feat: n to 1 retrieval legacy (#6554)

This commit is contained in:
zxhlyh
2024-07-24 12:50:48 +08:00
committed by GitHub
parent e4bb943fe5
commit 6fe9aa69cc
35 changed files with 1329 additions and 249 deletions

View File

@@ -1,18 +1,20 @@
'use client'
import type { FC } from 'react'
import React, { useState } from 'react'
import { useTranslation } from 'react-i18next'
import { RiDeleteBinLine } from '@remixicon/react'
import {
RiDeleteBinLine,
RiEditLine,
} from '@remixicon/react'
import SettingsModal from '../settings-modal'
import type { DataSet } from '@/models/datasets'
import { DataSourceType } from '@/models/datasets'
import { formatNumber } from '@/utils/format'
import FileIcon from '@/app/components/base/file-icon'
import { Settings01 } from '@/app/components/base/icons/src/vender/line/general'
import { Folder } from '@/app/components/base/icons/src/vender/solid/files'
import { Globe06 } from '@/app/components/base/icons/src/vender/solid/mapsAndTravel'
import Drawer from '@/app/components/base/drawer'
import useBreakpoints, { MediaType } from '@/hooks/use-breakpoints'
import Badge from '@/app/components/base/badge'
import { useKnowledge } from '@/hooks/use-knowledge'
type ItemProps = {
className?: string
@@ -27,12 +29,10 @@ const Item: FC<ItemProps> = ({
onSave,
onRemove,
}) => {
const { t } = useTranslation()
const media = useBreakpoints()
const isMobile = media === MediaType.mobile
const [showSettingsModal, setShowSettingsModal] = useState(false)
const { formatIndexingTechniqueAndMethod } = useKnowledge()
const handleSave = (newDataset: DataSet) => {
onSave(newDataset)
@@ -65,22 +65,17 @@ const Item: FC<ItemProps> = ({
<div className='grow'>
<div className='flex items-center h-[18px]'>
<div className='grow text-[13px] font-medium text-gray-800 truncate' title={config.name}>{config.name}</div>
<div className='shrink-0 text-xs text-gray-500'>
{formatNumber(config.word_count)} {t('appDebug.feature.dataSet.words')} · {formatNumber(config.document_count)} {t('appDebug.feature.dataSet.textBlocks')}
</div>
<Badge
text={formatIndexingTechniqueAndMethod(config.indexing_technique, config.retrieval_model_dict?.search_method)}
/>
</div>
{/* {
config.description && (
<div className='text-xs text-gray-500'>{config.description}</div>
)
} */}
</div>
<div className='hidden rounded-lg group-hover:flex items-center justify-end absolute right-0 top-0 bottom-0 pr-2 w-[124px] bg-gradient-to-r from-white/50 to-white to-50%'>
<div
className='flex items-center justify-center mr-1 w-6 h-6 hover:bg-black/5 rounded-md cursor-pointer'
onClick={() => setShowSettingsModal(true)}
>
<Settings01 className='w-4 h-4 text-gray-500' />
<RiEditLine className='w-4 h-4 text-gray-500' />
</div>
<div
className='group/action flex items-center justify-center w-6 h-6 hover:bg-[#FEE4E2] rounded-md cursor-pointer'

View File

@@ -44,7 +44,8 @@ const DatasetConfig: FC = () => {
const handleSave = (newDataset: DataSet) => {
const index = dataSet.findIndex(item => item.id === newDataset.id)
setDataSet([...dataSet.slice(0, index), newDataset, ...dataSet.slice(index + 1)])
const newDatasets = [...dataSet.slice(0, index), newDataset, ...dataSet.slice(index + 1)]
setDataSet(newDatasets)
formattingChangedDispatcher()
}
@@ -74,7 +75,7 @@ const DatasetConfig: FC = () => {
title={t('appDebug.feature.dataSet.title')}
headerRight={
<div className='flex items-center gap-1'>
{!isAgent && <ParamsConfig />}
{!isAgent && <ParamsConfig disabled={!hasData} selectedDatasets={dataSet} />}
<OperationBtn type="add" onClick={showSelectDataSet} />
</div>
}

