codefuse-chatbot/coagent/connector/chains/base_chain.py

123 lines
6.0 KiB
Python

from typing import List
from loguru import logger
import copy, os
from coagent.connector.agents import BaseAgent
from coagent.connector.schema import (
Memory, Role, Message, ActionStatus, ChainConfig,
load_role_configs
)
from coagent.connector.memory_manager import BaseMemoryManager
from coagent.connector.message_process import MessageUtils
from coagent.llm_models.llm_config import LLMConfig, EmbedConfig
from coagent.connector.configs.agent_config import AGETN_CONFIGS
role_configs = load_role_configs(AGETN_CONFIGS)
# from configs.model_config import JUPYTER_WORK_PATH
# from configs.server_config import SANDBOX_SERVER
class BaseChain:
def __init__(
self,
# chainConfig: ChainConfig,
agents: List[BaseAgent],
chat_turn: int = 1,
do_checker: bool = False,
sandbox_server: dict = {},
jupyter_work_path: str = "",
kb_root_path: str = "",
llm_config: LLMConfig = LLMConfig(),
embed_config: EmbedConfig = None,
log_verbose: str = "0"
) -> None:
# self.chainConfig = chainConfig
self.agents: List[BaseAgent] = agents
self.chat_turn = chat_turn
self.do_checker = do_checker
self.sandbox_server = sandbox_server
self.jupyter_work_path = jupyter_work_path
self.llm_config = llm_config
self.log_verbose = max(os.environ.get("log_verbose", "0"), log_verbose)
self.checker = BaseAgent(role=role_configs["checker"].role,
prompt_config=role_configs["checker"].prompt_config,
task = None, memory = None,
llm_config=llm_config, embed_config=embed_config,
sandbox_server=sandbox_server, jupyter_work_path=jupyter_work_path,
kb_root_path=kb_root_path
)
self.messageUtils = MessageUtils(None, sandbox_server, self.jupyter_work_path, embed_config, llm_config, kb_root_path, log_verbose)
# all memory created by agent until instance deleted
self.global_memory = Memory(messages=[])
def step(self, query: Message, history: Memory = None, background: Memory = None, memory_manager: BaseMemoryManager = None) -> Message:
'''execute chain'''
for output_message, local_memory in self.astep(query, history, background, memory_manager):
pass
return output_message, local_memory
def pre_print(self, query: Message, history: Memory = None, background: Memory = None, memory_manager: BaseMemoryManager = None) -> Message:
'''execute chain'''
for agent in self.agents:
agent.pre_print(query, history, background=background, memory_manager=memory_manager)
def astep(self, query: Message, history: Memory = None, background: Memory = None, memory_manager: BaseMemoryManager = None) -> Message:
'''execute chain'''
local_memory = Memory(messages=[])
input_message = copy.deepcopy(query)
step_nums = copy.deepcopy(self.chat_turn)
check_message = None
self.global_memory.append(input_message)
# local_memory.append(input_message)
while step_nums > 0:
for agent in self.agents:
for output_message in agent.astep(input_message, history, background=background, memory_manager=memory_manager):
# logger.debug(f"local_memory {local_memory + output_message}")
yield output_message, local_memory + output_message
output_message = self.messageUtils.inherit_extrainfo(input_message, output_message)
# according the output to choose one action for code_content or tool_content
output_message = self.messageUtils.parser(output_message)
yield output_message, local_memory + output_message
# output_message = self.step_router(output_message)
input_message = output_message
self.global_memory.append(output_message)
local_memory.append(output_message)
# when get finished signal can stop early
if output_message.action_status == ActionStatus.FINISHED or output_message.action_status == ActionStatus.STOPPED:
action_status = False
break
if output_message.action_status == ActionStatus.FINISHED:
break
if self.do_checker and self.chat_turn > 1:
for check_message in self.checker.astep(query, background=local_memory, memory_manager=memory_manager):
pass
check_message = self.messageUtils.parser(check_message)
check_message = self.messageUtils.inherit_extrainfo(output_message, check_message)
# logger.debug(f"{self.checker.role.role_name}: {check_message.role_content}")
if check_message.action_status == ActionStatus.FINISHED:
self.global_memory.append(check_message)
break
step_nums -= 1
#
output_message = check_message or output_message # 返回chain和checker的结果
output_message.input_query = query.input_query # chain和chain之间消息通信不改变问题
yield output_message, local_memory
def get_memory(self, content_key="role_content") -> Memory:
memory = self.global_memory
return memory.to_tuple_messages(content_key=content_key)
def get_memory_str(self, content_key="role_content") -> Memory:
memory = self.global_memory
return "\n".join([": ".join(i) for i in memory.to_tuple_messages(content_key=content_key)])
def get_agents_memory(self, content_key="role_content"):
return [agent.get_memory(content_key=content_key) for agent in self.agents]
def get_agents_memory_str(self, content_key="role_content"):
return "************".join([f"{agent.role.role_name}\n" + agent.get_memory_str(content_key=content_key) for agent in self.agents])