Mem0

from langgraph.graph import StateGraph
from mem0 import Memory

memory = Memory()

def memory_node(state):
    # 检索相关记忆
    memories = memory.search(
        query=state["query"],
        user_id=state["user_id"],
        limit=5
    )

    # 将记忆添加到状态
    state["memories"] = memories
    return state

def response_node(state):
    # 基于记忆生成回复
    context = "\n".join([m["memory"] for m in state["memories"]])
    # ... 生成回复逻辑
    return state

# 构建图
graph = StateGraph()
graph.add_node("memory", memory_node)
graph.add_node("response", response_node)
graph.add_edge("memory", "response")

from crewai import Agent, Task, Crew
from mem0 import Memory

memory = Memory()

# 创建具有记忆能力的代理
agent = Agent(
    role="Personal Assistant",
    goal="帮助用户管理日常任务",
    backstory="我是一个具有记忆能力的个人助手",
    tools=[
        # 自定义记忆工具
        lambda query: memory.search(query, user_id="user_123"),
        lambda data: memory.add(data, user_id="user_123")
    ]
)

task = Task(
    description="根据用户历史偏好推荐餐厅",
    agent=agent
)

crew = Crew(agents=[agent], tasks=[task])

import autogen
from mem0 import Memory

memory = Memory()

class MemoryAgent(autogen.ConversableAgent):
    def __init__(self, name, **kwargs):
        super().__init__(name, **kwargs)
        self.memory = memory

    def generate_reply(self, messages, sender, **kwargs):
        # 检索相关记忆
        last_message = messages[-1]["content"]
        memories = self.memory.search(
            query=last_message,
            user_id=sender.name,
            limit=3
        )

        # 将记忆添加到上下文
        context = f"相关记忆: {memories}"
        enhanced_messages = messages + [{"role": "system", "content": context}]

        # 生成回复
        reply = super().generate_reply(enhanced_messages, sender, **kwargs)

        # 存储新记忆
        self.memory.add(
            [{"role": "user", "content": last_message},
             {"role": "assistant", "content": reply}],
            user_id=sender.name
        )

        return reply