gyt/opendan
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

define bas environment

Browse Source
This commit is contained in:
tsukasa
2023-12-01 14:29:10 +08:00
parent 66e407add5
commit b7b968d5f7
18 changed files with 1700 additions and 1023 deletions
Download Patch File Download Diff File Expand all files Collapse all files
src/aios/agent/agent.py
+185 -165
View File
@@ -18,6 +18,7 @@ from ..proto.compute_task import ComputeTaskResult,ComputeTaskResultCode
from .agent_base import *
from .chatsession import *
from .ai_function import *
from ..environment.workspace_env import WorkspaceEnvironment, TodoListType
from ..frame.contact_manager import ContactManager,Contact,FamilyMember
from ..frame.compute_kernel import ComputeKernel
@@ -145,17 +146,22 @@ class AIAgent(BaseAIAgent):
self.contact_prompt_str = None
self.history_len = 10
self.review_todo_prompt = None
self.read_report_prompt = None
self.do_prompt = None
self.check_prompt = None
self.goal_to_todo_prompt = None
todo_prompts = {}
todo_prompts[TodoListType.TO_WORK] = {
"do": DEFAULT_AGENT_DO_PROMPT,
"check": DEFAULT_AGENT_SELF_CHECK_PROMPT,
"review": None,
}
todo_prompts[TodoListType.TO_LEARN] = {
"do": DEFAULT_AGENT_LEARN_PROMPT,
"check": None,
"review": None,
}
self.todo_prompts = todo_prompts
self.learn_token_limit = 4000
self.learn_prompt = AgentPrompt(DEFAULT_AGENT_LEARN_PROMPT)
self.chat_db = None
self.unread_msg = Queue() # msg from other agent
@@ -163,19 +169,18 @@ class AIAgent(BaseAIAgent):
self.owenr_bus = None
self.enable_function_list = None
@classmethod
def create_from_templete(cls,templete:AIAgentTemplete, fullname:str):
# Agent just inherit from templete on craete,if template changed,agent will not change
result_agent = AIAgent()
result_agent.llm_model_name = templete.llm_model_name
result_agent.max_token_size = templete.max_token_size
result_agent.template_id = templete.template_id
result_agent.agent_id = "agent#" + uuid.uuid4().hex
result_agent.fullname = fullname
result_agent.powerby = templete.author
result_agent.agent_prompt = templete.prompt
return result_agent
# @classmethod
# def create_from_templete(cls,templete:AIAgentTemplete, fullname:str):
# # Agent just inherit from templete on craete,if template changed,agent will not change
# result_agent = AIAgent()
# result_agent.llm_model_name = templete.llm_model_name
# result_agent.max_token_size = templete.max_token_size
# result_agent.template_id = templete.template_id
# result_agent.agent_id = "agent#" + uuid.uuid4().hex
# result_agent.fullname = fullname
# result_agent.powerby = templete.author
# result_agent.agent_prompt = templete.prompt
# return result_agent
def load_from_config(self,config:dict) -> bool:
if config.get("instance_id") is None:
@@ -200,11 +205,25 @@ class AIAgent(BaseAIAgent):
self.agent_think_prompt = AgentPrompt()
self.agent_think_prompt.load_from_config(config["think_prompt"])
if config.get("do_prompt") is not None:
self.do_prompt = AgentPrompt()
self.do_prompt.load_from_config(config["do_prompt"])
self.wake_up()
def load_todo_config(todo_type:str) -> bool:
todo_config = config.get(todo_type)
if todo_config is not None:
if todo_config.get("do") is not None:
prompt = AgentPrompt()
prompt.load_from_config(todo_config["do"])
self.todo_prompts[todo_type]["do"] = prompt
if todo_config.get("check") is not None:
prompt = AgentPrompt()
