@lobehub/chat 1.57.1 → 1.58.0

package/src/app/[variants]/(main)/settings/provider/features/ModelList/ModelConfigModal/index.tsx

@@ -1,13 +1,13 @@
 import { Modal } from '@lobehub/ui';
 import { Button, FormInstance } from 'antd';
 import isEqual from 'fast-deep-equal';
-import { memo, useState } from 'react';
+import { memo, use, useState } from 'react';
 import { useTranslation } from 'react-i18next';
 
-import { ModelProvider } from '@/libs/agent-runtime';
 import { aiModelSelectors, useAiInfraStore } from '@/store/aiInfra';
 
 import ModelConfigForm from '../CreateNewModelModal/Form';
+import { ProviderSettingsContext } from '../ProviderSettingsContext';
 
 interface ModelConfigModalProps {
   id: string;
@@ -28,6 +28,7 @@ const ModelConfigModal = memo<ModelConfigModalProps>(({ id, open, setOpen }) =>
   const closeModal = () => {
     setOpen(false);
   };
+  const { showDeployName } = use(ProviderSettingsContext);
 
   return (
     <Modal
@@ -66,7 +67,7 @@ const ModelConfigModal = memo<ModelConfigModalProps>(({ id, open, setOpen }) =>
         idEditable={false}
         initialValues={model}
         onFormInstanceReady={setFormInstance}
-        showAzureDeployName={editingProvider === ModelProvider.Azure || editingProvider === ModelProvider.Doubao}
+        showDeployName={showDeployName}
         type={model?.type}
       />
     </Modal>

package/src/app/[variants]/(main)/settings/provider/features/ModelList/ProviderSettingsContext.ts

@@ -2,7 +2,9 @@ import { createContext } from 'react';
 
 export interface ProviderSettingsContextValue {
   modelEditable?: boolean;
+  sdkType?: string;
   showAddNewModel?: boolean;
+  showDeployName?: boolean;
   showModelFetcher?: boolean;
 }
 

package/src/app/[variants]/(main)/settings/provider/features/ModelList/index.tsx

@@ -3,7 +3,6 @@
 import { Suspense, memo } from 'react';
 import { Flexbox } from 'react-layout-kit';
 
-import { ProviderSettingsContext } from '@/app/[variants]/(main)/settings/provider/features/ModelList/ProviderSettingsContext';
 import { useIsMobile } from '@/hooks/useIsMobile';
 import { aiModelSelectors, useAiInfraStore } from '@/store/aiInfra';
 
@@ -11,6 +10,7 @@ import DisabledModels from './DisabledModels';
 import EmptyModels from './EmptyModels';
 import EnabledModelList from './EnabledModelList';
 import ModelTitle from './ModelTitle';
+import { ProviderSettingsContext, ProviderSettingsContextValue } from './ProviderSettingsContext';
 import SearchResult from './SearchResult';
 import SkeletonList from './SkeletonList';
 
@@ -41,19 +41,18 @@ const Content = memo<ContentProps>(({ id }) => {
   );
 });
 
-interface ModelListProps {
+interface ModelListProps extends ProviderSettingsContextValue {
   id: string;
-  modelEditable?: boolean;
-  showAddNewModel?: boolean;
-  showModelFetcher?: boolean;
 }
 
 const ModelList = memo<ModelListProps>(
-  ({ id, showModelFetcher, showAddNewModel, modelEditable = true }) => {
+  ({ id, showModelFetcher, sdkType, showAddNewModel, showDeployName, modelEditable = true }) => {
     const mobile = useIsMobile();
 
     return (
-      <ProviderSettingsContext value={{ modelEditable, showAddNewModel, showModelFetcher }}>
+      <ProviderSettingsContext
+        value={{ modelEditable, sdkType, showAddNewModel, showDeployName, showModelFetcher }}
+      >
         <Flexbox gap={16} paddingInline={mobile ? 12 : 0}>
           <ModelTitle
             provider={id}

package/src/app/[variants]/(main)/settings/provider/features/ProviderConfig/index.tsx

