estel/dify
1
0
Fork 0
Code Issues Pull Requests Actions 1 Packages Projects Releases Wiki Activity

Fix/retrieval setting weight default value (#9622)

Browse Source
This commit is contained in:
zxhlyh
2024-10-22 18:31:39 +08:00
committed by GitHub
parent 7d7e0f9800
commit ff956cb546
7 changed files with 86 additions and 75 deletions
Download Patch File Download Diff File Expand all files Collapse all files

12
web/app/components/workflow/nodes/knowledge-retrieval/use-config.ts
View File

@@ -163,7 +163,7 @@ const useConfig = (id: string, payload: KnowledgeRetrievalNodeType) => {
draft.retrieval_mode = newMode
if (newMode === RETRIEVE_TYPE.multiWay) {
const multipleRetrievalConfig = draft.multiple_retrieval_config
draft.multiple_retrieval_config = getMultipleRetrievalConfig(multipleRetrievalConfig!, selectedDatasets)
draft.multiple_retrieval_config = getMultipleRetrievalConfig(multipleRetrievalConfig!, selectedDatasets, selectedDatasets, !!currentRerankModel)
}
else {
const hasSetModel = draft.single_retrieval_config?.model?.provider
@@ -180,14 +180,14 @@ const useConfig = (id: string, payload: KnowledgeRetrievalNodeType) => {
}
})
setInputs(newInputs)
}, [currentModel?.model, currentModel?.model_properties?.mode, currentProvider?.provider, inputs, setInputs, selectedDatasets])
}, [currentModel?.model, currentModel?.model_properties?.mode, currentProvider?.provider, inputs, setInputs, selectedDatasets, currentRerankModel])
const handleMultipleRetrievalConfigChange = useCallback((newConfig: MultipleRetrievalConfig) => {
const newInputs = produce(inputs, (draft) => {
draft.multiple_retrieval_config = getMultipleRetrievalConfig(newConfig!, selectedDatasets)
draft.multiple_retrieval_config = getMultipleRetrievalConfig(newConfig!, selectedDatasets, selectedDatasets, !!currentRerankModel)
})
setInputs(newInputs)
}, [inputs, setInputs, selectedDatasets])
}, [inputs, setInputs, selectedDatasets, currentRerankModel])
// datasets
useEffect(() => {
@@ -231,7 +231,7 @@ const useConfig = (id: string, payload: KnowledgeRetrievalNodeType) => {
if (payload.retrieval_mode === RETRIEVE_TYPE.multiWay && newDatasets.length > 0) {
const multipleRetrievalConfig = draft.multiple_retrieval_config
draft.multiple_retrieval_config = getMultipleRetrievalConfig(multipleRetrievalConfig!, newDatasets)
draft.multiple_retrieval_config = getMultipleRetrievalConfig(multipleRetrievalConfig!, newDatasets, selectedDatasets, !!currentRerankModel)
}
})
setInputs(newInputs)
@@ -243,7 +243,7 @@ const useConfig = (id: string, payload: KnowledgeRetrievalNodeType) => {
|| (allExternal && newDatasets.length > 1)
)
setRerankModelOpen(true)
}, [inputs, setInputs, payload.retrieval_mode])
}, [inputs, setInputs, payload.retrieval_mode, selectedDatasets, currentRerankModel])
const filterVar = useCallback((varPayload: Var) => {
return varPayload.type === VarType.string

49
web/app/components/workflow/nodes/knowledge-retrieval/utils.ts
View File

@@ -1,4 +1,7 @@
import { uniq } from 'lodash-es'
import {
uniq,
xorBy,
} from 'lodash-es'
import type { MultipleRetrievalConfig } from './types'
import type {
DataSet,
@@ -15,7 +18,9 @@ export const checkNodeValid = () => {
return true
}
export const getSelectedDatasetsMode = (datasets: DataSet[]) => {
export const getSelectedDatasetsMode = (datasets: DataSet[] = []) => {
if (datasets === null)
datasets = []
let allHighQuality = true
let allHighQualityVectorSearch = true
let allHighQualityFullTextSearch = true
@@ -85,7 +90,14 @@ export const getSelectedDatasetsMode = (datasets: DataSet[]) => {
} as SelectedDatasetsMode
}
export const getMultipleRetrievalConfig = (multipleRetrievalConfig: MultipleRetrievalConfig, selectedDatasets: DataSet[]) => {
export const getMultipleRetrievalConfig = (
multipleRetrievalConfig: MultipleRetrievalConfig,
selectedDatasets: DataSet[],
originalDatasets: DataSet[],
isValidRerankModel?: boolean,
) => {
const shouldSetWeightDefaultValue = xorBy(selectedDatasets, originalDatasets, 'id').length > 0
const {
allHighQuality,
allHighQualityVectorSearch,
@@ -123,6 +135,37 @@ export const getMultipleRetrievalConfig = (multipleRetrievalConfig: MultipleRetr
result.reranking_mode = RerankingModeEnum.WeightedScore
if (allHighQuality && !inconsistentEmbeddingModel && (reranking_mode === RerankingModeEnum.WeightedScore || reranking_mode === undefined) && allInternal && !weights) {
if (!isValidRerankModel)
result.reranking_mode = RerankingModeEnum.WeightedScore
else
result.reranking_mode = RerankingModeEnum.RerankingModel
result.weights = {
vector_setting: {
vector_weight: allHighQualityVectorSearch
? DEFAULT_WEIGHTED_SCORE.allHighQualityVectorSearch.semantic
: allHighQualityFullTextSearch
? DEFAULT_WEIGHTED_SCORE.allHighQualityFullTextSearch.semantic
: DEFAULT_WEIGHTED_SCORE.other.semantic,
embedding_provider_name: selectedDatasets[0].embedding_model_provider,
embedding_model_name: selectedDatasets[0].embedding_model,
},
keyword_setting: {
keyword_weight: allHighQualityVectorSearch
? DEFAULT_WEIGHTED_SCORE.allHighQualityVectorSearch.keyword
: allHighQualityFullTextSearch
? DEFAULT_WEIGHTED_SCORE.allHighQualityFullTextSearch.keyword
: DEFAULT_WEIGHTED_SCORE.other.keyword,
},
}
}
if (shouldSetWeightDefaultValue && allHighQuality && !inconsistentEmbeddingModel && (reranking_mode === RerankingModeEnum.WeightedScore || reranking_mode === undefined || !isValidRerankModel) && allInternal && weights) {
if (!isValidRerankModel)
result.reranking_mode = RerankingModeEnum.WeightedScore
else
result.reranking_mode = RerankingModeEnum.RerankingModel
result.weights = {
vector_setting: {
vector_weight: allHighQualityVectorSearch