prompt.load_from_config(todo_config["check"])
self.todo_prompts[todo_type]["check"] = prompt
if todo_config.get("review_prompt") is not None:
prompt = AgentPrompt()
prompt.load_from_config(todo_config["review_prompt"])
self.todo_prompts[todo_type]["review"] = prompt
load_todo_config(TodoListType.TO_WORK)
load_todo_config(TodoListType.TO_LEARN)
if config.get("guest_prompt") is not None:
self.guest_prompt_str = config["guest_prompt"]
@@ -234,6 +253,9 @@ class AIAgent(BaseAIAgent):
self.enable_timestamp = bool(config["enable_timestamp"])
if config.get("history_len"):
self.history_len = int(config.get("history_len"))
self.wake_up()
return True
def get_id(self) -> str:
@@ -372,7 +394,7 @@ class AIAgent(BaseAIAgent):
# self._format_msg_by_env_value(prompt)
# inner_functions,function_token_len = self._get_inner_functions()
# system_prompt_len = prompt.get_prompt_token_len()
# system_prompt_len = self.token_len(prompt=prompt)
# input_len = len(msg.body)
# history_prmpt,history_token_len = await self._get_prompt_from_session_for_groupchat(chatsession,system_prompt_len + function_token_len,input_len)
@@ -411,10 +433,10 @@ class AIAgent(BaseAIAgent):
# resp_msg = msg.create_group_resp_msg(self.agent_id,final_result)
# chatsession.append(msg)
# chatsession.append(resp_msg)
# return resp_msg
# return None
def get_workspace_by_msg(self,msg:AgentMsg) -> WorkspaceEnvironment:
return self.agent_workspace
@@ -528,7 +550,7 @@ class AIAgent(BaseAIAgent):
have_known_info = True
known_info_str += f"## todo\n{todos_str}\n"
inner_functions,function_token_len = BaseAIAgent.get_inner_functions(self.owner_env)
system_prompt_len = prompt.get_prompt_token_len()
system_prompt_len = self.token_len(prompt=prompt)
input_len = len(msg.body)
if msg.msg_type == AgentMsgType.TYPE_GROUPMSG:
history_str,history_token_len = await self._get_prompt_from_session_for_groupchat(chatsession,system_prompt_len + function_token_len,input_len)
@@ -600,9 +622,7 @@ class AIAgent(BaseAIAgent):
return None
async def _get_history_prompt_for_think(self,chatsession:AIChatSession,summary:str,system_token_len:int,pos:int)->(AgentPrompt,int):
history_len = (self.max_token_size * 0.7) - system_token_len
messages = chatsession.read_history(self.history_len,pos,"natural") # read
@@ -716,9 +736,7 @@ class AIAgent(BaseAIAgent):
worksapce.set_work_summary(self.agent_id,task_result.result_str)
# 尝试完成自己的TOOD （不依赖任何其他Agnet）
async def do_my_work(self) -> None:
async def _llm_run_todo_list(self, todo_list_type: TodoListType):
workspace : WorkspaceEnvironment = self.get_workspace_by_msg(None)
logger.info(f"agent {self.agent_id} do my work start!")
@@ -726,134 +744,154 @@ class AIAgent(BaseAIAgent):
#if await self.need_review_todolist():
# await self._llm_review_todolist(workspace)
todo_list = await workspace.get_todo_list(self.agent_id)
todo_list = workspace.todo_list[todo_list_type]
need_todo = todo_list.get_todo_list(self.agent_id)
check_count = 0
do_count = 0
review_count = 0
for todo in todo_list:
for todo in need_todo:
if self.agent_energy <= 0:
break
do_prompts = self._can_do_todo(todo_list_type, todo)
if do_prompts:
prompt : AgentPrompt = AgentPrompt()
prompt.append(self.agent_prompt)
prompt.append(workspace.get_role_prompt(self.agent_id))
prompt.append(do_prompts)
prompt.append(todo.to_prompt())
do_result : AgentTodoResult = await self._llm_do_todo(todo, prompt, workspace)