View File

@@ -1,10 +1,12 @@
'use client'
import React from 'react'
import { memo, useMemo } from 'react'
import type { FC } from 'react'
import { useTranslation } from 'react-i18next'
import {
RiQuestionLine,
} from '@remixicon/react'
import WeightedScore from './weighted-score'
import TopKItem from '@/app/components/base/param-item/top-k-item'
import ScoreThresholdItem from '@/app/components/base/param-item/score-threshold-item'
import RadioCard from '@/app/components/base/radio-card/simple'
@@ -16,13 +18,20 @@ import {
import type {
DatasetConfigs,
} from '@/models/debug'
import ModelSelector from '@/app/components/header/account-setting/model-provider-page/model-selector'
import { useModelListAndDefaultModelAndCurrentProviderAndModel } from '@/app/components/header/account-setting/model-provider-page/hooks'
import type { ModelConfig } from '@/app/components/workflow/types'
import ModelParameterModal from '@/app/components/header/account-setting/model-provider-page/model-parameter-modal'
import TooltipPlus from '@/app/components/base/tooltip-plus'
import { ModelTypeEnum } from '@/app/components/header/account-setting/model-provider-page/declarations'
import type {
DataSet,
WeightedScoreEnum,
} from '@/models/datasets'
import { RerankingModeEnum } from '@/models/datasets'
import cn from '@/utils/classnames'
import { useSelectedDatasetsMode } from '@/app/components/workflow/nodes/knowledge-retrieval/hooks'
import Switch from '@/app/components/base/switch'
type Props = {
datasetConfigs: DatasetConfigs
@@ -31,6 +40,7 @@ type Props = {
singleRetrievalModelConfig?: ModelConfig
onSingleRetrievalModelChange?: (config: ModelConfig) => void
onSingleRetrievalModelParamsChange?: (config: ModelConfig) => void
selectedDatasets?: DataSet[]
}
const ConfigContent: FC<Props> = ({
@@ -40,8 +50,10 @@ const ConfigContent: FC<Props> = ({
singleRetrievalModelConfig: singleRetrievalConfig = {} as ModelConfig,
onSingleRetrievalModelChange = () => { },
onSingleRetrievalModelParamsChange = () => { },
selectedDatasets = [],
}) => {
const { t } = useTranslation()
const selectedDatasetsMode = useSelectedDatasetsMode(selectedDatasets)
const type = datasetConfigs.retrieval_model
const setType = (value: RETRIEVE_TYPE) => {
onChange({
@@ -54,7 +66,7 @@ const ConfigContent: FC<Props> = ({
defaultModel: rerankDefaultModel,
} = useModelListAndDefaultModelAndCurrentProviderAndModel(ModelTypeEnum.rerank)
const rerankModel = (() => {
if (datasetConfigs.reranking_model) {
if (datasetConfigs.reranking_model?.reranking_provider_name) {
return {
provider_name: datasetConfigs.reranking_model.reranking_provider_name,
model_name: datasetConfigs.reranking_model.reranking_model_name,
@@ -93,14 +105,73 @@ const ConfigContent: FC<Props> = ({
})
}
const handleWeightedScoreChange = (value: { type: WeightedScoreEnum; value: number[] }) => {
const configs = {
...datasetConfigs,
weights: {
...datasetConfigs.weights!,
weight_type: value.type,
vector_setting: {
...datasetConfigs.weights!.vector_setting!,
vector_weight: value.value[0],
},
keyword_setting: {
keyword_weight: value.value[1],
},
},
}
onChange(configs)
}
const handleRerankModeChange = (mode: RerankingModeEnum) => {
