fix(MessageOperations): Improve message pause functionality and error handling

- Update pauseMessage method to handle both askId and messageId
- Add loading state reset when pausing messages
- Enhance error handling in providers with abort error detection
- Modify ApiService to handle aborted requests gracefully
- Add comprehensive isAbortError utility function

src/renderer/src/hooks/useMessageOperations.ts

@@ -10,6 +10,7 @@ import {
   selectTopicLoading,
   selectTopicMessages,
   setStreamMessage,
+  setTopicLoading,
   updateMessage,
   updateMessages
 } from '@renderer/store/messages'
@@ -155,14 +156,18 @@ export function useMessageOperations(topic: Topic) {
    * 暂停消息生成
    */
   const pauseMessage = useCallback(
-    async (messageId: string) => {
+    // 存的是用户消息的id，也就是助手消息的askId
+    async (askId: string, messageId: string) => {
       // 1. 调用 abort
-      abortCompletion(messageId)
+      abortCompletion(askId)
-
+      console.log('messageId', messageId)
       // 2. 更新消息状态
       await editMessage(messageId, { status: 'paused' })
 
-      // 3. 清理流式消息
+      // 3.更改loading状态
+      dispatch(setTopicLoading({ topicId: topic.id, loading: false }))
+
+      // 4. 清理流式消息
       clearStreamMessageAction(messageId)
     },
     [editMessage, clearStreamMessageAction]
@@ -173,15 +178,13 @@ export function useMessageOperations(topic: Topic) {
     const streamMessages = store.getState().messages.streamMessagesByTopic[topic.id]
     if (streamMessages) {
       // 获取所有流式消息的 askId
-      const askIds = new Set(
+      const askIds = Object.values(streamMessages)
-        Object.values(streamMessages)
+        .map((msg) => [msg.askId, msg.id])
-          .map((msg) => msg.askId)
+        .filter(([askId, id]) => askId && id)
-          .filter(Boolean)
-      )
 
       // 对每个 askId 执行暂停
-      for (const askId of askIds) {
+      for (const [askId, id] of askIds) {
-        await pauseMessage(askId)
+        await pauseMessage(askId, id)
       }
     }
   }, [topic.id, pauseMessage])

src/renderer/src/providers/AnthropicProvider.ts

@@ -208,7 +208,7 @@ export default class AnthropicProvider extends BaseProvider {
     const { signal } = abortController
     const toolResponses: MCPToolResponse[] = []
 
-    const processStream = async (body: MessageCreateParamsNonStreaming) => {
+    const processStream = (body: MessageCreateParamsNonStreaming) => {
       return new Promise<void>((resolve, reject) => {
         const toolCalls: ToolUseBlock[] = []
         let hasThinkingContent = false
@@ -326,7 +326,12 @@ export default class AnthropicProvider extends BaseProvider {
       })
     }
 
-    await processStream(body).finally(cleanup)
+    await processStream(body)
+      .catch((error) => {
+        // 不加这个错误抛不出来
+        throw error
+      })
+      .finally(cleanup)
   }
 
   public async translate(message: Message, assistant: Assistant, onResponse?: (text: string) => void) {

src/renderer/src/providers/BaseProvider.ts

@@ -160,13 +160,20 @@ export default abstract class BaseProvider {
       addAbortController(messageId, () => abortController.abort())
     }
 
+    const cleanup = () => {
+      if (messageId) {
+        removeAbortController(messageId)
+      }
+    }
+
+    abortController.signal.addEventListener('abort', () => {
+      // 兼容
+      cleanup()
+    })
+
     return {
       abortController,
-      cleanup: () => {
+      cleanup
-        if (messageId) {
-          removeAbortController(messageId)
-        }
-      }
     }
   }
 }

src/renderer/src/providers/GeminiProvider.ts

@@ -197,9 +197,10 @@ export default class GeminiProvider extends BaseProvider {
     const messageContents = await this.getMessageContents(userLastMessage!)
 
