div>
       </div>

web/app/components/datasets/create/step-two/hooks/use-document-creation.ts

@@ -9,6 +9,7 @@ import type {
   CustomFile,
   FullDocumentDetail,
   ProcessRule,
+  SummaryIndexSetting as SummaryIndexSettingType,
 } from '@/models/datasets'
 import type { RetrievalConfig, RETRIEVE_METHOD } from '@/types/app'
 import { useCallback } from 'react'
@@ -141,6 +142,7 @@ export const useDocumentCreation = (options: UseDocumentCreationOptions) => {
     retrievalConfig: RetrievalConfig,
     embeddingModel: DefaultModel,
     indexingTechnique: string,
+    summaryIndexSetting?: SummaryIndexSettingType,
   ): CreateDocumentReq | null => {
     if (isSetting) {
       return {
@@ -148,6 +150,7 @@ export const useDocumentCreation = (options: UseDocumentCreationOptions) => {
         doc_form: currentDocForm,
         doc_language: docLanguage,
         process_rule: processRule,
+        summary_index_setting: summaryIndexSetting,
         retrieval_model: retrievalConfig,
         embedding_model: embeddingModel.model,
         embedding_model_provider: embeddingModel.provider,
@@ -164,6 +167,7 @@ export const useDocumentCreation = (options: UseDocumentCreationOptions) => {
       },
       indexing_technique: indexingTechnique,
       process_rule: processRule,
+      summary_index_setting: summaryIndexSetting,
       doc_form: currentDocForm,
       doc_language: docLanguage,
       retrieval_model: retrievalConfig,

web/app/components/datasets/create/step-two/hooks/use-indexing-config.ts

@@ -1,4 +1,5 @@
 import type { DefaultModel } from '@/app/components/header/account-setting/model-provider-page/declarations'
+import type { SummaryIndexSetting as SummaryIndexSettingType } from '@/models/datasets'
 import type { RetrievalConfig } from '@/types/app'
 import { useEffect, useMemo, useState } from 'react'
 import { checkShowMultiModalTip } from '@/app/components/datasets/settings/utils'
@@ -29,6 +30,7 @@ export type UseIndexingConfigOptions = {
   initialRetrievalConfig?: RetrievalConfig
   isAPIKeySet: boolean
   hasSetIndexType: boolean
+  initialSummaryIndexSetting?: SummaryIndexSettingType
 }
 
 export const useIndexingConfig = (options: UseIndexingConfigOptions) => {
@@ -38,6 +40,7 @@ export const useIndexingConfig = (options: UseIndexingConfigOptions) => {
     initialRetrievalConfig,
     isAPIKeySet,
     hasSetIndexType,
+    initialSummaryIndexSetting,
   } = options
 
   // Rerank model
@@ -115,6 +118,11 @@ export const useIndexingConfig = (options: UseIndexingConfigOptions) => {
   // Get effective indexing technique
   const getIndexingTechnique = () => initialIndexType || indexType
 
+  // Summary index setting
+  const [summaryIndexSetting, setSummaryIndexSetting] = useState<SummaryIndexSettingType | undefined>(
+    initialSummaryIndexSetting ?? undefined,
+  )
+
   return {
     // Index type
     indexType,
@@ -137,6 +145,10 @@ export const useIndexingConfig = (options: UseIndexingConfigOptions) => {
 
     // Computed
     showMultiModalTip,
+
+    // Summary index setting
+    summaryIndexSetting,
+    setSummaryIndexSetting,
   }
 }
 

