跳过使用Open ai 接口时，DeepSeek-R1 模型的深度思考内容 (#54) (#95)

* 跳过使用Open ai 接口时，DeepSeek-R1 模型的深度思考内容 (#54)

* 跳过 DeepSeek-R1 模型的深度思考内容

* 跳过 DeepSeek-R1 模型的深度思考内容

* 新增LM Studio本地大模型API接口

* 优化代码，遇到Bad Case安全处理

* update:优化

---------

Co-authored-by: Sinyo <38577585+SinyoWong@users.noreply.github.com>
Co-authored-by: hrz <1710360675@qq.com>

config.yaml

@@ -147,6 +147,12 @@ LLM:
     user_id: 你的user_id
     base_url: "https://api.coze.cn/open_api/v2/chat" # 服务地址
     personal_access_token: 你的coze个人令牌
+  LMStudioLLM:
+    # 定义LLM API类型
+    type: openai
+    model_name: deepseek-r1-distill-llama-8b@q4_k_m # 使用的模型名称，需要预先在社区下载
+    url: http://localhost:1234/v1 # LM Studio服务地址
+    api_key: lm-studio # LM Studio服务的固定API Key
   HomeAssistant:
     # 定义LLM API类型
     type: homeassistant

core/providers/llm/openai/openai.py

@@ -25,12 +25,25 @@ class LLMProvider(LLMProviderBase):
                 messages=dialogue,
                 stream=True
             )
+            
+            is_active = True
             for chunk in responses:
-                # 检查是否存在有效的choice且content不为空
-                if chunk.choices and len(chunk.choices) > 0:
-                    delta = chunk.choices[0].delta
-                    content = getattr(delta, 'content', '')
-                    if content:  # 仅在content非空时生成
+                try:
+                    # 检查是否存在有效的choice且content不为空
+                    delta = chunk.choices[0].delta if getattr(chunk, 'choices', None) else None
+                    content = delta.content if hasattr(delta, 'content') else ''
+                except IndexError:
+                    content = ''
+                if content:
+                    # 处理标签跨多个chunk的情况
+                    if '<think>' in content:
+                        is_active = False
+                        content = content.split('<think>')[0]
+                    if '</think>' in content:
+                        is_active = True
+                        content = content.split('</think>')[-1]
+                    if is_active:
                         yield content
+
         except Exception as e:
             logger.bind(tag=TAG).error(f"Error in response generation: {e}")