Merge branch 'feat/rag-2' of https://github.com/langgenius/dify into feat/rag-2

api/core/app/apps/pipeline/pipeline_generator.py

@@ -639,7 +639,8 @@ class PipelineGenerator(BaseAppGenerator):
                 raise GenerateTaskStoppedError()
             else:
                 logger.exception(
-                    f"Fails to process generate task pipeline, task_id: {application_generate_entity.task_id}"
+                    "Fails to process generate task pipeline, task_id: %r",
+                    application_generate_entity.task_id,
                 )
                 raise e
 

api/core/helper/name_generator.py

@@ -17,7 +17,7 @@ def generate_provider_name(
             f"{credential_type.get_name()}",
         )
     except Exception as e:
-        logger.warning(f"Error generating next provider name for {fallback_context}: {str(e)}")
+        logger.warning("Error generating next provider name for %r: %r", fallback_context, e)
         return f"{credential_type.get_name()} 1"
 
 

api/services/rag_pipeline/pipeline_template/remote/remote_retrieval.py

@@ -20,7 +20,7 @@ class RemotePipelineTemplateRetrieval(PipelineTemplateRetrievalBase):
         try:
             result = self.fetch_pipeline_template_detail_from_dify_official(pipeline_id)
         except Exception as e:
-            logger.warning(f"fetch recommended app detail from dify official failed: {e}, switch to built-in.")
+            logger.warning("fetch recommended app detail from dify official failed: %r, switch to built-in.", e)
             result = BuildInRecommendAppRetrieval.fetch_recommended_app_detail_from_builtin(pipeline_id)
         return result
 
@@ -28,7 +28,7 @@ class RemotePipelineTemplateRetrieval(PipelineTemplateRetrievalBase):
         try:
             result = self.fetch_pipeline_templates_from_dify_official(language)
         except Exception as e:
-            logger.warning(f"fetch pipeline templates from dify official failed: {e}, switch to built-in.")
+            logger.warning("fetch pipeline templates from dify official failed: %r, switch to built-in.", e)
             result = BuildInRecommendAppRetrieval.fetch_recommended_apps_from_builtin(language)
         return result
 

web/app/components/workflow/nodes/knowledge-base/components/retrieval-setting/hooks.tsx

@@ -46,7 +46,7 @@ export const useRetrievalSetting = (indexMethod?: IndexMethodEnum) => {
   }, [t])
   const InvertedIndexOption: Option = useMemo(() => {
     return {
-      id: RetrievalSearchMethodEnum.keywordSearch,
+      id: RetrievalSearchMethodEnum.invertedIndex,
       icon: HybridSearch as any,
       title: t('dataset.retrieval.invertedIndex.title'),
       description: t('dataset.retrieval.invertedIndex.description'),

web/app/components/workflow/nodes/knowledge-base/components/retrieval-setting/search-method-option.tsx

@@ -194,7 +194,7 @@ const SearchMethodOption = ({
           isScoreThresholdEnabled={isScoreThresholdEnabled}
           onScoreThresholdEnabledChange={onScoreThresholdEnabledChange}
           readonly={readonly}
-          hiddenScoreThreshold={searchMethod === RetrievalSearchMethodEnum.keywordSearch}
+          hiddenScoreThreshold={searchMethod === RetrievalSearchMethodEnum.invertedIndex}
         />
       </div>
     </OptionCard>

web/app/components/workflow/nodes/knowledge-base/hooks/use-config.ts

@@ -15,6 +15,7 @@ import type {
   KnowledgeBaseNodeType,
   RerankingModel,
 } from '../types'
+import { isHighQualitySearchMethod } from '../utils'
 
 export const useConfig = (id: string) => {
   const store = useStoreApi()
@@ -36,10 +37,18 @@ export const useConfig = (id: string) => {
 
   const handleChunkStructureChange = useCallback((chunkStructure: ChunkStructureEnum) => {
     const nodeData = getNodeData()
-    const { indexing_technique } = nodeData?.data
+    const {
+      indexing_technique,
+      retrieval_model,
+    } = nodeData?.data
+    const { search_method } = retrieval_model || {}
     handleNodeDataUpdate({
       chunk_structure: chunkStructure,
       indexing_technique: (chunkStructure === ChunkStructureEnum.parent_child || chunkStructure === ChunkStructureEnum.question_answer) ? IndexMethodEnum.QUALIFIED : indexing_technique,
+      retrieval_model: {
+        ...retrieval_model,
+        search_method: ((chunkStructure === ChunkStructureEnum.parent_child || chunkStructure === ChunkStructureEnum.question_answer) && !isHighQualitySearchMethod(search_method)) ? RetrievalSearchMethodEnum.semantic : search_method,
+      },
     })
   }, [handleNodeDataUpdate, getNodeData])
 
@@ -50,7 +59,7 @@ export const useConfig = (id: string) => {
       draft.indexing_technique = indexMethod
 
       if (indexMethod === IndexMethodEnum.ECONOMICAL)
-        draft.retrieval_model.search_method = RetrievalSearchMethodEnum.keywordSearch
+        draft.retrieval_model.search_method = RetrievalSearchMethodEnum.invertedIndex
       else if (indexMethod === IndexMethodEnum.QUALIFIED)
         draft.retrieval_model.search_method = RetrievalSearchMethodEnum.semantic
     }))

web/app/components/workflow/nodes/knowledge-base/hooks/use-settings-display.ts

@@ -13,6 +13,6 @@ export const useSettingsDisplay = () => {
     [RetrievalSearchMethodEnum.semantic]: t('dataset.retrieval.semantic_search.title'),
     [RetrievalSearchMethodEnum.fullText]: t('dataset.retrieval.full_text_search.title'),
     [RetrievalSearchMethodEnum.hybrid]: t('dataset.retrieval.hybrid_search.title'),
-    [RetrievalSearchMethodEnum.keywordSearch]: t('dataset.retrieval.invertedIndex.title'),
+    [RetrievalSearchMethodEnum.invertedIndex]: t('dataset.retrieval.invertedIndex.title'),
   }
 }

web/app/components/workflow/nodes/knowledge-base/utils.ts

@@ -0,0 +1,9 @@
+import {
+  RetrievalSearchMethodEnum,
+} from './types'
+
+export const isHighQualitySearchMethod = (searchMethod: RetrievalSearchMethodEnum) => {
+  return searchMethod === RetrievalSearchMethodEnum.semantic
+    || searchMethod === RetrievalSearchMethodEnum.hybrid
+    || searchMethod === RetrievalSearchMethodEnum.fullText
+}