web/app/components/datasets/create/step-two/index.tsx

@@ -72,6 +72,7 @@ const StepTwo: FC<StepTwoProps> = ({
     initialRetrievalConfig: currentDataset?.retrieval_model_dict,
     isAPIKeySet,
     hasSetIndexType,
+    initialSummaryIndexSetting: currentDataset?.summary_index_setting,
   })
   const preview = usePreviewState({ dataSourceType, files, notionPages, websitePages, documentDetail, datasetId })
   const creation = useDocumentCreation({
@@ -156,7 +157,7 @@ const StepTwo: FC<StepTwoProps> = ({
     })
     if (!isValid)
       return
-    const params = creation.buildCreationParams(currentDocForm, docLanguage, segmentation.getProcessRule(currentDocForm), indexing.retrievalConfig, indexing.embeddingModel, indexing.getIndexingTechnique())
+    const params = creation.buildCreationParams(currentDocForm, docLanguage, segmentation.getProcessRule(currentDocForm), indexing.retrievalConfig, indexing.embeddingModel, indexing.getIndexingTechnique(), indexing.summaryIndexSetting)
     if (!params)
       return
     await creation.executeCreation(params, indexing.indexType, indexing.retrievalConfig)
@@ -217,6 +218,8 @@ const StepTwo: FC<StepTwoProps> = ({
             onPreview={updatePreview}
             onReset={segmentation.resetToDefaults}
             locale={locale}
+            summaryIndexSetting={indexing.summaryIndexSetting}
+            onSummaryIndexSettingChange={indexing.setSummaryIndexSetting}
           />
         )}
         {showParentChildOption && (
@@ -236,6 +239,8 @@ const StepTwo: FC<StepTwoProps> = ({
             onRuleToggle={segmentation.toggleRule}
             onPreview={updatePreview}
             onReset={segmentation.resetToDefaults}
+            summaryIndexSetting={indexing.summaryIndexSetting}
+            onSummaryIndexSettingChange={indexing.setSummaryIndexSetting}
           />
         )}
         <Divider className="my-5" />

web/app/components/datasets/documents/detail/completed/common/batch-action.tsx

@@ -6,6 +6,7 @@ import { useTranslation } from 'react-i18next'
 import Button from '@/app/components/base/button'
 import Confirm from '@/app/components/base/confirm'
 import Divider from '@/app/components/base/divider'
+import { SearchLinesSparkle } from '@/app/components/base/icons/src/vender/knowledge'
 import { cn } from '@/utils/classnames'
 
 const i18nPrefix = 'batchAction'
@@ -15,6 +16,7 @@ type IBatchActionProps = {
   onBatchEnable: () => void
   onBatchDisable: () => void
   onBatchDelete: () => Promise<void>
+  onBatchSummary?: () => void
   onArchive?: () => void
   onEditMetadata?: () => void
   onBatchReIndex?: () => void
@@ -26,6 +28,7 @@ const BatchAction: FC<IBatchActionProps> = ({
   selectedIds,
   onBatchEnable,
   onBatchDisable,
+  onBatchSummary,
   onArchive,
   onBatchDelete,
   onEditMetadata,
@@ -82,7 +85,16 @@ const BatchAction: FC<IBatchActionProps> = ({
             <span className="px-0.5">{t('metadata.metadata', { ns: 'dataset' })}</span>
           </Button>
         )}
-
+        {onBatchSummary && (
+          <Button
+            variant="ghost"
+            className="gap-x-0.5 px-3"
+            onClick={onBatchSummary}
+          >
+            <SearchLinesSparkle className="size-4" />
+            <span className="px-0.5">{t('list.action.summary', { ns: 'datasetDocuments' })}</span>
+          </Button>
+        )}
         {onArchive && (
           <Button
             variant="ghost"

web/app/components/datasets/documents/detail/completed/common/summary-label.tsx

@@ -0,0 +1,26 @@
+import { memo } from 'react'
+import { useTranslation } from 'react-i18next'
+import { cn } from '@/utils/classnames'
+
+type SummaryLabelProps = {
+  summary?: string
+  className?: string
+}
+const SummaryLabel = ({
+  summary,
+  className,
+}: SummaryLabelProps) => {
+  const { t } = useTranslation()
+
+  return (
+    <div className={cn('space-y-1', className)}>
+      <div className="system-xs-medium-uppercase mt-2 flex items-center justify-between text-text-tertiary">
+        {t('segment.summary', { ns: 'datasetDocuments' })}
+        <div className="ml-2 h-px grow bg-divider-regular"></div>
+      </div>
+      <div className="body-xs-regular text-text-tertiary">{summary}</div>
+    </div>
+  )
+}
+
+export default memo(SummaryLabel)