todo.last_do_time = datetime.datetime.now().timestamp()
todo.retry_count += 1
match do_result.result_code:
case AgentTodoResult.TODO_RESULT_CODE_LLM_ERROR:
continue
case AgentTodoResult.TODO_RESULT_CODE_OK:
await todo_list.update_todo(todo.todo_id,AgentTodo.TODO_STATE_WAITING_CHECK)
case AgentTodoResult.TODO_RESULT_CODE_EXEC_OP_ERROR:
await todo_list.update_todo(todo.todo_id,AgentTodo.TODO_STATE_EXEC_FAILED)
if await self.need_review_todo(todo,workspace):
review_result = await self._llm_review_todo(todo,workspace)
todo.last_review_time = datetime.datetime.now().timestamp()
await todo_list.append_worklog(todo,do_result)
self.agent_energy -= 2
do_count += 1
# review_result = await self._llm_review_todo(todo,workspace)
# todo.last_review_time = datetime.datetime.now().timestamp()
continue
elif await self.can_check(todo,workspace):
check_result : AgentTodoResult = await self._llm_check_todo(todo,workspace)
check_prompts = self._can_check_todo(todo_list_type, todo)
if check_prompts:
prompt : AgentPrompt = AgentPrompt()
prompt.append(self.agent_prompt)
prompt.append(workspace.get_role_prompt(self.agent_id))
prompt.append(check_prompts)
if todo.last_check_result:
prompt.append(AgentPrompt(todo.last_check_result))
prompt.append(todo.detail)
prompt.append(todo.result)
check_result: AgentTodoResult = await self._llm_check_todo(todo, prompt, workspace)
todo.last_check_time = datetime.datetime.now().timestamp()
match check_result.result_code:
case AgentTodoResult.TODO_RESULT_CODE_LLM_ERROR:
continue
case AgentTodoResult.TODO_RESULT_CODE_OK:
await workspace.update_todo(todo.todo_id,AgentTodo.TODO_STATE_DONE)
await todo_list.update_todo(todo.todo_id,AgentTodo.TODO_STATE_DONE)
case AgentTodoResult.TODO_RESULT_CODE_EXEC_OP_ERROR:
await workspace.update_todo(todo.todo_id,AgentTodo.TDDO_STATE_CHECKFAILED)
await todo_list.update_todo(todo.todo_id,AgentTodo.TDDO_STATE_CHECKFAILED)
await workspace.append_worklog(todo,check_result)
await todo_list.append_worklog(todo, check_result)
self.agent_energy -= 1
check_count += 1
elif await self.can_do(todo,workspace):
do_result : AgentTodoResult = await self._llm_do(todo,workspace)
todo.last_do_time = datetime.datetime.now().timestamp()
todo.retry_count += 1
continue
review_prompts = self._can_review_todo(todo_list_type, todo)
if review_prompts:
prompt.append(workspace.get_prompt())
prompt.append(workspace.get_role_prompt(self.agent_id))
prompt.append(review_prompts)
todo_tree = todo_list.get_todo_tree("/")
prompt.append(AgentPrompt(todo_tree))
do_result : AgentTodoResult = await self._llm_review_todo(todo, prompt, workspace)
todo.last_review_time = datetime.datetime.now().timestamp()
match do_result.result_code:
case AgentTodoResult.TODO_RESULT_CODE_LLM_ERROR:
continue
case AgentTodoResult.TODO_RESULT_CODE_OK:
await workspace.update_todo(todo.todo_id,AgentTodo.TODO_STATE_WAITING_CHECK)
case AgentTodoResult.TODO_RESULT_CODE_EXEC_OP_ERROR:
await workspace.update_todo(todo.todo_id,AgentTodo.TODO_STATE_EXEC_FAILED)
continue
case AgentTodoResult.TODO_RESULT_CODE_OK:
await todo_list.update_todo(todo.todo_id,AgentTodo.TODO_STATE_REVIEWED)
await workspace.append_worklog(todo,do_result)
self.agent_energy -= 2
do_count += 1
await todo_list.append_worklog(todo,do_result)
self.agent_energy -= 1
review_count += 1
continue
logger.info(f"agent {self.agent_id} ,check:{check_count} todo,do:{do_count} todo.")
