feat: Store pdfparse in local (#5534)

projects/app/src/pageComponents/app/detail/SimpleApp/ChatTest.tsx

@@ -45,7 +45,7 @@ const ChatTest = ({ appForm, setRenderEdit }: Props) => {
     setRenderEdit(!datasetCiteData);
   }, [datasetCiteData, setRenderEdit]);
 
-  const { ChatContainer, restartChat, loading } = useChatTest({
+  const { ChatContainer, restartChat } = useChatTest({
     ...workflowData,
     chatConfig: appForm.chatConfig,
     isReady: true

projects/app/src/pageComponents/app/detail/SimpleApp/EditForm.tsx

@@ -71,7 +71,6 @@ const EditForm = ({
   const { appDetail } = useContextSelector(AppContext, (v) => v);
   const selectDatasets = useMemo(() => appForm?.dataset?.datasets, [appForm]);
   const [, startTst] = useTransition();
-  const { llmModelList, defaultModels } = useSystemStore();
 
   const {
     isOpen: isOpenDatasetSelect,

projects/app/src/pageComponents/dataset/detail/Import/Context.tsx

@@ -1,5 +1,5 @@
 import { useRouter } from 'next/router';
-import { type SetStateAction, useMemo, useState } from 'react';
+import { type SetStateAction, useEffect, useMemo, useState } from 'react';
 import { useTranslation } from 'next-i18next';
 import { createContext, useContextSelector } from 'use-context-selector';
 import {
@@ -20,6 +20,7 @@ import { DatasetPageContext } from '@/web/core/dataset/context/datasetPageContex
 import { DataChunkSplitModeEnum } from '@fastgpt/global/core/dataset/constants';
 import { chunkAutoChunkSize, getAutoIndexSize } from '@fastgpt/global/core/dataset/training/utils';
 import { type CollectionChunkFormType } from '../Form/CollectionChunkForm';
+import { useLocalStorageState } from 'ahooks';
 
 export type ImportFormType = {
   customPdfParse: boolean;
@@ -38,7 +39,7 @@ type DatasetImportContextType = {
 };
 
 export const defaultFormData: ImportFormType = {
-  customPdfParse: false,
+  customPdfParse: true,
 
   trainingType: DatasetCollectionDataProcessModeEnum.chunk,
 
@@ -198,12 +199,23 @@ const DatasetImportContextProvider = ({ children }: { children: React.ReactNode
 
   const vectorModel = datasetDetail.vectorModel;
 
+  const [localCustomPdfParse, setLocalCustomPdfParse] = useLocalStorageState(
+    'dataset_customPdfParse',
+    {
+      defaultValue: true
+    }
+  );
   const processParamsForm = useForm<ImportFormType>({
     defaultValues: (() => ({
       ...defaultFormData,
+      customPdfParse: localCustomPdfParse,
       indexSize: getAutoIndexSize(vectorModel)
     }))()
   });
+  const customPdfParse = processParamsForm.watch('customPdfParse');
+  useEffect(() => {
+    setLocalCustomPdfParse(customPdfParse);
+  }, [customPdfParse, setLocalCustomPdfParse]);
 
   const [sources, setSources] = useState<ImportSourceItemType[]>([]);
 

projects/app/src/pageComponents/dataset/detail/Import/diffSource/ReTraining.tsx

@@ -46,7 +46,7 @@ const ReTraining = () => {
       ]);
 
       processParamsForm.reset({
-        customPdfParse: collection.customPdfParse || false,
+        customPdfParse: collection.customPdfParse ?? defaultFormData.customPdfParse,
         trainingType: collection.trainingType,
 
         chunkTriggerType: collection.chunkTriggerType || defaultFormData.chunkTriggerType,