web/app/components/datasets/documents/detail/completed/common/summary-status.tsx

@@ -0,0 +1,47 @@
+import { memo, useMemo } from 'react'
+import { useTranslation } from 'react-i18next'
+import Badge from '@/app/components/base/badge'
+import { SearchLinesSparkle } from '@/app/components/base/icons/src/vender/knowledge'
+import Tooltip from '@/app/components/base/tooltip'
+
+type SummaryStatusProps = {
+  status: string
+}
+
+const SummaryStatus = ({ status }: SummaryStatusProps) => {
+  const { t } = useTranslation()
+
+  const tip = useMemo(() => {
+    if (status === 'COMPLETED') {
+      return t('list.summary.ready', { ns: 'datasetDocuments' })
+    }
+    if (status === 'GENERATING') {
+      return t('list.summary.generatingSummary', { ns: 'datasetDocuments' })
+    }
+    return ''
+  }, [status, t])
+
+  return (
+    <Tooltip
+      popupContent={tip}
+    >
+      {
+        status === 'COMPLETED' && (
+          <Badge>
+            <SearchLinesSparkle className="h-3 w-3" />
+          </Badge>
+        )
+      }
+      {
+        status === 'GENERATING' && (
+          <Badge className="border-text-accent-secondary text-text-accent-secondary">
+            <SearchLinesSparkle className="mr-0.5 h-3 w-3" />
+            <span>{t('list.summary.generating', { ns: 'datasetDocuments' })}</span>
+          </Badge>
+        )
+      }
+    </Tooltip>
+  )
+}
+
+export default memo(SummaryStatus)

web/app/components/datasets/documents/detail/completed/common/summary-text.tsx

@@ -0,0 +1,35 @@
+import { memo } from 'react'
+import { useTranslation } from 'react-i18next'
+import Textarea from 'react-textarea-autosize'
+import { cn } from '@/utils/classnames'
+
+type SummaryTextProps = {
+  value?: string
+  onChange?: (value: string) => void
+  disabled?: boolean
+}
+const SummaryText = ({
+  value,
+  onChange,
+  disabled,
+}: SummaryTextProps) => {
+  const { t } = useTranslation()
+
+  return (
+    <div className="space-y-1">
+      <div className="system-xs-medium-uppercase text-text-tertiary">{t('segment.summary', { ns: 'datasetDocuments' })}</div>
+      <Textarea
+        className={cn(
+          'body-sm-regular w-full resize-none bg-transparent leading-6 text-text-secondary outline-none',
+        )}
+        placeholder={t('segment.summaryPlaceholder', { ns: 'datasetDocuments' })}
+        minRows={1}
+        value={value ?? ''}
+        onChange={e => onChange?.(e.target.value)}
+        disabled={disabled}
+      />
+    </div>
+  )
+}
+
+export default memo(SummaryText)

web/app/components/datasets/documents/detail/completed/index.tsx

@@ -322,6 +322,7 @@ const Completed: FC<ICompletedProps> = ({
     answer: string,
     keywords: string[],
     attachments: FileEntity[],
+    summary?: string,
     needRegenerate = false,
   ) => {
     const params: SegmentUpdater = { content: '', attachment_ids: [] }
@@ -351,6 +352,9 @@ const Completed: FC<ICompletedProps> = ({
       params.attachment_ids = attachments.map(item => item.uploadedId!)
     }
 
+    if (summary)
+      params.summary = summary
+
     if (needRegenerate)
       params.regenerate_child_chunks = needRegenerate
 
@@ -364,6 +368,7 @@ const Completed: FC<ICompletedProps> = ({
           if (seg.id === segmentId) {
             seg.answer = res.data.answer
             seg.content = res.data.content
+            seg.summary = res.data.summary
             seg.sign_content = res.data.sign_content
             seg.keywords = res.data.keywords
             seg.attachments = res.data.attachments