def _can_review_todo(self, todo_list_type: TodoListType, todo:AgentTodo) -> AgentPrompt:
do_prompts = self.todo_prompts[todo_list_type].get("review")
if not do_prompts:
return None
def get_review_todo_prompt(self,todo:AgentTodo) -> AgentPrompt:
return self.review_todo_prompt
if todo.can_review() is False:
return None
async def _llm_review_todo(self,todo:AgentTodo,workspace:WorkspaceEnvironment):
prompt = AgentPrompt()
return do_prompts
prompt.append(workspace.get_prompt())
prompt.append(workspace.get_role_prompt(self.agent_id))
prompt.append(self.get_review_todo_prompt(todo))
todo_tree = workspace.get_todo_tree("/")
prompt.append(AgentPrompt(todo_tree))
inner_functions,_ = BaseAIAgent.get_inner_functions(self.owner_env)
task_result:ComputeTaskResult = await self.do_llm_complection(prompt,inner_functions=inner_functions)
if task_result.result_code != ComputeTaskResultCode.OK:
logger.error(f"_llm_review_todos compute error:{task_result.error_str}")
return
return
def get_do_prompt(self,todo:AgentTodo) -> AgentPrompt:
return self.do_prompt
def get_prompt_from_todo(self,todo:AgentTodo) -> AgentPrompt:
json_str = json.dumps(todo.raw_obj)
return AgentPrompt(json_str)
async def need_review_todo(self,todo:AgentTodo,workspace:WorkspaceEnvironment) -> bool:
return False
async def can_check(self,todo:AgentTodo,workspace:WorkspaceEnvironment) -> bool:
if self.get_check_prompt(todo) is None:
return False
def _can_check_todo(self, todo_list_type: TodoListType, todo:AgentTodo) -> AgentPrompt:
do_prompts = self.todo_prompts[todo_list_type].get("check")
if not do_prompts:
return None
if todo.can_check() is False:
return False
return None
if todo.checker is not None:
if todo.checker != self.agent_id:
return False
return None
else:
if self.can_do_unassigned_task is False:
return False
return None
else:
todo.checker = self.agent_id
return True
return do_prompts
async def can_do(self,todo:AgentTodo,workspace:WorkspaceEnvironment) -> bool:
async def _can_do_todo(self, todo_list_type: TodoListType, todo:AgentTodo) -> AgentPrompt:
do_prompts = self.todo_prompts[todo_list_type].get("do")
if not do_prompts:
return None
if todo.can_do() is False:
return False
return None
if todo.worker is not None:
if todo.worker != self.agent_id:
return False
return None
else:
if self.can_do_unassigned_task is False:
return False
return None
else:
todo.worker = self.agent_id
return True
return do_prompts
async def _llm_do(self,todo:AgentTodo,workspace:WorkspaceEnvironment) -> AgentTodoResult:
async def _llm_do_todo(self, todo: AgentTodo, prompt: AgentPrompt, workspace: WorkspaceEnvironment) -> AgentTodoResult:
result = AgentTodoResult()
prompt : AgentPrompt = AgentPrompt()
#prompt.append(self.agent_prompt)
prompt.append(workspace.get_role_prompt(self.agent_id))
do_prompt = workspace.get_do_prompt(todo)
if do_prompt is None:
do_prompt = self.get_do_prompt(todo)
prompt.append(do_prompt)
# There are general methods for executing todos, as well as customized ones that are more efficient for specific types of TODOS.
