Feat: Support vLLM #4316 (#5395)

### What problem does this PR solve? Feat: Support vLLM #4316 ### Type of change - [x] New Feature (non-breaking change which adds functionality)
2025-12-23 06:46:40 +08:00 · 2025-02-26 18:33:43 +08:00
parent fefea3a2a5
commit d9dd1171a3
9 changed files with 235 additions and 115 deletions
--- a/web/src/pages/user-setting/constants.tsx
+++ b/web/src/pages/user-setting/constants.tsx
@ -6,6 +6,7 @@ import {
  ProfileIcon,
  TeamIcon,
 } from '@/assets/icon/Icon';
+import { LLMFactory } from '@/constants/llm';
 import { UserSettingRouteKey } from '@/constants/setting';
 import { MonitorOutlined } from '@ant-design/icons';

@ -22,17 +23,18 @@ export const UserSettingIconMap = {
 export * from '@/constants/setting';

 export const LocalLlmFactories = [
-  'Ollama',
-  'Xinference',
-  'LocalAI',
-  'LM-Studio',
-  'OpenAI-API-Compatible',
-  'TogetherAI',
-  'Replicate',
-  'OpenRouter',
-  'HuggingFace',
-  'GPUStack',
-  'ModelScope',
+  LLMFactory.Ollama,
+  LLMFactory.Xinference,
+  LLMFactory.LocalAI,
+  LLMFactory.LMStudio,
+  LLMFactory.OpenAiAPICompatible,
+  LLMFactory.TogetherAI,
+  LLMFactory.Replicate,
+  LLMFactory.OpenRouter,
+  LLMFactory.HuggingFace,
+  LLMFactory.GPUStack,
+  LLMFactory.ModelScope,
+  LLMFactory.VLLM,
 ];

 export enum TenantRole {
--- a/web/src/pages/user-setting/setting-model/api-key-modal/index.tsx
+++ b/web/src/pages/user-setting/setting-model/api-key-modal/index.tsx
@ -1,4 +1,5 @@
 import { IModalManagerChildrenProps } from '@/components/modal-manager';
+import { LLMFactory } from '@/constants/llm';
 import { useTranslate } from '@/hooks/common-hooks';
 import { Form, Input, Modal } from 'antd';
 import { useEffect } from 'react';
@ -18,7 +19,7 @@ type FieldType = {
  group_id?: string;
 };

-const modelsWithBaseUrl = ['OpenAI', 'Azure-OpenAI'];
+const modelsWithBaseUrl = [LLMFactory.OpenAI, LLMFactory.AzureOpenAI];

 const ApiKeyModal = ({
  visible,
--- a/web/src/pages/user-setting/setting-model/index.tsx
+++ b/web/src/pages/user-setting/setting-model/index.tsx
@ -1,6 +1,7 @@
 import { ReactComponent as MoreModelIcon } from '@/assets/svg/more-model.svg';
 import { LlmIcon } from '@/components/svg-icon';
 import { useTheme } from '@/components/theme-provider';
+import { LLMFactory } from '@/constants/llm';
 import { useSetModalState, useTranslate } from '@/hooks/common-hooks';
 import { LlmItem, useSelectLlmList } from '@/hooks/llm-hooks';
 import { CloseCircleOutlined, SettingOutlined } from '@ant-design/icons';
@ -94,14 +95,14 @@ const ModelCard = ({ item, clickApiKey }: IModelCardProps) => {
              <Button onClick={handleApiKeyClick}>
                <Flex align="center" gap={4}>
                  {isLocalLlmFactory(item.name) ||
-                  item.name === 'VolcEngine' ||
-                  item.name === 'Tencent Hunyuan' ||
-                  item.name === 'XunFei Spark' ||
-                  item.name === 'BaiduYiyan' ||
-                  item.name === 'Fish Audio' ||
-                  item.name === 'Tencent Cloud' ||
-                  item.name === 'Google Cloud' ||
-                  item.name === 'Azure OpenAI'
+                  item.name === LLMFactory.VolcEngine ||
+                  item.name === LLMFactory.TencentHunYuan ||
+                  item.name === LLMFactory.XunFeiSpark ||
+                  item.name === LLMFactory.BaiduYiYan ||
+                  item.name === LLMFactory.FishAudio ||
+                  item.name === LLMFactory.TencentCloud ||
+                  item.name === LLMFactory.GoogleCloud ||
+                  item.name === LLMFactory.AzureOpenAI
                    ? t('addTheModel')
                    : 'API-Key'}
                  <SettingOutlined />
@ -248,15 +249,15 @@ const UserSettingModel = () => {