@@ -127,7 +127,7 @@ const ProviderConfig = memo<ProviderConfigProps>(
       defaultShowBrowserRequest,
       disableBrowserRequest,
       showChecker = true,
-    } = settings;
+    } = settings || {};
     const { t } = useTranslation('modelProvider');
     const [form] = Form.useForm();
     const { cx, styles, theme } = useStyles();

package/src/config/aiModels/azureai.ts

@@ -0,0 +1,18 @@
+import { AIChatModelCard } from '@/types/aiModel';
+
+const azureChatModels: AIChatModelCard[] = [
+  {
+    abilities: {
+      reasoning: true,
+    },
+    contextWindowTokens: 128_000,
+    displayName: 'DeepSeek R1',
+    id: 'DeepSeek-R1',
+    maxOutput: 4096,
+    type: 'chat',
+  },
+];
+
+export const allModels = [...azureChatModels];
+
+export default allModels;

package/src/config/aiModels/index.ts

@@ -4,6 +4,7 @@ import { default as ai21 } from './ai21';
 import { default as ai360 } from './ai360';
 import { default as anthropic } from './anthropic';
 import { default as azure } from './azure';
+import { default as azureai } from './azureai';
 import { default as baichuan } from './baichuan';
 import { default as bedrock } from './bedrock';
 import { default as cloudflare } from './cloudflare';
@@ -68,6 +69,7 @@ export const LOBE_DEFAULT_MODEL_LIST = buildDefaultModelList({
   ai360,
   anthropic,
   azure,
+  azureai,
   baichuan,
   bedrock,
   cloudflare,
@@ -113,6 +115,7 @@ export { default as ai21 } from './ai21';
 export { default as ai360 } from './ai360';
 export { default as anthropic } from './anthropic';
 export { default as azure } from './azure';
+export { default as azureai } from './azureai';
 export { default as baichuan } from './baichuan';
 export { default as bedrock } from './bedrock';
 export { default as cloudflare } from './cloudflare';

package/src/config/modelProviders/azure.ts

@@ -58,10 +58,11 @@ const Azure: ModelProviderCard = {
     'Azure 提供多种先进的AI模型，包括GPT-3.5和最新的GPT-4系列，支持多种数据类型和复杂任务，致力于安全、可靠和可持续的AI解决方案。',
   id: 'azure',
   modelsUrl: 'https://learn.microsoft.com/azure/ai-services/openai/concepts/models',
-  name: 'Azure',
+  name: 'Azure OpenAI',
   settings: {
     defaultShowBrowserRequest: true,
     sdkType: 'azure',
+    showDeployName: true,
   },
   url: 'https://azure.microsoft.com',
 };

package/src/config/modelProviders/azureai.ts

@@ -0,0 +1,19 @@
+import { ModelProviderCard } from '@/types/llm';
+
+// ref: https://learn.microsoft.com/azure/ai-services/openai/concepts/models
+const Azure: ModelProviderCard = {
+  chatModels: [],
+  description:
+    'Azure 提供多种先进的AI模型，包括GPT-3.5和最新的GPT-4系列，支持多种数据类型和复杂任务，致力于安全、可靠和可持续的AI解决方案。',
+  id: 'azureai',
+  modelsUrl: 'https://ai.azure.com/explore/models',
+  name: 'Azure AI',
+  settings: {
+    defaultShowBrowserRequest: true,
+    sdkType: 'azureai',
+    showDeployName: true,
+  },
+  url: 'https://ai.azure.com',
+};
+
+export default Azure;

package/src/config/modelProviders/index.ts

@@ -4,6 +4,7 @@ import Ai21Provider from './ai21';
 import Ai360Provider from './ai360';
 import AnthropicProvider from './anthropic';
 import AzureProvider from './azure';
+import AzureAIProvider from './azureai';
 import BaichuanProvider from './baichuan';
 import BedrockProvider from './bedrock';
 import CloudflareProvider from './cloudflare';
@@ -92,6 +93,7 @@ export const LOBE_DEFAULT_MODEL_LIST: ChatModelCard[] = [
 export const DEFAULT_MODEL_PROVIDER_LIST = [
   OpenAIProvider,
   { ...AzureProvider, chatModels: [] },
+  AzureAIProvider,
   OllamaProvider,
   VLLMProvider,
   AnthropicProvider,
@@ -148,6 +150,7 @@ export { default as Ai21ProviderCard } from './ai21';
 export { default as Ai360ProviderCard } from './ai360';
 export { default as AnthropicProviderCard } from './anthropic';
 export { default as AzureProviderCard } from './azure';
+export { default as AzureAIProviderCard } from './azureai';
 export { default as BaichuanProviderCard } from './baichuan';
 export { default as BedrockProviderCard } from './bedrock';
 export { default as CloudflareProviderCard } from './cloudflare';

