@lobehub/chat 1.14.2 → 1.14.4

package/src/server/routers/async/file.ts

@@ -15,7 +15,12 @@ import { ModelProvider } from '@/libs/agent-runtime';
 import { asyncAuthedProcedure, asyncRouter as router } from '@/libs/trpc/async';
 import { S3 } from '@/server/modules/S3';
 import { ChunkService } from '@/server/services/chunk';
-import { AsyncTaskError, AsyncTaskErrorType, AsyncTaskStatus } from '@/types/asyncTask';
+import {
+  AsyncTaskError,
+  AsyncTaskErrorType,
+  AsyncTaskStatus,
+  IAsyncTaskError,
+} from '@/types/asyncTask';
 import { safeParseJSON } from '@/utils/safeParseJSON';
 
 const fileProcedure = asyncAuthedProcedure.use(async (opts) => {
@@ -55,10 +60,12 @@ export const fileRouter = router({
       try {
         const timeoutPromise = new Promise((_, reject) => {
           setTimeout(() => {
-            reject({
-              body: { detail: 'embedding task is timeout, please try again' },
-              name: AsyncTaskErrorType.Timeout,
-            } as AsyncTaskError);
+            reject(
+              new AsyncTaskError(
+                AsyncTaskErrorType.Timeout,
+                'embedding task is timeout, please try again',
+              ),
+            );
           }, ASYNC_TASK_TIMEOUT);
         });
 
@@ -76,40 +83,47 @@ export const fileRouter = router({
           const chunks = await ctx.chunkModel.getChunksTextByFileId(input.fileId);
           const requestArray = chunk(chunks, CHUNK_SIZE);
 
-          await pMap(
-            requestArray,
-            async (chunks, index) => {
-              const agentRuntime = await initAgentRuntimeWithUserPayload(
-                ModelProvider.OpenAI,
-                ctx.jwtPayload,
-              );
-
-              const number = index + 1;
-              console.log(`执行第 ${number} 个任务`);
-
-              console.time(`任务[${number}]: embeddings`);
-
-              const embeddings = await agentRuntime.embeddings({
-                dimensions: 1024,
-                input: chunks.map((c) => c.text),
-                model: input.model,
-              });
-              console.timeEnd(`任务[${number}]: embeddings`);
-
-              const items: NewEmbeddingsItem[] =
-                embeddings?.map((e) => ({
-                  chunkId: chunks[e.index].id,
-                  embeddings: e.embedding,
-                  fileId: input.fileId,
-                  model: input.model,
-                })) || [];
+          try {
+            await pMap(
+              requestArray,
+              async (chunks, index) => {
+                const agentRuntime = await initAgentRuntimeWithUserPayload(
+                  ModelProvider.OpenAI,
+                  ctx.jwtPayload,
+                );
 
-              console.time(`任务[${number}]: insert db`);
-              await ctx.embeddingModel.bulkCreate(items);
-              console.timeEnd(`任务[${number}]: insert db`);
-            },
-            { concurrency: CONCURRENCY },
-          );
+                const number = index + 1;
+                console.log(`执行第 ${number} 个任务`);
+
+                console.time(`任务[${number}]: embeddings`);
+
+                const embeddings = await agentRuntime.embeddings({
+                  dimensions: 1024,
+                  input: chunks.map((c) => c.text),
+                  model: input.model,
+                });
+                console.timeEnd(`任务[${number}]: embeddings`);
+
+                const items: NewEmbeddingsItem[] =
+                  embeddings?.map((e) => ({
+                    chunkId: chunks[e.index].id,
+                    embeddings: e.embedding,
+                    fileId: input.fileId,
+                    model: input.model,
+                  })) || [];
+
+                console.time(`任务[${number}]: insert db`);
+                await ctx.embeddingModel.bulkCreate(items);
+                console.timeEnd(`任务[${number}]: insert db`);
+              },
+              { concurrency: CONCURRENCY },
+            );
+          } catch (e) {
+            throw {
+              message: JSON.stringify(e),
+              name: AsyncTaskErrorType.EmbeddingError,
+            };
+          }
 
           const duration = Date.now() - startAt;
           // update the task status to success
@@ -125,8 +139,9 @@ export const fileRouter = router({
         return await Promise.race([embeddingPromise(), timeoutPromise]);
       } catch (e) {
         console.error('embeddingChunks error', e);
+
         await ctx.asyncTaskModel.update(input.taskId, {
-          error: e,
+          error: new AsyncTaskError((e as Error).name, (e as Error).message),
           status: AsyncTaskStatus.Error,
         });
 
