feat: add /api/v1/language endpoint for Go/Python runtime detection (#16952)

Both backends serve GET /api/v1/language. Frontend calls it once and
caches. By this way, front end can know the backend is go or python and
thus can determine which part of logic to load.

---------

Co-authored-by: Claude <noreply@anthropic.com>

web/src/hooks/use-user-setting-request.tsx

@@ -10,6 +10,11 @@ import {
 } from '@/interfaces/database/user-setting';
 import { ISetLangfuseConfigRequestBody } from '@/interfaces/request/system';
 import { DEFAULT_LANGUAGE_CODE, supportedLanguages } from '@/locales/config';
+import kbService from '@/services/knowledge-service';
+import {
+  getBackendLanguage,
+  subscribeBackendLanguage,
+} from '@/utils/backend-runtime';
 import userService, {
   addTenantUser,
   agreeTenant,
@@ -19,6 +24,7 @@ import userService, {
 } from '@/services/user-service';
 import { useMutation, useQuery, useQueryClient } from '@tanstack/react-query';
 import { useCallback, useMemo, useState } from 'react';
+import { useSyncExternalStore } from 'react';
 import { useTranslation } from 'react-i18next';
 
 import { useWarnEmptyModel } from './use-warn-empty-model';
@@ -38,6 +44,7 @@ export const enum UserSettingApiAction {
   AgreeTenant = 'agreeTenant',
   SetLangfuseConfig = 'setLangfuseConfig',
   DeleteLangfuseConfig = 'deleteLangfuseConfig',
+  ListPipelines = 'listPipelines',
   FetchLangfuseConfig = 'fetchLangfuseConfig',
 }
 
@@ -102,6 +109,27 @@ export const useSelectParserList = (): Array<{
   const { data: tenantInfo } = useFetchTenantInfo(true);
   const { t } = useTranslation();
 
+  // Detect backend runtime language (Go vs Python) so we can choose
+  // the matching parser-list code path at runtime.
+  // fetchBackendLanguage / getBackendLanguage handle their own caching
+  // internally; no need for an extra useQuery layer.
+  const backendLang = useSyncExternalStore(
+    subscribeBackendLanguage,
+    getBackendLanguage,
+    getBackendLanguage,
+  );
+
+  // Go backend: fetch pipeline catalog dynamically.
+  const { data: pipelineListData } = useQuery({
+    queryKey: [UserSettingApiAction.ListPipelines],
+    queryFn: async () => {
+      const { data } = await kbService.listPipelines();
+      return data;
+    },
+    staleTime: Infinity,
+    enabled: backendLang === 'go',
+  });
+
   const defaultParsers = useMemo(
     () => [
       { value: 'naive', label: t('knowledgeConfiguration.parserLabel.naive') },
@@ -135,6 +163,25 @@ export const useSelectParserList = (): Array<{
   );
 
   const parserList = useMemo(() => {
+    // Go backend: prefer the dynamic pipeline catalog from the API.
+    if (backendLang === 'go') {
+      const pipelineList: Array<{ parser_id: string; title: string; dsl: Record<string, any> }> =
+        pipelineListData?.data ?? [];
+      if (pipelineList.length > 0) {
+        const labelFromAPI = (parserId: string, title: string) => {
+          const key = `knowledgeConfiguration.parserLabel.${parserId}`;
+          const translated = t(key);
+          return translated !== key ? translated : title;
+        };
+        return pipelineList.map((item) => ({
+          value: item.parser_id,
+          label: labelFromAPI(item.parser_id, item.title),
+        }));
+      }
+    }
+
+    // Python backend (or fallback): use tenant-level parser_ids or
+    // the hardcoded default parsers.
     const parserArray: Array<string> = tenantInfo?.parser_ids?.split(',') ?? [];
     const filteredArray = parserArray.filter((x) => x.trim() !== '');
 
@@ -146,7 +193,7 @@ export const useSelectParserList = (): Array<{
       const arr = x.split(':');
       return { value: arr[0], label: arr[1] };
     });
-  }, [tenantInfo, defaultParsers]);
+  }, [tenantInfo, defaultParsers, backendLang, pipelineListData, t]);
 
   return parserList;
 };

web/src/locales/en.ts

@@ -778,6 +778,7 @@ Example: A 1 KB message with 1024-dim embedding uses ~9 KB. The 5 MB default lim
         'Re-parse existing documents for the new column roles to take effect.',
       parserLabel: {
         naive: 'General',
+        general: 'General',
         qa: 'Q&A',
         resume: 'Resume',
         manual: 'Manual',

web/src/services/knowledge-service.ts

@@ -19,6 +19,7 @@ const {
   documentThumbnails,
   documentIngest,
   listTagByKnowledgeIds,
+  listPipelines,
   setMeta,
   getMeta,
   getMetaKeys,
@@ -66,6 +67,10 @@ const methods = {
     url: retrievalTestShare,
     method: 'post',
   },
+  listPipelines: {
+    url: listPipelines,
+    method: 'get',
+  },
   pipelineRerun: {
     url: api.pipelineRerun,
     method: 'post',

web/src/utils/api.ts

@@ -181,6 +181,7 @@ export default {
     `${restAPIv1}/datasets/${datasetId}/ingestions/${logId}`,
   fetchPipelineDatasetLogs: (datasetId: string) =>
     `${restAPIv1}/datasets/${datasetId}/ingestions`,
+  listPipelines: `${restAPIv1}/pipelines?type=builtin`,
   runIndex: (datasetId: string, indexType: string) =>
     `${restAPIv1}/datasets/${datasetId}/index?type=${indexType.toLowerCase()}`,
   traceIndex: (datasetId: string, indexType: string) =>

web/src/utils/backend-runtime.ts

@@ -0,0 +1,40 @@
+/**
+ * Backend runtime language detection.
+ *
+ * Fetches /api/v1/language once at module load (app start) and caches the
+ * result. Components subscribe via React's useSyncExternalStore; there is no
+ * per-component fetch or useEffect.
+ *
+ * Pattern: module-level fetch + listener set, same subscription approach as
+ * enterprise billingStatus.ts.
+ */
+
+type Listener = () => void;
+const listeners = new Set<Listener>();
+
+let backendLanguage: string | null = null;
+
+// Kick off the fetch at module load — app start, not component mount.
+const promise: Promise<string> = fetch('/api/v1/language')
+  .then((r) => r.json())
+  .then((body: { data?: { language?: string } }) => {
+    backendLanguage = body.data?.language === 'go' ? 'go' : 'python';
+    listeners.forEach((fn) => fn());
+    return backendLanguage;
+  })
+  .catch(() => {
+    backendLanguage = 'python';
+    listeners.forEach((fn) => fn());
+    return 'python';
+  });
+
+export const fetchBackendLanguage = (): Promise<string> => promise;
+
+export const getBackendLanguage = (): string | null => backendLanguage;
+
+export const subscribeBackendLanguage = (listener: Listener): (() => void) => {
+  listeners.add(listener);
+  return () => {
+    listeners.delete(listener);
+  };
+};