onChange({
...datasetConfigs,
reranking_mode: mode,
})
}
const model = singleRetrievalConfig
const rerankingModeOptions = [
{
value: RerankingModeEnum.WeightedScore,
label: t('dataset.weightedScore.title'),
tips: t('dataset.weightedScore.description'),
},
{
value: RerankingModeEnum.RerankingModel,
label: t('common.modelProvider.rerankModel.key'),
tips: t('common.modelProvider.rerankModel.tip'),
},
]
const showWeightedScore = selectedDatasetsMode.allHighQuality
&& !selectedDatasetsMode.inconsistentEmbeddingModel
const showWeightedScorePanel = showWeightedScore && datasetConfigs.reranking_mode === RerankingModeEnum.WeightedScore && datasetConfigs.weights
const selectedRerankMode = datasetConfigs.reranking_mode || RerankingModeEnum.RerankingModel
const showRerankModel = useMemo(() => {
if (datasetConfigs.reranking_enable === false && selectedDatasetsMode.allEconomic)
return false
return true
}, [datasetConfigs.reranking_enable, selectedDatasetsMode.allEconomic])
return (
<div>
<div className='system-xl-semibold text-text-primary'>{t('dataset.retrievalSettings')}</div>
<div className='mt-2 space-y-3'>
<RadioCard
icon={<NTo1Retrieval className='shrink-0 mr-3 w-9 h-9 rounded-lg' />}
title={t('appDebug.datasetConfig.retrieveOneWay.title')}
title={(
<div className='flex items-center'>
{t('appDebug.datasetConfig.retrieveOneWay.title')}
<TooltipPlus popupContent={<div className='w-[320px]'>{t('dataset.nTo1RetrievalLegacy')}</div>}>
<div className='ml-1 flex items-center px-[5px] h-[18px] rounded-[5px] border border-text-accent-secondary system-2xs-medium-uppercase text-text-accent-secondary'>legacy</div>
</TooltipPlus>
</div>
)}
description={t('appDebug.datasetConfig.retrieveOneWay.description')}
isChosen={type === RETRIEVE_TYPE.oneWay}
onChosen={() => { setType(RETRIEVE_TYPE.oneWay) }}
@@ -115,43 +186,152 @@ const ConfigContent: FC<Props> = ({
</div>
{type === RETRIEVE_TYPE.multiWay && (
<>
<div className='mt-6'>
<div className='leading-[32px] text-[13px] font-medium text-gray-900'>{t('common.modelProvider.rerankModel.key')}</div>
<div>
<ModelSelector
defaultModel={rerankModel && { provider: rerankModel?.provider_name, model: rerankModel?.model_name }}
onSelect={(v) => {
onChange({
...datasetConfigs,
reranking_model: {
reranking_provider_name: v.provider,
reranking_model_name: v.model,
},
})
}}
modelList={rerankModelList}
/>
</div>
</div>
<div className='mt-4 space-y-4'>
<TopKItem
value={datasetConfigs.top_k}
onChange={handleParamChange}
enable={true}
/>
<ScoreThresholdItem
value={datasetConfigs.score_threshold as number}
onChange={handleParamChange}
enable={datasetConfigs.score_threshold_enabled}
hasSwitch={true}
onSwitchChange={handleSwitch}
/>
<div className='mb-2 mt-4 h-[1px] bg-divider-subtle'></div>
<div
className='flex items-center mb-2 h-6 system-md-semibold text-text-secondary'
>
{t('dataset.rerankSettings')}
</div>
{
selectedDatasetsMode.inconsistentEmbeddingModel
&& (
<div className='mt-4 system-xs-regular text-text-warning'>
{t('dataset.inconsistentEmbeddingModelTip')}
</div>
)
}
{
selectedDatasetsMode.mixtureHighQualityAndEconomic
&& (
<div className='mt-4 system-xs-regular text-text-warning'>
{t('dataset.mixtureHighQualityAndEconomicTip')}