@@ -175,10 +190,12 @@ export const fileRouter = router({
 
         const timeoutPromise = new Promise((_, reject) => {
           setTimeout(() => {
-            reject({
-              body: { detail: 'chunking task is timeout, please try again' },
-              name: AsyncTaskErrorType.Timeout,
-            } as AsyncTaskError);
+            reject(
+              new AsyncTaskError(
+                AsyncTaskErrorType.Timeout,
+                'chunking task is timeout, please try again',
+              ),
+            );
           }, ASYNC_TASK_TIMEOUT);
         });
 
@@ -201,6 +218,15 @@ export const fileRouter = router({
 
           const duration = Date.now() - startAt;
 
+          // if no chunk found, throw error
+          if (chunks.length === 0) {
+            throw {
+              message:
+                'No chunk found in this file. it may due to current chunking method can not parse file accurately',
+              name: AsyncTaskErrorType.NoChunkError,
+            };
+          }
+
           await ctx.chunkModel.bulkCreate(chunks, input.fileId);
 
           if (chunkResult.unstructuredChunks) {
@@ -228,9 +254,9 @@ export const fileRouter = router({
       } catch (e) {
         const error = e as any;
 
-        const asyncTaskError: AsyncTaskError = error.body
-          ? { body: safeParseJSON(error.body) ?? error.body, name: error.name }
-          : { body: { detail: error.message }, name: (error as Error).name };
+        const asyncTaskError = error.body
+          ? ({ body: safeParseJSON(error.body) ?? error.body, name: error.name } as IAsyncTaskError)
+          : new AsyncTaskError((error as Error).name, error.message);
 
         console.error('[Chunking Error]', asyncTaskError);
         await ctx.asyncTaskModel.update(input.taskId, {

package/src/server/services/chunk/index.ts

@@ -3,7 +3,12 @@ import { AsyncTaskModel } from '@/database/server/models/asyncTask';
 import { FileModel } from '@/database/server/models/file';
 import { ChunkContentParams, ContentChunk } from '@/server/modules/ContentChunk';
 import { createAsyncServerClient } from '@/server/routers/async';
-import { AsyncTaskStatus, AsyncTaskType } from '@/types/asyncTask';
+import {
+  AsyncTaskError,
+  AsyncTaskErrorType,
+  AsyncTaskStatus,
+  AsyncTaskType,
+} from '@/types/asyncTask';
 
 export class ChunkService {
   private userId: string;
@@ -40,7 +45,19 @@ export class ChunkService {
     const asyncCaller = await createAsyncServerClient(this.userId, payload);
 
     // trigger embedding task asynchronously
-    await asyncCaller.file.embeddingChunks.mutate({ fileId, taskId: asyncTaskId });
+    try {
+      await asyncCaller.file.embeddingChunks.mutate({ fileId, taskId: asyncTaskId });
+    } catch (e) {
+      console.error('[embeddingFileChunks] error:', e);
+
+      await this.asyncTaskModel.update(asyncTaskId, {
+        error: new AsyncTaskError(
+          AsyncTaskErrorType.TaskTriggerError,
+          'trigger chunk embedding async task error. Please check your app is public available or check your proxy settings is set correctly.',
+        ),
+        status: AsyncTaskStatus.Error,
+      });
+    }
 
     return asyncTaskId;
   }
@@ -67,7 +84,19 @@ export class ChunkService {
     const asyncCaller = await createAsyncServerClient(this.userId, payload);
 