package/src/database/server/models/aiProvider.ts

@@ -1,4 +1,5 @@
 import { and, asc, desc, eq } from 'drizzle-orm/expressions';
+import { isEmpty } from 'lodash-es';
 
 import { LobeChatDatabase } from '@/database/type';
 import { ModelProvider } from '@/libs/agent-runtime';
@@ -207,6 +208,7 @@ export class AiProviderModel {
       ...result,
       fetchOnClient: typeof result.fetchOnClient === 'boolean' ? result.fetchOnClient : undefined,
       keyVaults,
+      settings: isEmpty(result.settings) ? undefined : result.settings,
     } as AiProviderDetailItem;
   };
 

package/src/libs/agent-runtime/AgentRuntime.ts

@@ -7,6 +7,7 @@ import { LobeAi21AI } from './ai21';
 import { LobeAi360AI } from './ai360';
 import { LobeAnthropicAI } from './anthropic';
 import { LobeAzureOpenAI } from './azureOpenai';
+import { LobeAzureAI } from './azureai';
 import { LobeBaichuanAI } from './baichuan';
 import { LobeBedrockAI, LobeBedrockAIParams } from './bedrock';
 import { LobeCloudflareAI, LobeCloudflareParams } from './cloudflare';
@@ -141,6 +142,7 @@ class AgentRuntime {
       ai360: Partial<ClientOptions>;
       anthropic: Partial<ClientOptions>;
       azure: { apiKey?: string; apiVersion?: string; baseURL?: string };
+      azureai: { apiKey?: string; apiVersion?: string; baseURL?: string };
       baichuan: Partial<ClientOptions>;
       bedrock: Partial<LobeBedrockAIParams>;
       cloudflare: Partial<LobeCloudflareParams>;
@@ -201,6 +203,11 @@ class AgentRuntime {
         break;
       }
 
+      case ModelProvider.AzureAI: {
+        runtimeModel = new LobeAzureAI(params.azureai);
+        break;
+      }
+
       case ModelProvider.ZhiPu: {
         runtimeModel = new LobeZhipuAI(params.zhipu);
         break;
@@ -307,7 +314,7 @@ class AgentRuntime {
       }
 
       case ModelProvider.Novita: {
-        runtimeModel = new LobeNovitaAI(params.novita ?? {});
+        runtimeModel = new LobeNovitaAI(params.novita);
         break;
       }
 
@@ -317,7 +324,7 @@ class AgentRuntime {
       }
 
       case ModelProvider.Baichuan: {
-        runtimeModel = new LobeBaichuanAI(params.baichuan ?? {});
+        runtimeModel = new LobeBaichuanAI(params.baichuan);
         break;
       }
 
@@ -327,12 +334,12 @@ class AgentRuntime {
       }
 
       case ModelProvider.Ai360: {
-        runtimeModel = new LobeAi360AI(params.ai360 ?? {});
+        runtimeModel = new LobeAi360AI(params.ai360);
         break;
       }
 
       case ModelProvider.SiliconCloud: {
-        runtimeModel = new LobeSiliconCloudAI(params.siliconcloud ?? {});
+        runtimeModel = new LobeSiliconCloudAI(params.siliconcloud);
         break;
       }
 
@@ -372,12 +379,12 @@ class AgentRuntime {
       }
 
       case ModelProvider.Jina: {
-        runtimeModel = new LobeJinaAI(params.jina ?? {});
+        runtimeModel = new LobeJinaAI(params.jina);
         break;
       }
 
       case ModelProvider.Cloudflare: {
-        runtimeModel = new LobeCloudflareAI(params.cloudflare ?? {});
+        runtimeModel = new LobeCloudflareAI(params.cloudflare);
         break;
       }
 