</div>
)
}
{
showWeightedScore && (
<div className='flex items-center justify-between'>
{
rerankingModeOptions.map(option => (
<div
key={option.value}
className={cn(
'flex items-center justify-center w-[calc((100%-8px)/2)] h-8 rounded-lg border border-components-option-card-option-border bg-components-option-card-option-bg cursor-pointer system-sm-medium text-text-secondary',
selectedRerankMode === option.value && 'border-[1.5px] border-components-option-card-option-selected-border bg-components-option-card-option-selected-bg text-text-primary',
)}
onClick={() => handleRerankModeChange(option.value)}
>
<div className='truncate'>{option.label}</div>
<TooltipPlus
popupContent={<div className='w-[200px]'>{option.tips}</div>}
hideArrow
>
<RiQuestionLine className='ml-0.5 w-3.5 h-4.5 text-text-quaternary' />
</TooltipPlus>
</div>
))
}
</div>
)
}
{
!showWeightedScorePanel && (
<div className='mt-2'>
<div className='flex items-center'>
{
selectedDatasetsMode.allEconomic && (
<Switch
size='md'
defaultValue={showRerankModel}
onChange={(v) => {
onChange({
...datasetConfigs,
reranking_enable: v,
})
}}
/>
)
}
<div className='ml-2 leading-[32px] text-[13px] font-medium text-gray-900'>{t('common.modelProvider.rerankModel.key')}</div>
<TooltipPlus popupContent={<div className="w-[200px]">{t('common.modelProvider.rerankModel.tip')}</div>}>
<RiQuestionLine className='ml-0.5 w-[14px] h-[14px] text-gray-400' />
</TooltipPlus>
</div>
<div>
<ModelSelector
defaultModel={rerankModel && { provider: rerankModel?.provider_name, model: rerankModel?.model_name }}
onSelect={(v) => {
onChange({
...datasetConfigs,
reranking_model: {
reranking_provider_name: v.provider,
reranking_model_name: v.model,
},
})
}}
modelList={rerankModelList}
/>
</div>
</div>
)
}
{
showWeightedScorePanel
&& (
<div className='mt-2 space-y-4'>
<WeightedScore
value={{
type: datasetConfigs.weights!.weight_type,
value: [
datasetConfigs.weights!.vector_setting.vector_weight,
datasetConfigs.weights!.keyword_setting.keyword_weight,
],
}}
onChange={handleWeightedScoreChange}
/>
<TopKItem
value={datasetConfigs.top_k}
onChange={handleParamChange}
enable={true}
/>
<ScoreThresholdItem
value={datasetConfigs.score_threshold as number}
onChange={handleParamChange}
enable={datasetConfigs.score_threshold_enabled}
hasSwitch={true}
onSwitchChange={handleSwitch}
/>
</div>
)
}
{
!showWeightedScorePanel
&& (
<div className='mt-4 space-y-4'>
<TopKItem
value={datasetConfigs.top_k}
onChange={handleParamChange}
enable={true}
/>
{
showRerankModel && (
<ScoreThresholdItem
value={datasetConfigs.score_threshold as number}
onChange={handleParamChange}
enable={datasetConfigs.score_threshold_enabled}
hasSwitch={true}
onSwitchChange={handleSwitch}
/>
)
}
</div>
)
}
</>
)}
{isInWorkflow && type === RETRIEVE_TYPE.oneWay && (
<div className='mt-6'>
<div className='mt-4'>
<div className='flex items-center space-x-0.5'>
<div className='leading-[32px] text-[13px] font-medium text-gray-900'>{t('common.modelProvider.systemReasoningModel.key')}</div>
<TooltipPlus
@@ -180,4 +360,4 @@ const ConfigContent: FC<Props> = ({
</div >
)
}
export default React.memo(ConfigContent)
export default memo(ConfigContent)