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

@@ -19,13 +19,14 @@ import { useDocumentContext } from '../../context'
 import ChildSegmentList from '../child-segment-list'
 import Dot from '../common/dot'
 import { SegmentIndexTag } from '../common/segment-index-tag'
+import SummaryLabel from '../common/summary-label'
 import Tag from '../common/tag'
 import ParentChunkCardSkeleton from '../skeleton/parent-chunk-card-skeleton'
 import ChunkContent from './chunk-content'
 
 type ISegmentCardProps = {
   loading: boolean
-  detail?: SegmentDetailModel & { document?: { name: string } }
+  detail?: SegmentDetailModel & { document?: { name: string }, status?: string }
   onClick?: () => void
   onChangeSwitch?: (enabled: boolean, segId?: string) => Promise<void>
   onDelete?: (segId: string) => Promise<void>
@@ -43,7 +44,7 @@ type ISegmentCardProps = {
 }
 
 const SegmentCard: FC<ISegmentCardProps> = ({
-  detail = {},
+  detail = { status: '' },
   onClick,
   onChangeSwitch,
   onDelete,
@@ -67,6 +68,7 @@ const SegmentCard: FC<ISegmentCardProps> = ({
     word_count,
     hit_count,
     answer,
+    summary,
     keywords,
     child_chunks = [],
     created_at,
@@ -237,6 +239,11 @@ const SegmentCard: FC<ISegmentCardProps> = ({
         className={contentOpacity}
       />
       {images.length > 0 && <ImageList images={images} size="md" className="py-1" />}
+      {
+        summary && (
+          <SummaryLabel summary={summary} className="mt-2" />
+        )
+      }
       {isGeneralMode && (
         <div className={cn('flex flex-wrap items-center gap-2 py-1.5', contentOpacity)}>
           {keywords?.map(keyword => <Tag key={keyword} text={keyword} />)}

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

@@ -25,6 +25,7 @@ import Dot from './common/dot'
 import Keywords from './common/keywords'
 import RegenerationModal from './common/regeneration-modal'
 import { SegmentIndexTag } from './common/segment-index-tag'
+import SummaryText from './common/summary-text'
 import { useSegmentListContext } from './index'
 
 type ISegmentDetailProps = {
@@ -35,6 +36,7 @@ type ISegmentDetailProps = {
     a: string,
     k: string[],
     attachments: FileEntity[],
+    summary?: string,
     needRegenerate?: boolean,
   ) => void
   onCancel: () => void
@@ -57,6 +59,7 @@ const SegmentDetail: FC<ISegmentDetailProps> = ({
   const { t } = useTranslation()
   const [question, setQuestion] = useState(isEditMode ? segInfo?.content || '' : segInfo?.sign_content || '')
   const [answer, setAnswer] = useState(segInfo?.answer || '')
+  const [summary, setSummary] = useState(segInfo?.summary || '')
   const [attachments, setAttachments] = useState<FileEntity[]>(() => {
     return segInfo?.attachments?.map(item => ({
       id: uuid4(),
@@ -91,8 +94,8 @@ const SegmentDetail: FC<ISegmentDetailProps> = ({
   }, [onCancel])
 
   const handleSave = useCallback(() => {
-    onUpdate(segInfo?.id || '', question, answer, keywords, attachments)
-  }, [onUpdate, segInfo?.id, question, answer, keywords, attachments])
+    onUpdate(segInfo?.id || '', question, answer, keywords, attachments, summary, false)
+  }, [onUpdate, segInfo?.id, question, answer, keywords, attachments, summary])
 
   const handleRegeneration = useCallback(() => {
     setShowRegenerationModal(true)
@@ -111,8 +114,8 @@ const SegmentDetail: FC<ISegmentDetailProps> = ({
   }, [onCancel, onModalStateChange])
 
   const onConfirmRegeneration = useCallback(() => {
-    onUpdate(segInfo?.id || '', question, answer, keywords, attachments, true)
-  }, [onUpdate, segInfo?.id, question, answer, keywords, attachments])
+    onUpdate(segInfo?.id || '', question, answer, keywords, attachments, summary, true)
+  }, [onUpdate, segInfo?.id, question, answer, keywords, attachments, summary])
 
