248 lines
9.2 KiB
Python
248 lines
9.2 KiB
Python
import logging
|
||
import sys
|
||
import time
|
||
import json
|
||
|
||
from Config.Config import VOLC_ACCESSKEY, VOLC_SECRETKEY, VOLC_API_KEY
|
||
from Volcengine.Kit.VikingDBMemoryService import VikingDBMemoryService, MEMORY_COLLECTION_NAME
|
||
from volcenginesdkarkruntime import Ark
|
||
|
||
# 控制日志输出
|
||
logger = logging.getLogger('ChatWithMemory')
|
||
logger.setLevel(logging.INFO)
|
||
|
||
# 只添加一次处理器，避免重复日志
|
||
if not logger.handlers:
|
||
handler = logging.StreamHandler()
|
||
handler.setFormatter(logging.Formatter('%(asctime)s - %(name)s - %(levelname)s - %(message)s'))
|
||
logger.addHandler(handler)
|
||
|
||
|
||
def initialize_services():
|
||
"""初始化服务和LLM客户端"""
|
||
ak = VOLC_ACCESSKEY
|
||
sk = VOLC_SECRETKEY
|
||
ark_api_key = VOLC_API_KEY
|
||
|
||
if not all([ak, sk, ark_api_key]):
|
||
raise ValueError("必须在环境变量中设置 VOLC_ACCESSKEY, VOLC_SECRETKEY, 和 ARK_API_KEY。")
|
||
|
||
memory_service = VikingDBMemoryService(
|
||
ak=ak,
|
||
sk=sk,
|
||
host="api-knowledgebase.mlp.cn-beijing.volces.com",
|
||
region="cn-beijing"
|
||
)
|
||
llm_client = Ark(
|
||
base_url="https://ark.cn-beijing.volces.com/api/v3",
|
||
api_key=ark_api_key,
|
||
)
|
||
return memory_service, llm_client
|
||
|
||
|
||
def search_relevant_memories(memory_service, collection_name, user_id, query):
|
||
"""搜索与用户查询相关的记忆，并在索引构建中时重试。"""
|
||
logger.info(f"正在搜索与 '{query}' 相关的记忆...")
|
||
retry_attempt = 0
|
||
while True:
|
||
try:
|
||
filter_params = {
|
||
"user_id": [user_id],
|
||
"memory_type": ["sys_event_v1", "sys_profile_v1"]
|
||
}
|
||
response = memory_service.search_memory(
|
||
collection_name=collection_name,
|
||
query=query,
|
||
filter=filter_params,
|
||
limit=3
|
||
)
|
||
|
||
memories = []
|
||
if response.get('data', {}).get('count', 0) > 0:
|
||
for result in response['data']['result_list']:
|
||
if 'memory_info' in result and result['memory_info']:
|
||
memories.append({
|
||
'memory_info': result['memory_info'],
|
||
'score': result['score']
|
||
})
|
||
|
||
if memories:
|
||
if retry_attempt > 0:
|
||
logger.info("重试后搜索成功。")
|
||
logger.info(f"找到 {len(memories)} 条相关记忆:")
|
||
for i, memory in enumerate(memories, 1):
|
||
logger.info(
|
||
f" {i}. (相关度: {memory['score']:.3f}): {json.dumps(memory['memory_info'], ensure_ascii=False, indent=2)}")
|
||
else:
|
||
logger.info("未找到相关记忆。")
|
||
return memories
|
||
|
||
except Exception as e:
|
||
error_message = str(e)
|
||
if "1000023" in error_message:
|
||
retry_attempt += 1
|
||
logger.info(f"记忆索引正在构建中。将在60秒后重试... (尝试次数 {retry_attempt})")
|
||
time.sleep(60)
|
||
else:
|
||
logger.info(f"搜索记忆时出错 (不可重试): {e}")
|
||
return []
|
||
|
||
|
||
def handle_conversation_turn(memory_service, llm_client, collection_name, user_id, user_message, conversation_history):
|
||
"""处理一轮对话，包括记忆搜索和LLM响应。"""
|
||
logger.info("\n" + "=" * 60)
|
||
logger.info(f"用户: {user_message}")
|
||
|
||
relevant_memories = search_relevant_memories(memory_service, collection_name, user_id, user_message)
|
||
|
||
system_prompt = "你是一个富有同情心、善于倾听的AI伙伴，拥有长期记忆能力。你的目标是为用户提供情感支持和温暖的陪伴。"
|
||
if relevant_memories:
|
||
memory_context = "\n".join(
|
||
[f"- {json.dumps(mem['memory_info'], ensure_ascii=False)}" for mem in relevant_memories])
|
||
system_prompt += f"\n\n这是我们过去的一些对话记忆，请参考：\n{memory_context}\n\n请利用这些信息来更好地理解和回应用户。"
|
||
|
||
logger.info("AI正在思考...")