View File

@@ -1,29 +1,73 @@
'use client'
import type { FC } from 'react'
import { memo, useState } from 'react'
import { memo, useEffect, useState } from 'react'
import { useTranslation } from 'react-i18next'
import { useContext } from 'use-context-selector'
import { RiEqualizer2Line } from '@remixicon/react'
import ConfigContent from './config-content'
import cn from '@/utils/classnames'
import { Settings04 } from '@/app/components/base/icons/src/vender/line/general'
import ConfigContext from '@/context/debug-configuration'
import Modal from '@/app/components/base/modal'
import Button from '@/app/components/base/button'
import { RETRIEVE_TYPE } from '@/types/app'
import Toast from '@/app/components/base/toast'
import { DATASET_DEFAULT } from '@/config'
import { useModelListAndDefaultModelAndCurrentProviderAndModel } from '@/app/components/header/account-setting/model-provider-page/hooks'
import { ModelTypeEnum } from '@/app/components/header/account-setting/model-provider-page/declarations'
import type { DataSet } from '@/models/datasets'
import type { DatasetConfigs } from '@/models/debug'
import {
getMultipleRetrievalConfig,
getSelectedDatasetsMode,
} from '@/app/components/workflow/nodes/knowledge-retrieval/utils'
const ParamsConfig: FC = () => {
type ParamsConfigProps = {
disabled?: boolean
selectedDatasets: DataSet[]
}
const ParamsConfig = ({
disabled,
selectedDatasets,
}: ParamsConfigProps) => {
const { t } = useTranslation()
const [open, setOpen] = useState(false)
const {
datasetConfigs,
setDatasetConfigs,
rerankSettingModalOpen,
setRerankSettingModalOpen,
} = useContext(ConfigContext)
const [tempDataSetConfigs, setTempDataSetConfigs] = useState(datasetConfigs)
useEffect(() => {
const {
allEconomic,
} = getSelectedDatasetsMode(selectedDatasets)
const { datasets, retrieval_model, score_threshold_enabled, ...restConfigs } = datasetConfigs
let rerankEnable = restConfigs.reranking_enable
if (allEconomic && !restConfigs.reranking_model?.reranking_provider_name && rerankEnable === undefined)
rerankEnable = false
setTempDataSetConfigs({
...getMultipleRetrievalConfig({
top_k: restConfigs.top_k,
score_threshold: restConfigs.score_threshold,
reranking_model: restConfigs.reranking_model && {
provider: restConfigs.reranking_model.reranking_provider_name,
model: restConfigs.reranking_model.reranking_model_name,
},
reranking_mode: restConfigs.reranking_mode,
weights: restConfigs.weights,
reranking_enable: rerankEnable,
}, selectedDatasets),
reranking_model: restConfigs.reranking_model && {
reranking_provider_name: restConfigs.reranking_model.reranking_provider_name,
reranking_model_name: restConfigs.reranking_model.reranking_model_name,
},
retrieval_model,
score_threshold_enabled,
datasets,
})
}, [selectedDatasets, datasetConfigs])
const {
defaultModel: rerankDefaultModel,
currentModel: isRerankDefaultModelVaild,
@@ -55,45 +99,68 @@ const ParamsConfig: FC = () => {
} as any
}
setDatasetConfigs(config)
setOpen(false)
setRerankSettingModalOpen(false)
}
const handleSetTempDataSetConfigs = (newDatasetConfigs: DatasetConfigs) => {
const { datasets, retrieval_model, score_threshold_enabled, ...restConfigs } = newDatasetConfigs
const retrievalConfig = getMultipleRetrievalConfig({
top_k: restConfigs.top_k,
score_threshold: restConfigs.score_threshold,
reranking_model: restConfigs.reranking_model && {
provider: restConfigs.reranking_model.reranking_provider_name,
model: restConfigs.reranking_model.reranking_model_name,
},
reranking_mode: restConfigs.reranking_mode,
weights: restConfigs.weights,
reranking_enable: restConfigs.reranking_enable,
}, selectedDatasets)
setTempDataSetConfigs({
...retrievalConfig,
reranking_model: restConfigs.reranking_model && {
reranking_provider_name: restConfigs.reranking_model.reranking_provider_name,
reranking_model_name: restConfigs.reranking_model.reranking_model_name,
},
retrieval_model,
score_threshold_enabled,
datasets,
})
}
return (
<div>
<div
className={cn('flex items-center rounded-md h-7 px-3 space-x-1 text-gray-700 cursor-pointer hover:bg-gray-200', open && 'bg-gray-200')}
<Button
variant='ghost'
size='small'
className={cn('h-7', rerankSettingModalOpen && 'bg-components-button-ghost-bg-hover')}
onClick={() => {
setTempDataSetConfigs({
...datasetConfigs,
top_k: datasetConfigs.top_k || DATASET_DEFAULT.top_k,
score_threshold: datasetConfigs.score_threshold || DATASET_DEFAULT.score_threshold,
})
setOpen(true)
setRerankSettingModalOpen(true)
}}
disabled={disabled}
>
<Settings04 className="w-[14px] h-[14px]" />
<div className='text-xs font-medium'>
{t('appDebug.datasetConfig.params')}
</div>
</div>
<RiEqualizer2Line className='mr-1 w-3.5 h-3.5' />
{t('dataset.retrievalSettings')}
</Button>
{
open && (
rerankSettingModalOpen && (
<Modal
isShow={open}
isShow={rerankSettingModalOpen}
onClose={() => {
setOpen(false)
setRerankSettingModalOpen(false)
}}
className='sm:min-w-[528px]'
title={t('appDebug.datasetConfig.settingTitle')}
>
<ConfigContent
datasetConfigs={tempDataSetConfigs}
onChange={setTempDataSetConfigs}
onChange={handleSetTempDataSetConfigs}
selectedDatasets={selectedDatasets}
/>
<div className='mt-6 flex justify-end'>
<Button className='mr-2 flex-shrink-0' onClick={() => {
setOpen(false)
setRerankSettingModalOpen(false)
}}>{t('common.operation.cancel')}</Button>
<Button variant='primary' className='flex-shrink-0' onClick={handleSave} >{t('common.operation.save')}</Button>
</div>