  const ModalMap = useMemo(
    () => ({
-      Bedrock: showBedrockAddingModal,
-      VolcEngine: showVolcAddingModal,
-      'Tencent Hunyuan': showHunyuanAddingModal,
-      'XunFei Spark': showSparkAddingModal,
-      BaiduYiyan: showyiyanAddingModal,
-      'Fish Audio': showFishAudioAddingModal,
-      'Tencent Cloud': showTencentCloudAddingModal,
-      'Google Cloud': showGoogleAddingModal,
-      'Azure-OpenAI': showAzureAddingModal,
+      [LLMFactory.Bedrock]: showBedrockAddingModal,
+      [LLMFactory.VolcEngine]: showVolcAddingModal,
+      [LLMFactory.TencentHunYuan]: showHunyuanAddingModal,
+      [LLMFactory.XunFeiSpark]: showSparkAddingModal,
+      [LLMFactory.BaiduYiYan]: showyiyanAddingModal,
+      [LLMFactory.FishAudio]: showFishAudioAddingModal,
+      [LLMFactory.TencentCloud]: showTencentCloudAddingModal,
+      [LLMFactory.GoogleCloud]: showGoogleAddingModal,
+      [LLMFactory.AzureOpenAI]: showAzureAddingModal,
    }),
    [
      showBedrockAddingModal,
@ -396,63 +397,63 @@ const UserSettingModel = () => {
        hideModal={hideVolcAddingModal}
        onOk={onVolcAddingOk}
        loading={volcAddingLoading}
-        llmFactory={'VolcEngine'}
+        llmFactory={LLMFactory.VolcEngine}
      ></VolcEngineModal>
      <HunyuanModal
        visible={HunyuanAddingVisible}
        hideModal={hideHunyuanAddingModal}
        onOk={onHunyuanAddingOk}
        loading={HunyuanAddingLoading}
-        llmFactory={'Tencent Hunyuan'}
+        llmFactory={LLMFactory.TencentHunYuan}
      ></HunyuanModal>
      <GoogleModal
        visible={GoogleAddingVisible}
        hideModal={hideGoogleAddingModal}
        onOk={onGoogleAddingOk}
        loading={GoogleAddingLoading}
-        llmFactory={'Google Cloud'}
+        llmFactory={LLMFactory.GoogleCloud}
      ></GoogleModal>
      <TencentCloudModal
        visible={TencentCloudAddingVisible}
        hideModal={hideTencentCloudAddingModal}
        onOk={onTencentCloudAddingOk}
        loading={TencentCloudAddingLoading}
-        llmFactory={'Tencent Cloud'}
+        llmFactory={LLMFactory.TencentCloud}
      ></TencentCloudModal>
      <SparkModal
        visible={SparkAddingVisible}
        hideModal={hideSparkAddingModal}
        onOk={onSparkAddingOk}
        loading={SparkAddingLoading}
-        llmFactory={'XunFei Spark'}
+        llmFactory={LLMFactory.XunFeiSpark}
      ></SparkModal>
      <YiyanModal
        visible={yiyanAddingVisible}
        hideModal={hideyiyanAddingModal}
        onOk={onyiyanAddingOk}
        loading={yiyanAddingLoading}
-        llmFactory={'BaiduYiyan'}
+        llmFactory={LLMFactory.BaiduYiYan}
      ></YiyanModal>
      <FishAudioModal
        visible={FishAudioAddingVisible}
        hideModal={hideFishAudioAddingModal}
        onOk={onFishAudioAddingOk}
        loading={FishAudioAddingLoading}
-        llmFactory={'Fish Audio'}
+        llmFactory={LLMFactory.FishAudio}
      ></FishAudioModal>
      <BedrockModal
        visible={bedrockAddingVisible}
        hideModal={hideBedrockAddingModal}
        onOk={onBedrockAddingOk}
        loading={bedrockAddingLoading}
-        llmFactory={'Bedrock'}
+        llmFactory={LLMFactory.Bedrock}
      ></BedrockModal>
      <AzureOpenAIModal
        visible={AzureAddingVisible}
        hideModal={hideAzureAddingModal}
        onOk={onAzureAddingOk}
        loading={AzureAddingLoading}
-        llmFactory={'Azure-OpenAI'}
+        llmFactory={LLMFactory.AzureOpenAI}
      ></AzureOpenAIModal>
    </section>
  );
--- a/web/src/pages/user-setting/setting-model/ollama-modal/index.tsx
+++ b/web/src/pages/user-setting/setting-model/ollama-modal/index.tsx
@ -1,3 +1,4 @@
+import { LLMFactory } from '@/constants/llm';
 import { useTranslate } from '@/hooks/common-hooks';
 import { IModalProps } from '@/interfaces/common';
 import { IAddLlmRequestBody } from '@/interfaces/request/llm';
@ -18,19 +19,23 @@ type FieldType = IAddLlmRequestBody & { vision: boolean };
 const { Option } = Select;