     const start_time_millsec = new Date().getTime()
-
+    const { abortController, cleanup } = this.createAbortController(userLastMessage?.id)
+    const { signal } = abortController
     if (!streamOutput) {
-      const { response } = await chat.sendMessage(messageContents.parts)
+      const { response } = await chat.sendMessage(messageContents.parts, { signal })
       const time_completion_millsec = new Date().getTime() - start_time_millsec
       onChunk({
         text: response.candidates?.[0].content.parts[0].text,
@@ -218,13 +219,8 @@ export default class GeminiProvider extends BaseProvider {
       return
     }
 
-    const lastUserMessage = userMessages.findLast((m) => m.role === 'user')
-    const { abortController, cleanup } = this.createAbortController(lastUserMessage?.id)
-    const { signal } = abortController
-
     const userMessagesStream = await chat.sendMessageStream(messageContents.parts, { signal })
     let time_first_token_millsec = 0
-
     const processStream = async (stream: GenerateContentStreamResult) => {
       for await (const chunk of stream.stream) {
         if (window.keyv.get(EVENT_NAMES.CHAT_COMPLETION_PAUSED)) break
@@ -297,7 +293,6 @@ export default class GeminiProvider extends BaseProvider {
         })
       }
     }
-
     await processStream(userMessagesStream).finally(cleanup)
   }
 

src/renderer/src/services/ApiService.ts

@@ -3,7 +3,7 @@ import i18n from '@renderer/i18n'
 import store from '@renderer/store'
 import { setGenerating } from '@renderer/store/runtime'
 import { Assistant, Message, Model, Provider, Suggestion } from '@renderer/types'
-import { formatMessageError } from '@renderer/utils/error'
+import { formatMessageError, isAbortError } from '@renderer/utils/error'
 import { cloneDeep, findLast, isEmpty } from 'lodash'
 
 import AiProvider from '../providers/AiProvider'
@@ -116,12 +116,18 @@ export async function fetchChatCompletion({
       // Set metrics.completion_tokens
       if (message.metrics && message?.usage?.completion_tokens) {
         if (!message.metrics?.completion_tokens) {
-          message.metrics.completion_tokens = message.usage.completion_tokens
+          message = {
+            ...message,
+            metrics: {
+              ...message.metrics,
+              completion_tokens: message.usage.completion_tokens
+            }
+          }
         }
       }
     }
   } catch (error: any) {
-    console.log('error', error)
+    if (isAbortError(error)) return
     message.status = 'error'
     message.error = formatMessageError(error)
   }

src/renderer/src/store/messages.ts

@@ -345,7 +345,6 @@ export const sendMessage =
               onResponse: async (msg) => {
                 // 允许在回调外维护一个最新的消息状态，每次都更新这个对象，但只通过节流函数分发到Redux
                 const updateMessage = { ...msg, status: msg.status || 'pending', content: msg.content || '' }
-                // 创建节流函数，限制Redux更新频率
                 // 使用节流函数更新Redux
                 throttledDispatch(
                   assistant,

src/renderer/src/utils/error.ts

@@ -62,3 +62,30 @@ export function formatMessageError(error: any): Record<string, any> {
 export function getErrorMessage(error: any): string {
   return error?.message || error?.toString() || ''
 }
+
+export const isAbortError = (error: any): boolean => {
+  // 检查错误消息
+  if (error?.message === 'Request was aborted.') {
+    return true
+  }
+
+  // 检查是否为 DOMException 类型的中止错误
+  if (error instanceof DOMException && error.name === 'AbortError') {
+    return true
+  }
+  console.log(
+    typeof error === 'object',
+    error.message === 'Request was aborted.' || error?.message?.includes('signal is aborted without reason')
+  )
+  // 检查 OpenAI 特定的错误结构
+  if (
+    (error &&
+      typeof error === 'object' &&
+      (error.message === 'Request was aborted.' || error?.message?.includes('signal is aborted without reason'))) ||
+    error.stack?.includes('OpenAI.makeRequest')
+  ) {
+    return true
+  }
+
+  return false
+}