View File

@@ -0,0 +1,112 @@
import { memo, useCallback } from 'react'
import { useTranslation } from 'react-i18next'
import {
DEFAULT_WEIGHTED_SCORE,
WeightedScoreEnum,
} from '@/models/datasets'
import Slider from '@/app/components/base/slider'
import cn from '@/utils/classnames'
const formatNumber = (value: number) => {
if (value > 0 && value < 1)
return `0.${value * 10}`
else if (value === 1)
return '1.0'
return value
}
type Value = {
type: WeightedScoreEnum
value: number[]
}
type WeightedScoreProps = {
value: Value
onChange: (value: Value) => void
}
const WeightedScore = ({
value,
onChange = () => {},
}: WeightedScoreProps) => {
const { t } = useTranslation()
const options = [
{
value: WeightedScoreEnum.SemanticFirst,
label: t('dataset.weightedScore.semanticFirst'),
},
{
value: WeightedScoreEnum.KeywordFirst,
label: t('dataset.weightedScore.keywordFirst'),
},
{
value: WeightedScoreEnum.Customized,
label: t('dataset.weightedScore.customized'),
},
]
const disabled = value.type !== WeightedScoreEnum.Customized
const handleTypeChange = useCallback((type: WeightedScoreEnum) => {
const result = { ...value, type }
if (type === WeightedScoreEnum.SemanticFirst)
result.value = [DEFAULT_WEIGHTED_SCORE.semanticFirst.semantic, DEFAULT_WEIGHTED_SCORE.semanticFirst.keyword]
if (type === WeightedScoreEnum.KeywordFirst)
result.value = [DEFAULT_WEIGHTED_SCORE.keywordFirst.semantic, DEFAULT_WEIGHTED_SCORE.keywordFirst.keyword]
onChange(result)
}, [value, onChange])
return (
<div>
<div className='flex items-center mb-1 space-x-4'>
{
options.map(option => (
<div
key={option.value}
className='flex py-1.5 max-w-[calc((100%-32px)/3)] system-sm-regular text-text-secondary cursor-pointer'
onClick={() => handleTypeChange(option.value)}
>
<div
className={cn(
'shrink-0 mr-2 w-4 h-4 bg-components-radio-bg border border-components-radio-border rounded-full shadow-xs',
value.type === option.value && 'border-[5px] border-components-radio-border-checked',
)}
></div>
<div className='truncate' title={option.label}>{option.label}</div>
</div>
))
}
</div>
<div className='flex items-center px-3 h-9 space-x-3 rounded-lg border border-components-panel-border'>
<div className='shrink-0 flex items-center w-[90px] system-xs-semibold-uppercase text-util-colors-blue-blue-500'>
<div className='mr-1 truncate uppercase' title={t('dataset.weightedScore.semantic') || ''}>
{t('dataset.weightedScore.semantic')}
</div>
{formatNumber(value.value[0])}
</div>
<Slider
className={cn('grow h-0.5 bg-gradient-to-r from-[#53B1FD] to-[#2ED3B7]', disabled && 'cursor-not-allowed')}
max={1.0}
min={0}
step={0.1}
value={value.value[0]}
onChange={v => onChange({ type: value.type, value: [v, (10 - v * 10) / 10] })}
disabled={disabled}
thumbClassName={cn(disabled && '!cursor-not-allowed')}
trackClassName='!bg-transparent'
/>
<div className='shrink-0 flex items-center justify-end w-[90px] system-xs-semibold-uppercase text-util-colors-cyan-cyan-500'>
{formatNumber(value.value[1])}
<div className='ml-1 truncate uppercase' title={t('dataset.weightedScore.keyword') || ''}>
{t('dataset.weightedScore.keyword')}
</div>
</div>
</div>
</div>
)
}
export default memo(WeightedScore)

