使用Ollama Llama-2时遇AttributeError：无bind_functions属性

报错原因：LangChain的ChatOllama类没有bind_functions方法，该方法是OpenAI集成专属的。要在Ollama模型上实现函数/工具绑定，需使用LangChain为Ollama提供的bind_tools方法，并调整对应的输出解析器。

修改步骤及代码示例： #

1. 导入正确的依赖类
   确保使用langchain_ollama.ChatOllama（旧版langchain.llms.Ollama不支持工具调用），同时替换输出解析器为适配Ollama的类型：

   from langchain_ollama import ChatOllama
   from langchain_core.prompts import ChatPromptTemplate, MessagesPlaceholder
   from langchain_core.output_parsers import JsonOutputToolsParser
   

2. 替换绑定方法与解析器
   将llm.bind_functions改为llm.bind_tools，并使用JsonOutputToolsParser解析工具调用输出：

   修改后的完整代码：

   # Setup your supervisor, conditional logic, and integrate the nodes into the workflow
   members = ["Web_Searcher", "Web_Searcher_Quality","Blog_Searcher", "Blog_Searcher_Quality", "Content_Writer"]
   system_prompt = (
       """As a supervisor, your role is to oversee a dialogue between these
       workers: {members}. and execute all the agents one by one,
       determine which worker should take the next action. Each worker is responsible for
       executing a specific task and reporting back their findings and progress. Once all tasks are complete,
       indicate with 'FINISH'."""
   )
   
   options = ["FINISH"] + members
   function_def = {
       "name": "route",
       "description": "Select the next role.",
       "parameters": {
           "type": "object",
           "properties": {"next": {"type": "string", "enum": options}},
           "required": ["next"],
       },
   }
   
   prompt = ChatPromptTemplate.from_messages([
       ("system", system_prompt),
       MessagesPlaceholder(variable_name="messages"),
       ("system", "Given the conversation above, who should act next? Or should we FINISH? Select one of: {options}"),
   ]).partial(options=str(options), members=", ".join(members))
   
   # 初始化Ollama模型
   llm = ChatOllama(model="llama2", temperature=0)
   # 使用bind_tools替代bind_functions，并用JsonOutputToolsParser解析输出
   manager_chain = (prompt | llm.bind_tools(tools=[function_def], tool_choice="route") | JsonOutputToolsParser())
   

3. 关键说明

   • bind_tools是LangChain为支持工具调用的模型（含Ollama）提供的统一绑定方法，替代OpenAI专属的bind_functions
   • JsonOutputToolsParser适配Ollama工具调用的输出格式，可正确提取工具参数
   • 需确保LangChain版本在0.1.10及以上，bind_tools是较新版本新增的功能

内容的提问来源于stack exchange，提问作者Prangya Paramita