     // trigger parse file task asynchronously
-    asyncCaller.file.parseFileToChunks.mutate({ fileId: fileId, taskId: asyncTaskId });
+    asyncCaller.file.parseFileToChunks
+      .mutate({ fileId: fileId, taskId: asyncTaskId })
+      .catch(async (e) => {
+        console.error('[ParseFileToChunks] error:', e);
+
+        await this.asyncTaskModel.update(asyncTaskId, {
+          error: new AsyncTaskError(
+            AsyncTaskErrorType.TaskTriggerError,
+            'trigger file parse async task error. Please check your app is public available or check your proxy settings is set correctly.',
+          ),
+          status: AsyncTaskStatus.Error,
+        });
+      });
 
     return asyncTaskId;
   }

package/src/types/asyncTask.ts

@@ -11,21 +11,40 @@ export enum AsyncTaskStatus {
 }
 
 export enum AsyncTaskErrorType {
-  SDKError = 'SDKError',
+  EmbeddingError = 'EmbeddingError',
+  /**
+   * the chunk parse result it empty
+   */
+  NoChunkError = 'NoChunkError',
   ServerError = 'ServerError',
+  /**
+   * this happens when the task is not trigger successfully
+   */
+  TaskTriggerError = 'TaskTriggerError',
   Timeout = 'TaskTimeout',
 }
 
-export interface AsyncTaskError {
+export interface IAsyncTaskError {
   body: string | { detail: string };
   name: string;
 }
 
+export class AsyncTaskError implements IAsyncTaskError {
+  constructor(name: string, message: string) {
+    this.name = name;
+    this.body = { detail: message };
+  }
+
+  name: string;
+
+  body: { detail: string };
+}
+
 export interface FileParsingTask {
   chunkCount?: number | null;
-  chunkingError?: AsyncTaskError | null;
+  chunkingError?: IAsyncTaskError | null;
   chunkingStatus?: AsyncTaskStatus | null;
-  embeddingError?: AsyncTaskError | null;
+  embeddingError?: IAsyncTaskError | null;
   embeddingStatus?: AsyncTaskStatus | null;
   finishEmbedding?: boolean;
 }