View File

@@ -13,7 +13,8 @@ import type { DataSet } from '@/models/datasets'
import Button from '@/app/components/base/button'
import { fetchDatasets } from '@/service/datasets'
import Loading from '@/app/components/base/loading'
import { formatNumber } from '@/utils/format'
import Badge from '@/app/components/base/badge'
import { useKnowledge } from '@/hooks/use-knowledge'
export type ISelectDataSetProps = {
isShow: boolean
@@ -38,6 +39,7 @@ const SelectDataSet: FC<ISelectDataSetProps> = ({
const listRef = useRef<HTMLDivElement>(null)
const [page, setPage, getPage] = useGetState(1)
const [isNoMore, setIsNoMore] = useState(false)
const { formatIndexingTechniqueAndMethod } = useKnowledge()
useInfiniteScroll(
async () => {
@@ -45,7 +47,7 @@ const SelectDataSet: FC<ISelectDataSetProps> = ({
const { data, has_more } = await fetchDatasets({ url: '/datasets', params: { page } })
setPage(getPage() + 1)
setIsNoMore(!has_more)
const newList = [...(datasets || []), ...data]
const newList = [...(datasets || []), ...data.filter(item => item.indexing_technique)]
setDataSets(newList)
setLoaded(true)
if (!selected.find(item => !item.name))
@@ -136,14 +138,13 @@ const SelectDataSet: FC<ISelectDataSetProps> = ({
<span className='ml-1 shrink-0 px-1 border boder-gray-200 rounded-md text-gray-500 text-xs font-normal leading-[18px]'>{t('dataset.unavailable')}</span>
)}
</div>
<div className={cn('shrink-0 flex text-xs text-gray-500 overflow-hidden whitespace-nowrap', !item.embedding_available && 'opacity-50')}>
<span className='max-w-[100px] overflow-hidden text-ellipsis whitespace-nowrap'>{formatNumber(item.word_count)}</span>
{t('appDebug.feature.dataSet.words')}
<span className='px-0.5'>·</span>
<span className='max-w-[100px] min-w-[8px] overflow-hidden text-ellipsis whitespace-nowrap'>{formatNumber(item.document_count)} </span>
{t('appDebug.feature.dataSet.textBlocks')}
</div>
{
item.indexing_technique && (
<Badge
text={formatIndexingTechniqueAndMethod(item.indexing_technique, item.retrieval_model_dict?.search_method)}
/>
)
}
</div>
))}
</div>

View File

@@ -259,7 +259,7 @@ const SettingsModal: FC<SettingsModalProps> = ({
{/* Retrieval Method Config */}
<div className={rowClass}>
<div className={labelClass}>
<div className={cn(labelClass, 'w-auto min-w-[168px]')}>
<div>
<div>{t('datasetSettings.form.retrievalSetting.title')}</div>
<div className='leading-[18px] text-xs font-normal text-gray-500'>
@@ -268,7 +268,7 @@ const SettingsModal: FC<SettingsModalProps> = ({
</div>
</div>
</div>
<div className='w-[480px]'>
<div>
{indexMethod === 'high_quality'
? (
<RetrievalMethodConfig