   const onAttachmentsChange = useCallback((attachments: FileEntity[]) => {
     setAttachments(attachments)
@@ -197,6 +200,11 @@ const SegmentDetail: FC<ISegmentDetailProps> = ({
             value={attachments}
             onChange={onAttachmentsChange}
           />
+          <SummaryText
+            value={summary}
+            onChange={summary => setSummary(summary)}
+            disabled={!isEditMode}
+          />
           {isECOIndexing && (
             <Keywords
               className="w-full"

web/app/components/datasets/documents/list.tsx

@@ -27,7 +27,7 @@ import { useDatasetDetailContextWithSelector as useDatasetDetailContext } from '
 import useTimestamp from '@/hooks/use-timestamp'
 import { ChunkingMode, DataSourceType, DocumentActionType } from '@/models/datasets'
 import { DatasourceType } from '@/models/pipeline'
-import { useDocumentArchive, useDocumentBatchRetryIndex, useDocumentDelete, useDocumentDisable, useDocumentEnable } from '@/service/knowledge/use-document'
+import { useDocumentArchive, useDocumentBatchRetryIndex, useDocumentDelete, useDocumentDisable, useDocumentEnable, useDocumentSummary } from '@/service/knowledge/use-document'
 import { asyncRunSafe } from '@/utils'
 import { cn } from '@/utils/classnames'
 import { formatNumber } from '@/utils/format'
@@ -35,6 +35,7 @@ import FileTypeIcon from '../../base/file-uploader/file-type-icon'
 import ChunkingModeLabel from '../common/chunking-mode-label'
 import useBatchEditDocumentMetadata from '../metadata/hooks/use-batch-edit-document-metadata'
 import BatchAction from './detail/completed/common/batch-action'
+import SummaryStatus from './detail/completed/common/summary-status'
 import Operations from './operations'
 import RenameModal from './rename-modal'
 import StatusItem from './status-item'
@@ -218,6 +219,7 @@ const DocumentList: FC<IDocumentListProps> = ({
       onSelectedIdChange(uniq([...selectedIds, ...localDocs.map(doc => doc.id)]))
   }, [isAllSelected, localDocs, onSelectedIdChange, selectedIds])
   const { mutateAsync: archiveDocument } = useDocumentArchive()
+  const { mutateAsync: generateSummary } = useDocumentSummary()
   const { mutateAsync: enableDocument } = useDocumentEnable()
   const { mutateAsync: disableDocument } = useDocumentDisable()
   const { mutateAsync: deleteDocument } = useDocumentDelete()
@@ -230,6 +232,9 @@ const DocumentList: FC<IDocumentListProps> = ({
         case DocumentActionType.archive:
           opApi = archiveDocument
           break
+        case DocumentActionType.summary:
+          opApi = generateSummary
+          break
         case DocumentActionType.enable:
           opApi = enableDocument
           break
@@ -409,6 +414,13 @@ const DocumentList: FC<IDocumentListProps> = ({
                       >
                         <span className="grow-1 truncate text-sm">{doc.name}</span>
                       </Tooltip>
+                      {
+                        doc.summary_index_status && (
+                          <div className="ml-1 hidden shrink-0 group-hover:flex">
+                            <SummaryStatus status={doc.summary_index_status} />
+                          </div>
+                        )
+                      }
                       <div className="hidden shrink-0 group-hover:ml-auto group-hover:flex">
                         <Tooltip
                           popupContent={t('list.table.rename', { ns: 'datasetDocuments' })}
@@ -461,6 +473,7 @@ const DocumentList: FC<IDocumentListProps> = ({
           className="absolute bottom-16 left-0 z-20"
           selectedIds={selectedIds}
           onArchive={handleAction(DocumentActionType.archive)}
+          onBatchSummary={handleAction(DocumentActionType.summary)}
           onBatchEnable={handleAction(DocumentActionType.enable)}
           onBatchDisable={handleAction(DocumentActionType.disable)}
           onBatchDelete={handleAction(DocumentActionType.delete)}