package/src/libs/agent-runtime/azureai/index.ts

@@ -0,0 +1,109 @@
+import createClient, { ModelClient } from '@azure-rest/ai-inference';
+import { AzureKeyCredential } from '@azure/core-auth';
+import OpenAI from 'openai';
+
+import { LobeRuntimeAI } from '../BaseAI';
+import { AgentRuntimeErrorType } from '../error';
+import { ChatCompetitionOptions, ChatStreamPayload, ModelProvider } from '../types';
+import { AgentRuntimeError } from '../utils/createError';
+import { debugStream } from '../utils/debugStream';
+import { transformResponseToStream } from '../utils/openaiCompatibleFactory';
+import { StreamingResponse } from '../utils/response';
+import { OpenAIStream, createSSEDataExtractor } from '../utils/streams';
+
+export class LobeAzureAI implements LobeRuntimeAI {
+  client: ModelClient;
+
+  constructor(params?: { apiKey?: string; apiVersion?: string; baseURL?: string }) {
+    if (!params?.apiKey || !params?.baseURL)
+      throw AgentRuntimeError.createError(AgentRuntimeErrorType.InvalidProviderAPIKey);
+
+    this.client = createClient(params?.baseURL, new AzureKeyCredential(params?.apiKey));
+
+    this.baseURL = params?.baseURL;
+  }
+
+  baseURL: string;
+
+  async chat(payload: ChatStreamPayload, options?: ChatCompetitionOptions) {
+    const { messages, model, ...params } = payload;
+    // o1 series models on Azure OpenAI does not support streaming currently
+    const enableStreaming = model.includes('o1') ? false : (params.stream ?? true);
+    try {
+      const response = this.client.path('/chat/completions').post({
+        body: {
+          messages: messages as OpenAI.ChatCompletionMessageParam[],
+          model,
+          ...params,
+          stream: enableStreaming,
+          tool_choice: params.tools ? 'auto' : undefined,
+        },
+      });
+
+      if (enableStreaming) {
+        const stream = await response.asBrowserStream();
+
+        const [prod, debug] = stream.body!.tee();
+
+        if (process.env.DEBUG_AZURE_AI_CHAT_COMPLETION === '1') {
+          debugStream(debug).catch(console.error);
+        }
+
+        return StreamingResponse(
+          OpenAIStream(prod.pipeThrough(createSSEDataExtractor()), {
+            callbacks: options?.callback,
+          }),
+          {
+            headers: options?.headers,
+          },
+        );
+      } else {
+        const res = await response;
+
+        // the azure AI inference response is openai compatible
+        const stream = transformResponseToStream(res.body as OpenAI.ChatCompletion);
+        return StreamingResponse(OpenAIStream(stream, { callbacks: options?.callback }), {
+          headers: options?.headers,
+        });
+      }
+    } catch (e) {
+      let error = e as { [key: string]: any; code: string; message: string };
+
+      if (error.code) {
+        switch (error.code) {
+          case 'DeploymentNotFound': {
+            error = { ...error, deployId: model };
+          }
+        }
+      } else {
+        error = {
+          cause: error.cause,
+          message: error.message,
+          name: error.name,
+        } as any;
+      }
+
+      const errorType = error.code
+        ? AgentRuntimeErrorType.ProviderBizError
+        : AgentRuntimeErrorType.AgentRuntimeError;
+
+      throw AgentRuntimeError.chat({
+        endpoint: this.maskSensitiveUrl(this.baseURL),
+        error,
+        errorType,
+        provider: ModelProvider.Azure,
+      });
+    }
+  }
+
+  private maskSensitiveUrl = (url: string) => {
+    // 使用正则表达式匹配 'https://' 后面和 '.azure.com/' 前面的内容
+    const regex = /^(https:\/\/)([^.]+)(\.azure\.com\/.*)$/;
+
+    // 使用替换函数
+    return url.replace(regex, (match, protocol, subdomain, rest) => {
+      // 将子域名替换为 '***'
+      return `${protocol}***${rest}`;
+    });
+  };
+}

package/src/libs/agent-runtime/baichuan/index.test.ts

@@ -1,255 +1,13 @@
 // @vitest-environment node
-import OpenAI from 'openai';
-import { Mock, afterEach, beforeEach, describe, expect, it, vi } from 'vitest';
+import { ModelProvider } from '@/libs/agent-runtime';
 
