Refa: better UX for adding OCR model (#12034)

### What problem does this PR solve?

Better UX for adding OCR model.

### Type of change

- [x] Refactoring

web/src/locales/en.ts

@@ -1110,6 +1110,9 @@ Example: Virtual Hosted Style`,
       mcp: 'MCP',
       mineru: {
         modelNameRequired: 'Model name is required',
+        apiServerRequired: 'MinerU API Server Configuration is required',
+        serverUrlBackendLimit:
+          'MinerU Server URL Address is only available for the HTTP client backend',
         apiserver: 'MinerU API Server Configuration',
         outputDir: 'MinerU Output Directory Path',
         backend: 'MinerU Processing Backend Type',
@@ -1121,6 +1124,11 @@ Example: Virtual Hosted Style`,
           vlmTransformers: 'Vision Language Model with Transformers',
           vlmVllmEngine: 'Vision Language Model with vLLM Engine',
           vlmHttpClient: 'Vision Language Model via HTTP Client',
+          vlmMlxEngine: 'Vision Language Model with MLX Engine',
+          vlmVllmAsyncEngine:
+            'Vision Language Model with vLLM Async Engine (Experimental)',
+          vlmLmdeployEngine:
+            'Vision Language Model with LMDeploy Engine (Experimental)',
         },
       },
     },

web/src/locales/zh.ts

@@ -959,6 +959,8 @@ General：实体和关系提取提示来自 GitHub - microsoft/graphrag：基于
       mcp: 'MCP',
       mineru: {
         modelNameRequired: '模型名称为必填项',
+        apiServerRequired: 'MinerU API服务器配置为必填项',
+        serverUrlBackendLimit: '仅在backend 为vlm-http-client 时可填写',
         apiserver: 'MinerU API服务器配置',
         outputDir: 'MinerU输出目录路径',
         backend: 'MinerU处理后端类型',
@@ -970,6 +972,9 @@ General：实体和关系提取提示来自 GitHub - microsoft/graphrag：基于
           vlmTransformers: '基于Transformers的视觉语言模型',
           vlmVllmEngine: '基于vLLM引擎的视觉语言模型',
           vlmHttpClient: '通过HTTP客户端连接的视觉语言模型',
+          vlmMlxEngine: '基于MLX引擎的视觉语言模型',
+          vlmVllmAsyncEngine: '基于vLLM异步引擎的视觉语言模型（实验性）',
+          vlmLmdeployEngine: '基于LMDeploy引擎的视觉语言模型（实验性）',
         },
       },
     },

web/src/pages/user-setting/setting-model/hooks.tsx

@@ -472,10 +472,13 @@ export const useSubmitMinerU = () => {
 
   const onMineruOk = useCallback(
     async (payload: MinerUFormValues) => {
-      const cfg = {
+      const cfg: any = {
         ...payload,
         mineru_delete_output: payload.mineru_delete_output ?? true ? '1' : '0',
       };
+      if (payload.mineru_backend !== 'vlm-http-client') {
+        delete cfg.mineru_server_url;
+      }
       const req: IAddLlmRequestBody = {
         llm_factory: LLMFactory.MinerU,
         llm_name: payload.llm_name,

web/src/pages/user-setting/setting-model/modal/mineru-modal/index.tsx

@@ -16,7 +16,7 @@ import { IModalProps } from '@/interfaces/common';
 import { buildOptions } from '@/utils/form';
 import { zodResolver } from '@hookform/resolvers/zod';
 import { t } from 'i18next';
-import { useForm } from 'react-hook-form';
+import { useForm, useWatch } from 'react-hook-form';
 import { useTranslation } from 'react-i18next';
 import { z } from 'zod';
 import { LLMHeader } from '../../components/llm-header';
@@ -25,15 +25,18 @@ const FormSchema = z.object({
   llm_name: z.string().min(1, {
     message: t('setting.mineru.modelNameRequired'),
   }),
-  mineru_apiserver: z.string().optional(),
+  mineru_apiserver: z.string().url(),
   mineru_output_dir: z.string().optional(),
   mineru_backend: z.enum([
     'pipeline',
     'vlm-transformers',
     'vlm-vllm-engine',
     'vlm-http-client',
+    'vlm-mlx-engine',
+    'vlm-vllm-async-engine',
+    'vlm-lmdeploy-engine',
   ]),
-  mineru_server_url: z.string().optional(),
+  mineru_server_url: z.string().url().optional(),
   mineru_delete_output: z.boolean(),
 });
 
@@ -52,6 +55,9 @@ const MinerUModal = ({
     'vlm-transformers',
     'vlm-vllm-engine',
     'vlm-http-client',
+    'vlm-mlx-engine',
+    'vlm-vllm-async-engine',
+    'vlm-lmdeploy-engine',
   ]);
 
   const form = useForm<MinerUFormValues>({
@@ -62,6 +68,11 @@ const MinerUModal = ({
     },
   });
 
+  const backend = useWatch({
+    control: form.control,
+    name: 'mineru_backend',
+  });
+
   const handleOk = async (values: MinerUFormValues) => {
     const ret = await onOk?.(values as any);
     if (ret) {
@@ -93,6 +104,7 @@ const MinerUModal = ({
             <RAGFlowFormItem
               name="mineru_apiserver"
               label={t('setting.mineru.apiserver')}
+              required
             >
               <Input placeholder="http://host.docker.internal:9987" />
             </RAGFlowFormItem>
@@ -109,18 +121,25 @@ const MinerUModal = ({
               {(field) => (
                 <RAGFlowSelect
                   value={field.value}
-                  onChange={field.onChange}
+                  onChange={(value) => {
+                    field.onChange(value);
+                    if (value !== 'vlm-http-client') {
+                      form.setValue('mineru_server_url', undefined);
+                    }
+                  }}
                   options={backendOptions}
                   placeholder={t('setting.mineru.selectBackend')}
                 />
               )}
             </RAGFlowFormItem>
+            {backend === 'vlm-http-client' && (
               <RAGFlowFormItem
                 name="mineru_server_url"
                 label={t('setting.mineru.serverUrl')}
               >
                 <Input placeholder="http://your-vllm-server:30000" />
               </RAGFlowFormItem>
+            )}
             <RAGFlowFormItem
               name="mineru_delete_output"
               label={t('setting.mineru.deleteOutput')}