web/app/components/datasets/documents/operations.tsx

@@ -17,6 +17,7 @@ import * as React from 'react'
 import { useCallback, useState } from 'react'
 import { useTranslation } from 'react-i18next'
 import { useContext } from 'use-context-selector'
+import { SearchLinesSparkle } from '@/app/components/base/icons/src/vender/knowledge'
 import { DataSourceType, DocumentActionType } from '@/models/datasets'
 import {
   useDocumentArchive,
@@ -25,6 +26,7 @@ import {
   useDocumentEnable,
   useDocumentPause,
   useDocumentResume,
+  useDocumentSummary,
   useDocumentUnArchive,
   useSyncDocument,
   useSyncWebsite,
@@ -82,6 +84,7 @@ const Operations = ({
   const { mutateAsync: deleteDocument } = useDocumentDelete()
   const { mutateAsync: syncDocument } = useSyncDocument()
   const { mutateAsync: syncWebsite } = useSyncWebsite()
+  const { mutateAsync: generateSummary } = useDocumentSummary()
   const { mutateAsync: pauseDocument } = useDocumentPause()
   const { mutateAsync: resumeDocument } = useDocumentResume()
   const isListScene = scene === 'list'
@@ -107,6 +110,9 @@ const Operations = ({
         else
           opApi = syncWebsite
         break
+      case 'summary':
+        opApi = generateSummary
+        break
       case 'pause':
         opApi = pauseDocument
         break
@@ -220,6 +226,10 @@ const Operations = ({
                         <span className={s.actionName}>{t('list.action.sync', { ns: 'datasetDocuments' })}</span>
                       </div>
                     )}
+                    <div className={s.actionItem} onClick={() => onOperate('summary')}>
+                      <SearchLinesSparkle className="h-4 w-4 text-text-tertiary" />
+                      <span className={s.actionName}>{t('list.action.summary', { ns: 'datasetDocuments' })}</span>
+                    </div>
                     <Divider className="my-1" />
                   </>
                 )}

web/app/components/datasets/documents/types.ts

@@ -1 +1 @@
-export type OperationName = 'delete' | 'archive' | 'enable' | 'disable' | 'sync' | 'un_archive' | 'pause' | 'resume'
+export type OperationName = 'delete' | 'archive' | 'enable' | 'disable' | 'sync' | 'un_archive' | 'pause' | 'resume' | 'summary'

web/app/components/datasets/hit-testing/components/result-item.tsx

@@ -7,6 +7,7 @@ import * as React from 'react'
 import { useMemo } from 'react'
 import { useTranslation } from 'react-i18next'
 import { Markdown } from '@/app/components/base/markdown'
+import SummaryLabel from '@/app/components/datasets/documents/detail/completed/common/summary-label'
 import Tag from '@/app/components/datasets/documents/detail/completed/common/tag'
 import { extensionToFileType } from '@/app/components/datasets/hit-testing/utils/extension-to-file-type'
 import { cn } from '@/utils/classnames'
@@ -25,7 +26,7 @@ const ResultItem = ({
   payload,
 }: ResultItemProps) => {
   const { t } = useTranslation()
-  const { segment, score, child_chunks, files } = payload
+  const { segment, score, child_chunks, files, summary } = payload
   const data = segment
   const { position, word_count, content, sign_content, keywords, document } = data
   const isParentChildRetrieval = !!(child_chunks && child_chunks.length > 0)
@@ -98,6 +99,9 @@ const ResultItem = ({
             ))}
           </div>
         )}
+        {summary && (
+          <SummaryLabel summary={summary} className="mt-2" />
+        )}
       </div>
       {/* Foot */}
       <ResultItemFooter docType={fileType} docTitle={document.name} showDetailModal={showDetailModal} />

web/i18n/en-US/dataset-documents.json

@@ -30,6 +30,7 @@
   "list.action.pause": "Pause",
   "list.action.resume": "Resume",
   "list.action.settings": "Chunking Settings",
+  "list.action.summary": "Generate summary",
   "list.action.sync": "Sync",
   "list.action.unarchive": "Unarchive",
   "list.action.uploadFile": "Upload new file",
@@ -74,6 +75,9 @@
   "list.status.indexing": "Indexing",
   "list.status.paused": "Paused",
   "list.status.queuing": "Queuing",
+  "list.summary.generating": "Generating...",
+  "list.summary.generatingSummary": "Generating summary",
+  "list.summary.ready": "Summary ready",
   "list.table.header.action": "ACTION",
   "list.table.header.chunkingMode": "CHUNKING MODE",
   "list.table.header.fileName": "NAME",
@@ -328,5 +332,7 @@
   "segment.searchResults_one": "RESULT",
   "segment.searchResults_other": "RESULTS",
   "segment.searchResults_zero": "RESULT",
+  "segment.summary": "SUMMARY",
+  "segment.summaryPlaceholder": "Write a brief summary for better retrieval…",
   "segment.vectorHash": "Vector hash: "
 }