-import {
-  ChatStreamCallbacks,
-  LobeOpenAICompatibleRuntime,
-  ModelProvider,
-} from '@/libs/agent-runtime';
-
-import * as debugStreamModule from '../utils/debugStream';
+import { testProvider } from '../providerTestUtils';
 import { LobeBaichuanAI } from './index';
 
-const provider = ModelProvider.Baichuan;
-const defaultBaseURL = 'https://api.baichuan-ai.com/v1';
-
-const bizErrorType = 'ProviderBizError';
-const invalidErrorType = 'InvalidProviderAPIKey';
-
-// Mock the console.error to avoid polluting test output
-vi.spyOn(console, 'error').mockImplementation(() => {});
-
-let instance: LobeOpenAICompatibleRuntime;
-
-beforeEach(() => {
-  instance = new LobeBaichuanAI({ apiKey: 'test' });
-
-  // 使用 vi.spyOn 来模拟 chat.completions.create 方法
-  vi.spyOn(instance['client'].chat.completions, 'create').mockResolvedValue(
-    new ReadableStream() as any,
-  );
-});
-
-afterEach(() => {
-  vi.clearAllMocks();
-});
-
-describe('LobeBaichuanAI', () => {
-  describe('init', () => {
-    it('should correctly initialize with an API key', async () => {
-      const instance = new LobeBaichuanAI({ apiKey: 'test_api_key' });
-      expect(instance).toBeInstanceOf(LobeBaichuanAI);
-      expect(instance.baseURL).toEqual(defaultBaseURL);
-    });
-  });
-
-  describe('chat', () => {
-    describe('Error', () => {
-      it('should return OpenAIBizError with an openai error response when OpenAI.APIError is thrown', async () => {
-        // Arrange
-        const apiError = new OpenAI.APIError(
-          400,
-          {
-            status: 400,
-            error: {
-              message: 'Bad Request',
-            },
-          },
-          'Error message',
-          {},
-        );
-
-        vi.spyOn(instance['client'].chat.completions, 'create').mockRejectedValue(apiError);
-
-        // Act
-        try {
-          await instance.chat({
-            messages: [{ content: 'Hello', role: 'user' }],
-            model: 'Baichuan4',
-            temperature: 0,
-          });
-        } catch (e) {
-          expect(e).toEqual({
-            endpoint: defaultBaseURL,
-            error: {
-              error: { message: 'Bad Request' },
-              status: 400,
-            },
-            errorType: bizErrorType,
-            provider,
-          });
-        }
-      });
-
-      it('should throw AgentRuntimeError with NoOpenAIAPIKey if no apiKey is provided', async () => {
-        try {
-          new LobeBaichuanAI({});
-        } catch (e) {
-          expect(e).toEqual({ errorType: invalidErrorType });
-        }
-      });
-
-      it('should return OpenAIBizError with the cause when OpenAI.APIError is thrown with cause', async () => {
-        // Arrange
-        const errorInfo = {
-          stack: 'abc',
-          cause: {
-            message: 'api is undefined',
-          },
-        };
-        const apiError = new OpenAI.APIError(400, errorInfo, 'module error', {});
-
-        vi.spyOn(instance['client'].chat.completions, 'create').mockRejectedValue(apiError);
-
-        // Act
-        try {
-          await instance.chat({
-            messages: [{ content: 'Hello', role: 'user' }],
-            model: 'Baichuan4',
-            temperature: 0,
-          });
-        } catch (e) {
-          expect(e).toEqual({
-            endpoint: defaultBaseURL,
-            error: {
-              cause: { message: 'api is undefined' },
-              stack: 'abc',
-            },
-            errorType: bizErrorType,
-            provider,
-          });
-        }
-      });
-
-      it('should return OpenAIBizError with an cause response with desensitize Url', async () => {
-        // Arrange
-        const errorInfo = {
-          stack: 'abc',
-          cause: { message: 'api is undefined' },
-        };
-        const apiError = new OpenAI.APIError(400, errorInfo, 'module error', {});
-
-        instance = new LobeBaichuanAI({
-          apiKey: 'test',
-
-          baseURL: 'https://api.abc.com/v1',
-        });
-
-        vi.spyOn(instance['client'].chat.completions, 'create').mockRejectedValue(apiError);