package/src/components/ModelIcon/index.tsx

@@ -1,114 +0,0 @@
-import {
-  Adobe,
-  Ai21,
-  Ai360,
-  AiMass,
-  Aws,
-  Aya,
-  Azure,
-  Baichuan,
-  ByteDance,
-  ChatGLM,
-  Claude,
-  CodeGeeX,
-  Cohere,
-  Dbrx,
-  DeepSeek,
-  FishAudio,
-  Gemini,
-  Gemma,
-  Hunyuan,
-  LLaVA,
-  Meta,
-  Minimax,
-  Mistral,
-  Moonshot,
-  OpenAI,
-  OpenChat,
-  OpenRouter,
-  Perplexity,
-  Rwkv,
-  Spark,
-  Stability,
-  Stepfun,
-  Tongyi,
-  Wenxin,
-  Yi,
-} from '@lobehub/icons';
-import { memo } from 'react';
-
-interface ModelProviderIconProps {
-  model?: string;
-  size?: number;
-}
-
-const ModelIcon = memo<ModelProviderIconProps>(({ model: originModel, size = 12 }) => {
-  if (!originModel) return;
-
-  // lower case the origin model so to better match more model id case
-  const model = originModel.toLowerCase();
-
-  // currently supported models, maybe not in its own provider
-  if (model.includes('text-embedding-')) return <OpenAI.Avatar size={size} />;
-  if (model.includes('gpt-3')) return <OpenAI.Avatar size={size} type={'gpt3'} />;
-  if (model.includes('gpt-4')) return <OpenAI.Avatar size={size} type={'gpt4'} />;
-  if (model.includes('glm-') || model.includes('chatglm')) return <ChatGLM.Avatar size={size} />;
-  if (model.startsWith('codegeex')) return <CodeGeeX.Avatar size={size} />;
-  if (model.includes('deepseek')) return <DeepSeek.Avatar size={size} />;
-  if (model.includes('claude')) return <Claude.Avatar size={size} />;
-  if (model.includes('titan')) return <Aws.Avatar size={size} />;
-  if (model.includes('llama')) return <Meta.Avatar size={size} />;
-  if (model.includes('llava')) return <LLaVA.Avatar size={size} />;
-  if (model.includes('gemini')) return <Gemini.Avatar size={size} />;
-  if (model.includes('gemma')) return <Gemma.Avatar size={size} />;
-  if (model.includes('moonshot')) return <Moonshot.Avatar size={size} />;
-  if (model.includes('qwen')) return <Tongyi.Avatar background={Tongyi.colorPrimary} size={size} />;
-  if (model.includes('minmax') || model.includes('abab')) return <Minimax.Avatar size={size} />;
-  if (
-    model.includes('mistral') ||
-    model.includes('mixtral') ||
-    model.includes('codestral') ||
-    model.includes('mathstral')
-  )
-    return <Mistral.Avatar size={size} />;
-  if (model.includes('pplx') || model.includes('sonar')) return <Perplexity.Avatar size={size} />;
-  if (model.includes('yi-')) return <Yi.Avatar size={size} />;
-  if (model.startsWith('openrouter')) return <OpenRouter.Avatar size={size} />; // only for Cinematika and Auto
-  if (model.startsWith('openchat')) return <OpenChat.Avatar size={size} />;
-  if (model.includes('aya')) return <Aya.Avatar size={size} />;
-  if (model.includes('command')) return <Cohere.Avatar size={size} />;
-  if (model.includes('dbrx')) return <Dbrx.Avatar size={size} />;
-  if (model.includes('step')) return <Stepfun.Avatar size={size} />;
-  if (model.includes('taichu')) return <AiMass.Avatar size={size} />;
-  if (model.includes('360gpt')) return <Ai360.Avatar size={size} />;
-
-  // below: To be supported in providers, move up if supported
-  if (model.includes('baichuan'))
-    return <Baichuan.Avatar background={Baichuan.colorPrimary} size={size} />;
-  if (model.includes('rwkv')) return <Rwkv.Avatar size={size} />;
-  if (model.includes('ernie')) return <Wenxin.Avatar size={size} />;
-  if (model.includes('spark')) return <Spark.Avatar size={size} />;
-  if (model.includes('hunyuan')) return <Hunyuan.Avatar size={size} />;
-  // ref https://github.com/fishaudio/Bert-VITS2/blob/master/train_ms.py#L702
-  if (model.startsWith('d_') || model.startsWith('g_') || model.startsWith('wd_'))
-    return <FishAudio.Avatar size={size} />;
-  if (model.includes('skylark')) return <ByteDance.Avatar size={size} />;
-
-  if (
-    model.includes('stable-diffusion') ||
-    model.includes('stable-video') ||
-    model.includes('stable-cascade') ||
-    model.includes('sdxl') ||
-    model.includes('stablelm') ||
-    model.startsWith('stable-') ||
-    model.startsWith('sd3')
-  )
-    return <Stability.Avatar size={size} />;
-
-  if (model.includes('phi3') || model.includes('phi-3') || model.includes('wizardlm'))
-    return <Azure.Avatar size={size} />;
-  if (model.includes('firefly')) return <Adobe.Avatar size={size} />;
-  if (model.includes('jamba') || model.includes('j2-')) return <Ai21.Avatar size={size} />;
-});
-
-export default ModelIcon;

