修复错误代码

maubot_llmplus/aibot.py

@@ -127,7 +127,6 @@ class AiBotPlugin(AbsExtraConfigPlugin):
             platform = self.get_ai_platform()
 
             if platform.is_streaming_enabled():
-                await self.client.set_typing(event.room_id, timeout=0)
                 await self._handle_streaming(event, platform)
                 return
 
@@ -154,7 +153,7 @@ class AiBotPlugin(AbsExtraConfigPlugin):
         return None
 
     async def _handle_streaming(self, evt: MessageEvent, platform) -> None:
-        # 发送初始占位消息
+        # 发送初始占位消息；on_message 已设 typing=on，等收到第一个 chunk 再关掉
         placeholder = TextMessageEventContent(
             msgtype=MessageType.TEXT, body="▌", format=Format.HTML, formatted_body="▌"
         )
@@ -163,17 +162,17 @@ class AiBotPlugin(AbsExtraConfigPlugin):
 
         accumulated = ""
         last_edit_len = 0
-        EDIT_THRESHOLD = 100  # 每积累100个字符更新一次消息
+        first_chunk = True
+        EDIT_THRESHOLD = 100
 
         async def send_edit(content: TextMessageEventContent) -> None:
-            """顺序发送编辑消息：用shield确保send_message不被cancel，保护mautrix内部锁"""
+            # shield 防止 wait_for 超时时 cancel send_task，保护 mautrix 内部锁不残留
             send_task = asyncio.ensure_future(self.client.send_message(evt.room_id, content))
             try:
-                # shield防止wait_for超时cancel send_task本身，避免mautrix锁残留
                 await asyncio.wait_for(asyncio.shield(send_task), timeout=8.0)
             except asyncio.TimeoutError:
-                self.log.debug("Streaming: edit wait_for timed out, awaiting task completion")
+                self.log.debug("Streaming: edit timed out, waiting naturally")
-                await send_task  # send_task仍在运行，等待自然完成
+                await send_task
             except Exception as e:
                 self.log.warning(f"Streaming: edit error: {e}")
                 if not send_task.done():
@@ -181,6 +180,10 @@ class AiBotPlugin(AbsExtraConfigPlugin):
 
         try:
             async for chunk in platform.create_chat_completion_stream(self, evt):
+                if first_chunk:
+                    # 收到第一个 chunk 才关掉 typing，等待期间用户可见 typing 指示器
+                    await self.client.set_typing(evt.room_id, timeout=0)
+                    first_chunk = False
                 accumulated += chunk
                 if len(accumulated) - last_edit_len >= EDIT_THRESHOLD:
                     display = accumulated + " ▌"
@@ -198,10 +201,12 @@ class AiBotPlugin(AbsExtraConfigPlugin):
             self.log.exception(f"Streaming error: {e}")
             if not accumulated:
                 accumulated = f"Streaming error: {e}"
+        finally:
+            if first_chunk:
+                await self.client.set_typing(evt.room_id, timeout=0)
 
         self.log.debug(f"Streaming: loop done, total={len(accumulated)}")
 
-        # 输出最终完整内容
         if not accumulated:
             accumulated = "（无响应）"
         final_content = TextMessageEventContent(

maubot_llmplus/thrid_platform.py

@@ -1,5 +1,6 @@
 import asyncio
 import json
+import logging
 from collections import deque
 
 from typing import List
@@ -39,6 +40,7 @@ async def _read_openai_sse(response):
             pass
 
 
+
 class Deepseek(Platform):
 
     def __init__(self, config: BaseProxyConfig, http: ClientSession):
@@ -381,10 +383,11 @@ class Gemini(Platform):
         context = await maubot_llmplus.platforms.get_context(plugin, self, evt)
         request_body, headers = self._build_gemini_request(context)
 
-        endpoint = f"{self.url}/v1beta/models/{self.model}:streamGenerateContent"
+        endpoint = f"{self.url}/v1beta/models/{self.model}:streamGenerateContent?alt=sse"
         async with self.http.post(endpoint, headers=headers, data=json.dumps(request_body)) as response:
             if response.status != 200:
                 raise ValueError(f"Error: {await response.text()}")
+            # 与 Anthropic 保持一致：内联 while 循环，避免双层异步生成器代理
             while True:
                 try:
                     line_bytes = await asyncio.wait_for(response.content.readline(), timeout=60.0)
@@ -399,12 +402,21 @@ class Gemini(Platform):
                 try:
                     data = json.loads(data_str)
                     candidates = data.get("candidates", [])
-                    if candidates:
+                    if not candidates:
-                        parts = candidates[0].get("content", {}).get("parts", [])
+                        continue
+                    candidate = candidates[0]
+                    # 先 yield 文本，再判断是否结束（对齐 OpenAI [DONE] 逻辑）
+                    parts = candidate.get("content", {}).get("parts", [])
                     for part in parts:
                         text = part.get("text", "")
                         if text:
                             yield text
+                    finish_reason = candidate.get("finishReason")
+                    if finish_reason:
+                        logging.getLogger("instance/aibot").debug(
+                            f"Gemini stream finished: finishReason={finish_reason}"
+                        )
+                        break
                 except json.JSONDecodeError:
                     pass