View File

@@ -46,7 +46,7 @@ import { fetchDatasets } from '@/service/datasets'
import { useProviderContext } from '@/context/provider-context'
import { AgentStrategy, AppType, ModelModeType, RETRIEVE_TYPE, Resolution, TransferMethod } from '@/types/app'
import { PromptMode } from '@/models/debug'
import { ANNOTATION_DEFAULT, DEFAULT_AGENT_SETTING, DEFAULT_CHAT_PROMPT_CONFIG, DEFAULT_COMPLETION_PROMPT_CONFIG } from '@/config'
import { ANNOTATION_DEFAULT, DATASET_DEFAULT, DEFAULT_AGENT_SETTING, DEFAULT_CHAT_PROMPT_CONFIG, DEFAULT_COMPLETION_PROMPT_CONFIG } from '@/config'
import SelectDataSet from '@/app/components/app/configuration/dataset-config/select-dataset'
import { useModalContext } from '@/context/modal-context'
import useBreakpoints, { MediaType } from '@/hooks/use-breakpoints'
@@ -57,6 +57,10 @@ import { useTextGenerationCurrentProviderAndModelAndModelList } from '@/app/comp
import { fetchCollectionList } from '@/service/tools'
import { type Collection } from '@/app/components/tools/types'
import { useStore as useAppStore } from '@/app/components/app/store'
import {
getMultipleRetrievalConfig,
getSelectedDatasetsMode,
} from '@/app/components/workflow/nodes/knowledge-retrieval/utils'
type PublishConfig = {
modelConfig: ModelConfig
@@ -174,14 +178,14 @@ const Configuration: FC = () => {
}, [])
const [datasetConfigs, setDatasetConfigs] = useState<DatasetConfigs>({
retrieval_model: RETRIEVE_TYPE.oneWay,
retrieval_model: RETRIEVE_TYPE.multiWay,
reranking_model: {
reranking_provider_name: '',
reranking_model_name: '',
},
top_k: 2,
top_k: DATASET_DEFAULT.top_k,
score_threshold_enabled: false,
score_threshold: 0.7,
score_threshold: DATASET_DEFAULT.score_threshold,
datasets: {
datasets: [],
},
@@ -202,6 +206,7 @@ const Configuration: FC = () => {
const hasSetContextVar = !!contextVar
const [isShowSelectDataSet, { setTrue: showSelectDataSet, setFalse: hideSelectDataSet }] = useBoolean(false)
const selectedIds = dataSets.map(item => item.id)
const [rerankSettingModalOpen, setRerankSettingModalOpen] = useState(false)
const handleSelect = (data: DataSet[]) => {
if (isEqual(data.map(item => item.id), dataSets.map(item => item.id))) {
hideSelectDataSet()
@@ -209,6 +214,7 @@ const Configuration: FC = () => {
}
formattingChangedDispatcher()
let newDatasets = data
if (data.find(item => !item.name)) { // has not loaded selected dataset
const newSelected = produce(data, (draft: any) => {
data.forEach((item, index) => {
@@ -220,11 +226,45 @@ const Configuration: FC = () => {
})
})
setDataSets(newSelected)
newDatasets = newSelected
}
else {
setDataSets(data)
}
hideSelectDataSet()
const {
allEconomic,
mixtureHighQualityAndEconomic,
inconsistentEmbeddingModel,
} = getSelectedDatasetsMode(newDatasets)
if (allEconomic || mixtureHighQualityAndEconomic || inconsistentEmbeddingModel)
setRerankSettingModalOpen(true)
const { datasets, retrieval_model, score_threshold_enabled, ...restConfigs } = datasetConfigs
const retrievalConfig = getMultipleRetrievalConfig({
top_k: restConfigs.top_k,
score_threshold: restConfigs.score_threshold,
reranking_model: restConfigs.reranking_model && {
provider: restConfigs.reranking_model.reranking_provider_name,
model: restConfigs.reranking_model.reranking_model_name,
},
reranking_mode: restConfigs.reranking_mode,
weights: restConfigs.weights,
reranking_enable: restConfigs.reranking_enable,
}, newDatasets)
setDatasetConfigs({
...retrievalConfig,
reranking_model: restConfigs.reranking_model && {
reranking_provider_name: restConfigs.reranking_model.reranking_provider_name,
reranking_model_name: restConfigs.reranking_model.reranking_model_name,
},
retrieval_model,
score_threshold_enabled,
datasets,
})
}
const [isShowHistoryModal, { setTrue: showHistoryModal, setFalse: hideHistoryModal }] = useBoolean(false)
@@ -509,7 +549,7 @@ const Configuration: FC = () => {
syncToPublishedConfig(config)
setPublishedConfig(config)
setDatasetConfigs({
retrieval_model: RETRIEVE_TYPE.oneWay,
retrieval_model: RETRIEVE_TYPE.multiWay,
...modelConfig.dataset_configs,
})
setHasFetchedDetail(true)
@@ -744,6 +784,8 @@ const Configuration: FC = () => {
isShowVisionConfig,
visionConfig,
setVisionConfig: handleSetVisionConfig,
rerankSettingModalOpen,
setRerankSettingModalOpen,
}}
>
<>