 const llmFactoryToUrlMap = {
-  Ollama:
+  [LLMFactory.Ollama]:
    'https://github.com/infiniflow/ragflow/blob/main/docs/guides/deploy_local_llm.mdx',
-  Xinference: 'https://inference.readthedocs.io/en/latest/user_guide',
-  ModelScope: 'https://www.modelscope.cn/docs/model-service/API-Inference/intro',
-  LocalAI: 'https://localai.io/docs/getting-started/models/',
-  'LM-Studio': 'https://lmstudio.ai/docs/basics',
-  'OpenAI-API-Compatible': 'https://platform.openai.com/docs/models/gpt-4',
-  TogetherAI: 'https://docs.together.ai/docs/deployment-options',
-  Replicate: 'https://replicate.com/docs/topics/deployments',
-  OpenRouter: 'https://openrouter.ai/docs',
-  HuggingFace:
+  [LLMFactory.Xinference]:
+    'https://inference.readthedocs.io/en/latest/user_guide',
+  [LLMFactory.ModelScope]:
+    'https://www.modelscope.cn/docs/model-service/API-Inference/intro',
+  [LLMFactory.LocalAI]: 'https://localai.io/docs/getting-started/models/',
+  [LLMFactory.LMStudio]: 'https://lmstudio.ai/docs/basics',
+  [LLMFactory.OpenAiAPICompatible]:
+    'https://platform.openai.com/docs/models/gpt-4',
+  [LLMFactory.TogetherAI]: 'https://docs.together.ai/docs/deployment-options',
+  [LLMFactory.Replicate]: 'https://replicate.com/docs/topics/deployments',
+  [LLMFactory.OpenRouter]: 'https://openrouter.ai/docs',
+  [LLMFactory.HuggingFace]:
    'https://huggingface.co/docs/text-embeddings-inference/quick_tour',
-  GPUStack: 'https://docs.gpustack.ai/latest/quickstart',
+  [LLMFactory.GPUStack]: 'https://docs.gpustack.ai/latest/quickstart',
+  [LLMFactory.VLLM]: 'https://docs.vllm.ai/en/latest/',
 };
 type LlmFactory = keyof typeof llmFactoryToUrlMap;

@ -66,11 +71,11 @@ const OllamaModal = ({
    llmFactoryToUrlMap[llmFactory as LlmFactory] ||
    'https://github.com/infiniflow/ragflow/blob/main/docs/guides/deploy_local_llm.mdx';
  const optionsMap = {
-    HuggingFace: [
+    [LLMFactory.HuggingFace]: [
      { value: 'embedding', label: 'embedding' },
      { value: 'chat', label: 'chat' },
    ],
-    Xinference: [
+    [LLMFactory.Xinference]: [
      { value: 'chat', label: 'chat' },
      { value: 'embedding', label: 'embedding' },
      { value: 'rerank', label: 'rerank' },
@ -78,10 +83,8 @@ const OllamaModal = ({
      { value: 'speech2text', label: 'sequence2text' },
      { value: 'tts', label: 'tts' },
    ],
-    ModelScope: [
-      { value: 'chat', label: 'chat' },
-    ],
-    GPUStack: [
+    [LLMFactory.ModelScope]: [{ value: 'chat', label: 'chat' }],
+    [LLMFactory.GPUStack]: [
      { value: 'chat', label: 'chat' },
      { value: 'embedding', label: 'embedding' },
      { value: 'rerank', label: 'rerank' },