package/src/components/ModelProviderIcon/index.tsx

@@ -1,148 +0,0 @@
-import {
-  Ai360,
-  AiMass,
-  Anthropic,
-  Azure,
-  Baichuan,
-  Bedrock,
-  DeepSeek,
-  Google,
-  Groq,
-  LobeHub,
-  Minimax,
-  Mistral,
-  Moonshot,
-  Novita,
-  Ollama,
-  OpenAI,
-  OpenRouter,
-  Perplexity,
-  SiliconCloud,
-  Stepfun,
-  Together,
-  Tongyi,
-  ZeroOne,
-  Zhipu,
-} from '@lobehub/icons';
-import { memo } from 'react';
-import { Center } from 'react-layout-kit';
-
-import { ModelProvider } from '@/libs/agent-runtime';
-
-interface ModelProviderIconProps {
-  provider?: string;
-}
-
-const ModelProviderIcon = memo<ModelProviderIconProps>(({ provider }) => {
-  switch (provider) {
-    case 'lobehub': {
-      return <LobeHub.Color size={20} />;
-    }
-
-    case ModelProvider.ZhiPu: {
-      return <Zhipu size={20} />;
-    }
-
-    case ModelProvider.Bedrock: {
-      return <Bedrock size={20} />;
-    }
-
-    case ModelProvider.DeepSeek: {
-      return <DeepSeek size={20} />;
-    }
-
-    case ModelProvider.Google: {
-      return (
-        <Center height={20} width={20}>
-          <Google size={14} />
-        </Center>
-      );
-    }
-
-    case ModelProvider.Azure: {
-      return (
-        <Center height={20} width={20}>
-          <Azure size={14} />
-        </Center>
-      );
-    }
-
-    case ModelProvider.Moonshot: {
-      return <Moonshot size={20} />;
-    }
-
-    case ModelProvider.OpenAI: {
-      return <OpenAI size={20} />;
-    }
-
-    case ModelProvider.Ollama: {
-      return <Ollama size={20} />;
-    }
-
-    case ModelProvider.Perplexity: {
-      return <Perplexity size={20} />;
-    }
-
-    case ModelProvider.Minimax: {
-      return <Minimax size={20} />;
-    }
-
-    case ModelProvider.Mistral: {
-      return <Mistral size={20} />;
-    }
-
-    case ModelProvider.Anthropic: {
-      return <Anthropic size={20} />;
-    }
-
-    case ModelProvider.Groq: {
-      return <Groq size={20} />;
-    }
-
-    case ModelProvider.OpenRouter: {
-      return <OpenRouter size={20} />;
-    }
-
-    case ModelProvider.ZeroOne: {
-      return <ZeroOne size={20} />;
-    }
-
-    case ModelProvider.TogetherAI: {
-      return <Together size={20} />;
-    }
-
-    case ModelProvider.Qwen: {
-      return <Tongyi size={20} />;
-    }
-
-    case ModelProvider.Stepfun: {
-      return <Stepfun size={20} />;
-    }
-
-    case ModelProvider.Novita: {
-      return <Novita size={20} />;
-    }
-
-    case ModelProvider.Baichuan: {
-      return <Baichuan size={20} />;
-    }
-
-    case ModelProvider.Taichu: {
-      return <AiMass size={20} />;
-    }
-
-    case ModelProvider.Ai360: {
-      return <Ai360 size={20} />;
-    }
-
-    case ModelProvider.SiliconCloud: {
-      return <SiliconCloud size={20} />;
-    }
-
-    default: {
-      return null;
-    }
-  }
-});
-
-export default ModelProviderIcon;