|
||
|
||
try:
|
||
messages = [{"role": "system", "content": system_prompt}] + conversation_history + [
|
||
{"role": "user", "content": user_message}]
|
||
completion = llm_client.chat.completions.create(
|
||
model="doubao-seed-1-6-flash-250715",
|
||
messages=messages
|
||
)
|
||
assistant_reply = completion.choices[0].message.content
|
||
except Exception as e:
|
||
logger.info(f"LLM调用失败: {e}")
|
||
assistant_reply = "抱歉，我现在有点混乱，无法回应。我们可以稍后再聊吗？"
|
||
|
||
logger.info(f"伙伴: {assistant_reply}")
|
||
|
||
conversation_history.extend([
|
||
{"role": "user", "content": user_message},
|
||
{"role": "assistant", "content": assistant_reply}
|
||
])
|
||
return assistant_reply
|
||
|
||
|
||
def archive_conversation(memory_service, collection_name, user_id, assistant_id, conversation_history, topic_name):
|
||
"""将对话历史归档到记忆数据库。"""
|
||
if not conversation_history:
|
||
logger.info("没有对话可以归档。")
|
||
return False
|
||
|
||
logger.info(f"\n正在归档关于 '{topic_name}' 的对话...")
|
||
session_id = f"{topic_name}_{int(time.time())}"
|
||
metadata = {
|
||
"default_user_id": user_id,
|
||
"default_assistant_id": assistant_id,
|
||
"time": int(time.time() * 1000)
|
||
}
|
||
|
||
try:
|
||
memory_service.add_session(
|
||
collection_name=collection_name,
|
||
session_id=session_id,
|
||
messages=conversation_history,
|
||
metadata=metadata
|
||
)
|
||
logger.info(f"对话已成功归档，会话ID: {session_id}")
|
||
logger.info("正在等待记忆索引更新...")
|
||
return True
|
||
except Exception as e:
|
||
logger.info(f"归档对话失败: {e}")
|
||
return False
|
||
|
||
|
||
def main():
|
||
logger.info("开始测试大模型记忆功能...")
|
||
|
||
try:
|
||
# 使用initialize_services函数初始化服务和LLM客户端
|
||
memory_service, llm_client = initialize_services()
|
||
|
||
collection_name = MEMORY_COLLECTION_NAME
|
||
user_id = "liming"
|
||
assistant_id = "assistant"
|
||
|
||
# 告知大模型用户信息
|
||
logger.info("告知大模型用户信息...")
|
||
user_info = "李明，男生，15岁，家住长春"
|
||
|
||
# 记录信息到记忆体
|
||
logger.info("记录用户信息到记忆体...")
|
||
|
||
# 准备对话历史
|
||
conversation_history = []
|
||
|
||
# 使用正确的handle_conversation_turn方法参数
|
||
response = handle_conversation_turn(
|
||
memory_service=memory_service,
|
||
llm_client=llm_client,
|
||
collection_name=collection_name,
|
||
user_id=user_id,
|
||
user_message=f"请记住以下用户信息：{user_info}",
|
||
conversation_history=conversation_history
|
||
)
|
||
|
||
logger.info(f"模型回复: {response}")
|
||
|
||
# 归档对话
|
||
archive_conversation(
|
||
memory_service=memory_service,
|
||
collection_name=collection_name,
|
||
user_id=user_id,
|
||
assistant_id=assistant_id,
|
||
conversation_history=conversation_history,
|
||
topic_name="user_info"
|
||
)
|
||
|
||
# 等待一段时间确保索引更新
|
||
logger.info("等待索引更新...")
|
||
time.sleep(5)
|
||
|
||
# 验证大模型是否记住个人信息
|
||
logger.info("验证大模型是否记住个人信息...")
|
||
test_conversation_history = []
|
||
|
||
test_response = handle_conversation_turn(
|
||
memory_service=memory_service,
|
||
llm_client=llm_client,
|
||
collection_name=collection_name,
|
||
user_id=user_id,
|
||
user_message="请告诉我李明的个人信息",
|
||
conversation_history=test_conversation_history
|
||
)
|
||
|
||
logger.info(f"测试回复: {test_response}")
|
||
|
||
# 检查回复中是否包含关键信息
|
||
keywords = ["李明", "男", "15", "长春"]
|
||
found_keywords = [kw for kw in keywords if kw in test_response]
|
||
|
||
if len(found_keywords) == len(keywords):
|
||
logger.info("✅ 大模型成功记住了用户信息！")
|
||
else:
|
||
logger.info(f"❌ 大模型可能没有完全记住用户信息。找到的关键词: {found_keywords}")
|
||
|
||
# 尝试直接搜索记忆
|
||
logger.info("尝试直接搜索记忆...")
|
||
filter_params = {
|
||
"user_id": [user_id],
|
||
"memory_type": ["sys_event_v1", "sys_profile_v1"]
|
||
}
|
||
search_result = memory_service.search_memory(
|
||
collection_name=collection_name,
|
||
query="李明 15岁 长春 男生",
|
||
filter=filter_params,
|
||
limit=5
|
||
)
|
||
|
||
logger.info(f"搜索结果: {search_result}")
|
||
|
||
except Exception as e:
|
||
logger.error(f"操作失败: {e}")
|
||
sys.exit(1)
|
||
|
||
|
||
if __name__ == "__main__":
|
||
main() |