web/i18n/zh-Hans/dataset-documents.json

@@ -30,6 +30,7 @@
   "list.action.pause": "暂停",
   "list.action.resume": "恢复",
   "list.action.settings": "分段设置",
+  "list.action.summary": "生成摘要",
   "list.action.sync": "同步",
   "list.action.unarchive": "撤销归档",
   "list.action.uploadFile": "上传新文件",
@@ -74,6 +75,9 @@
   "list.status.indexing": "索引中",
   "list.status.paused": "已暂停",
   "list.status.queuing": "排队中",
+  "list.summary.generating": "生成中...",
+  "list.summary.generatingSummary": "生成摘要中",
+  "list.summary.ready": "摘要已生成",
   "list.table.header.action": "操作",
   "list.table.header.chunkingMode": "分段模式",
   "list.table.header.fileName": "名称",
@@ -328,5 +332,7 @@
   "segment.searchResults_one": "搜索结果",
   "segment.searchResults_other": "搜索结果",
   "segment.searchResults_zero": "搜索结果",
+  "segment.summary": "摘要",
+  "segment.summaryPlaceholder": "写一个简短的摘要，以便更好地检索…",
   "segment.vectorHash": "向量哈希："
 }

web/models/datasets.ts

@@ -401,6 +401,7 @@ export type InitialDocumentDetail = {
   total_segments?: number
   doc_form: ChunkingMode
   doc_language: string
+  summary_index_status?: string
 }
 
 export type SimpleDocumentDetail = InitialDocumentDetail & {
@@ -434,6 +435,7 @@ export type DocumentReq = {
   doc_form: ChunkingMode
   doc_language: string
   process_rule: ProcessRule
+  summary_index_setting?: SummaryIndexSetting
 }
 
 export type CreateDocumentReq = DocumentReq & {
@@ -585,6 +587,7 @@ export type SegmentDetailModel = {
   error: string | null
   stopped_at: number
   answer?: string
+  summary?: string
   child_chunks?: ChildChunkDetail[]
   updated_at: number
   attachments: Attachment[]
@@ -628,6 +631,7 @@ export type HitTesting = {
   tsne_position: TsnePosition
   child_chunks: HitTestingChildChunk[] | null
   files: Attachment[]
+  summary?: string
 }
 
 export type ExternalKnowledgeBaseHitTesting = {
@@ -707,6 +711,7 @@ export type RelatedAppResponse = {
 export type SegmentUpdater = {
   content: string
   answer?: string
+  summary?: string
   keywords?: string[]
   regenerate_child_chunks?: boolean
   attachment_ids?: string[]
@@ -788,6 +793,7 @@ export enum DocumentActionType {
   archive = 'archive',
   unArchive = 'un_archive',
   delete = 'delete',
+  summary = 'summary',
 }
 
 export type UpdateDocumentBatchParams = {

web/service/knowledge/use-document.ts

@@ -107,6 +107,18 @@ export const useSyncDocument = () => {
   })
 }
 
+export const useDocumentSummary = () => {
+  return useMutation({
+    mutationFn: ({ datasetId, documentIds, documentId }: UpdateDocumentBatchParams) => {
+      return post<CommonResponse>(`/datasets/${datasetId}/documents/generate-summary`, {
+        body: {
+          document_list: documentId ? [documentId] : documentIds!,
+        },
+      })
+    },
+  })
+}
+
 export const useSyncWebsite = () => {
   return useMutation({
     mutationFn: ({ datasetId, documentId }: UpdateDocumentBatchParams) => {