package/src/components/ModelTag/ModelIcon.tsx

@@ -1,105 +0,0 @@
-import {
-  AdobeFirefly,
-  Ai21,
-  Ai360,
-  AiMass,
-  Aws,
-  Aya,
-  Azure,
-  Baichuan,
-  ByteDance,
-  ChatGLM,
-  Claude,
-  CodeGeeX,
-  Cohere,
-  Dbrx,
-  DeepSeek,
-  FishAudio,
-  Gemini,
-  Gemma,
-  Hunyuan,
-  LLaVA,
-  Meta,
-  Minimax,
-  Mistral,
-  Moonshot,
-  OpenAI,
-  OpenChat,
-  OpenRouter,
-  Perplexity,
-  Rwkv,
-  Spark,
-  Stability,
-  Stepfun,
-  Tongyi,
-  Wenxin,
-  Yi,
-} from '@lobehub/icons';
-import { memo } from 'react';
-
-interface ModelIconProps {
-  model?: string;
-  size?: number;
-}
-
-const ModelIcon = memo<ModelIconProps>(({ model: originModel, size = 12 }) => {
-  if (!originModel) return;
-
-  // lower case the origin model so to better match more model id case
-  const model = originModel.toLowerCase();
-  
-  // currently supported models, maybe not in its own provider
-  if (model.startsWith('gpt')) return <OpenAI size={size} />;
-  if (model.startsWith('glm') || model.includes('chatglm')) return <ChatGLM size={size} />;
-  if (model.includes('codegeex')) return <CodeGeeX size={size} />;
-  if (model.includes('claude')) return <Claude size={size} />;
-  if (model.includes('deepseek')) return <DeepSeek size={size} />;
-  if (model.includes('titan')) return <Aws size={size} />;
-  if (model.includes('llama')) return <Meta size={size} />;
-  if (model.includes('llava')) return <LLaVA size={size} />;
-  if (model.includes('gemini')) return <Gemini size={size} />;
-  if (model.includes('gemma')) return <Gemma.Simple size={size} />;
-  if (model.includes('moonshot')) return <Moonshot size={size} />;
-  if (model.includes('qwen')) return <Tongyi size={size} />;
-  if (model.includes('minmax')) return <Minimax size={size} />;
-  if (model.includes('abab')) return <Minimax size={size} />;
-  if (model.includes('mistral') || model.includes('mixtral') || model.includes('codestral') || model.includes('mathstral')) return <Mistral size={size} />;
-  if (model.includes('pplx') || model.includes('sonar')) return <Perplexity size={size} />;
-  if (model.includes('yi-')) return <Yi size={size} />;
-  if (model.startsWith('openrouter')) return <OpenRouter size={size} />; // only for Cinematika and Auto
-  if (model.startsWith('openchat')) return <OpenChat size={size} />;
-  if (model.includes('aya')) return <Aya.Avatar size={size} />;
-  if (model.includes('command')) return <Cohere size={size} />;
-  if (model.includes('dbrx')) return <Dbrx size={size} />;
-  if (model.includes('step')) return <Stepfun size={size} />;
-  if (model.includes('taichu')) return <AiMass size={size} />;
-  if (model.includes('360gpt')) return <Ai360 size={size} />;
-
-  // below: To be supported in providers, move up if supported
-  if (model.includes('baichuan')) return <Baichuan size={size} />;
-  if (model.includes('rwkv')) return <Rwkv size={size} />;
-  if (model.includes('ernie')) return <Wenxin size={size} />;
-  if (model.includes('spark')) return <Spark size={size} />;
-  if (model.includes('hunyuan')) return <Hunyuan size={size} />;
-  // ref https://github.com/fishaudio/Bert-VITS2/blob/master/train_ms.py#L702
-  if (model.startsWith('d_') || model.startsWith('g_') || model.startsWith('wd_'))
-    return <FishAudio size={size} />;
-  if (model.includes('skylark')) return <ByteDance size={size} />;
-
-  if (
-    model.includes('stable-diffusion') ||
-    model.includes('stable-video') ||
-    model.includes('stable-cascade') ||
-    model.includes('sdxl') ||
-    model.includes('stablelm') ||
-    model.startsWith('stable-') ||
-    model.startsWith('sd3')
-  )
-    return <Stability size={size} />;
-
-  if (model.includes('phi3') || model.includes('phi-3') || model.includes('wizardlm')) return <Azure size={size} />;
-  if (model.includes('firefly')) return <AdobeFirefly size={size} />;
-  if (model.includes('jamba') || model.includes('j2-')) return <Ai21 size={size} />;
-});
-
-export default ModelIcon;

package/src/components/ModelTag/index.tsx

@@ -1,13 +0,0 @@
-import { Tag } from '@lobehub/ui';
-import { memo } from 'react';
-
-import ModelIcon from './ModelIcon';
-
-interface ModelTagProps {
-  model: string;
-}
-const ModelTag = memo<ModelTagProps>(({ model }) => (
-  <Tag icon={<ModelIcon model={model} />}>{model}</Tag>
-));
-
-export default ModelTag;

package/src/features/AgentSetting/AgentTTS/options.ts

@@ -1,16 +0,0 @@
-import { SelectProps } from 'antd';
-
-export const ttsOptions: SelectProps['options'] = [
-  {
-    label: 'OpenAI',
-    value: 'openai',
-  },
-  {
-    label: 'Edge Speech',
-    value: 'edge',
-  },
-  {
-    label: 'Microsoft Speech',
-    value: 'microsoft',
-  },
-];