-
-        // Act
-        try {
-          await instance.chat({
-            messages: [{ content: 'Hello', role: 'user' }],
-            model: 'Baichuan4',
-            temperature: 0,
-          });
-        } catch (e) {
-          expect(e).toEqual({
-            endpoint: 'https://api.***.com/v1',
-            error: {
-              cause: { message: 'api is undefined' },
-              stack: 'abc',
-            },
-            errorType: bizErrorType,
-            provider,
-          });
-        }
-      });
-
-      it('should throw an InvalidBaichuanAPIKey error type on 401 status code', async () => {
-        // Mock the API call to simulate a 401 error
-        const error = new Error('Unauthorized') as any;
-        error.status = 401;
-        vi.mocked(instance['client'].chat.completions.create).mockRejectedValue(error);
-
-        try {
-          await instance.chat({
-            messages: [{ content: 'Hello', role: 'user' }],
-            model: 'Baichuan4',
-            temperature: 0,
-          });
-        } catch (e) {
-          // Expect the chat method to throw an error with InvalidBaichuanAPIKey
-          expect(e).toEqual({
-            endpoint: defaultBaseURL,
-            error: new Error('Unauthorized'),
-            errorType: invalidErrorType,
-            provider,
-          });
-        }
-      });
-
-      it('should return AgentRuntimeError for non-OpenAI errors', async () => {
-        // Arrange
-        const genericError = new Error('Generic Error');
-
-        vi.spyOn(instance['client'].chat.completions, 'create').mockRejectedValue(genericError);
-
-        // Act
-        try {
-          await instance.chat({
-            messages: [{ content: 'Hello', role: 'user' }],
-            model: 'Baichuan4',
-            temperature: 0,
-          });
-        } catch (e) {
-          expect(e).toEqual({
-            endpoint: defaultBaseURL,
-            errorType: 'AgentRuntimeError',
-            provider,
-            error: {
-              name: genericError.name,
-              cause: genericError.cause,
-              message: genericError.message,
-              stack: genericError.stack,
-            },
-          });
-        }
-      });
-    });
-
-    describe('DEBUG', () => {
-      it('should call debugStream and return StreamingTextResponse when DEBUG_BAICHUAN_CHAT_COMPLETION is 1', async () => {
-        // Arrange
-        const mockProdStream = new ReadableStream() as any; // 模拟的 prod 流
-        const mockDebugStream = new ReadableStream({
-          start(controller) {
-            controller.enqueue('Debug stream content');
-            controller.close();
-          },
-        }) as any;
-        mockDebugStream.toReadableStream = () => mockDebugStream; // 添加 toReadableStream 方法
-
-        // 模拟 chat.completions.create 返回值，包括模拟的 tee 方法
-        (instance['client'].chat.completions.create as Mock).mockResolvedValue({
-          tee: () => [mockProdStream, { toReadableStream: () => mockDebugStream }],
-        });
-
-        // 保存原始环境变量值
-        const originalDebugValue = process.env.DEBUG_BAICHUAN_CHAT_COMPLETION;
-
-        // 模拟环境变量
-        process.env.DEBUG_BAICHUAN_CHAT_COMPLETION = '1';
-        vi.spyOn(debugStreamModule, 'debugStream').mockImplementation(() => Promise.resolve());
-
-        // 执行测试
-        // 运行你的测试函数，确保它会在条件满足时调用 debugStream
-        // 假设的测试函数调用，你可能需要根据实际情况调整
-        await instance.chat({
-          messages: [{ content: 'Hello', role: 'user' }],
-          model: 'Baichuan4',
-          stream: true,
-          temperature: 0,
-        });
-
-        // 验证 debugStream 被调用
-        expect(debugStreamModule.debugStream).toHaveBeenCalled();
-
-        // 恢复原始环境变量值
-        process.env.DEBUG_BAICHUAN_CHAT_COMPLETION = originalDebugValue;
-      });
-    });
-  });
+testProvider({
+  Runtime: LobeBaichuanAI,
+  provider: ModelProvider.Baichuan,
+  defaultBaseURL: 'https://api.baichuan-ai.com/v1',
+  chatDebugEnv: 'DEBUG_BAICHUAN_CHAT_COMPLETION',
+  chatModel: 'hunyuan-lite',
 });