# Based on experience, an Agent can autonomously master/organize execution methods for a greater variety of TODO types.
#prompt.append(work_log_prompt)
prompt.append(self.get_prompt_from_todo(todo))
task_result:ComputeTaskResult = await self.do_llm_complection(prompt)
if task_result.error_str is not None:
logger.error(f"_llm_do compute error:{task_result.error_str}")
@@ -875,7 +913,7 @@ class AIAgent(BaseAIAgent):
resp = await AIBus.get_default_bus().post_message(msg)
logging.info(f"agent {self.agent_id} send msg to {msg.target} result:{resp}")
op_errors,have_error = await workspace.exec_op_list(llm_result.op_list,self.agent_id)
op_errors, have_error = await workspace.exec_op_list(llm_result.op_list, self.agent_id)
if have_error:
result.result_code = AgentTodoResult.TODO_RESULT_CODE_EXEC_OP_ERROR
#result.error_str = error_str
@@ -883,37 +921,30 @@ class AIAgent(BaseAIAgent):
return result
async def append_toddo_result(self,todo,worksapce,llm_result,result_str):
pass
def get_check_prompt(self,todo:AgentTodo) -> AgentPrompt:
return self.check_prompt
async def _llm_check_todo(self, todo:AgentTodo,workspace:WorkspaceEnvironment) :
if self.get_check_prompt(todo) is None:
return None
prompt : AgentPrompt = AgentPrompt()
prompt.append(self.agent_prompt)
prompt.append(workspace.get_role_prompt(self.agent_id))
prompt.append(self.get_check_prompt(todo))
if todo.last_check_result:
prompt.append(AgentPrompt(todo.last_check_result))
prompt.append(todo.detail)
prompt.append(todo.result)
async def _llm_check_todo(self, todo: AgentTodo, prompt: AgentPrompt, workspace: WorkspaceEnvironment) -> AgentTodoResult:
result = AgentTodoResult()
inner_functions,_ = BaseAIAgent.get_inner_functions(workspace)
task_result:ComputeTaskResult = await self.do_llm_complection(prompt,inner_functions=inner_functions,is_json_resp=True)
if task_result.result_code != ComputeTaskResultCode.OK:
logger.error(f"_llm_check_todo compute error:{task_result.error_str}")
return False
if task_result.result_str == "OK":
return True
if task_result.error_str is not None:
logger.error(f"_llm_do compute error:{task_result.error_str}")
result.result_code = AgentTodoResult.TODO_RESULT_CODE_LLM_ERROR
result.error_str = task_result.error_str
return result
result.result_str = task_result.result_str
todo.last_check_result = task_result.result_str
return False
return result
async def _llm_review_todo(self, todo:AgentTodo, prompt: AgentPrompt, workspace: WorkspaceEnvironment):
inner_functions,_ = BaseAIAgent.get_inner_functions(self.owner_env)
task_result:ComputeTaskResult = await self.do_llm_complection(prompt,inner_functions=inner_functions)
if task_result.result_code != ComputeTaskResultCode.OK:
logger.error(f"_llm_review_todos compute error:{task_result.error_str}")
return
return
# 尝试自我学习，会主动获取、读取资料并进行整理
# LLM的本质能力是处理海量知识，应该让LLM能基于知识把自己的工作处理的更好
@@ -1121,16 +1152,15 @@ class AIAgent(BaseAIAgent):
used_energy = await self.think_chatsession(session_id)
self.agent_energy -= used_energy
todo_logs = await self.get_todo_logs()
for todo_log in todo_logs:
if self.agent_energy <= 0:
break
used_energy = await self.think_todo_log(todo_log)
self.agent_energy -= used_energy
# todo_logs = await self.get_todo_logs()
# for todo_log in todo_logs:
# if self.agent_energy <= 0:
# break
# used_energy = await self.think_todo_log(todo_log)
# self.agent_energy -= used_energy
return
async def think_todo_log(self,todo_log:AgentWorkLog):
pass
@@ -1146,7 +1176,7 @@ class AIAgent(BaseAIAgent):
prompt:AgentPrompt = AgentPrompt()
#prompt.append(self._get_agent_prompt())
prompt.append(await self._get_agent_think_prompt())
system_prompt_len = prompt.get_prompt_token_len()
system_prompt_len = self.token_len(prompt=prompt)
#think env?
history_prompt,next_pos = await self._get_history_prompt_for_think(chatsession,summary,system_prompt_len,cur_pos)
prompt.append(history_prompt)
@@ -1220,11 +1250,7 @@ class AIAgent(BaseAIAgent):
def need_self_think(self) -> bool:
return False
def need_self_learn(self) -> bool:
if self.learn_prompt is not None:
return True
return False
def wake_up(self) -> None:
if self.agent_task is None:
self.agent_task = asyncio.create_task(self._on_timer())
@@ -1248,26 +1274,20 @@ class AIAgent(BaseAIAgent):
continue
# complete & check todo
if self.need_work():
await self.do_my_work()
# review other's todo
# self.review_other_works()
await self._llm_run_todo_list(TodoListType.TO_WORK)
await self._llm_run_todo_list(TodoListType.TO_LEARN)
if self.need_self_think():
await self.do_self_think()
if self.need_self_learn():
await self.do_self_learn()
# review other's todo
# self.review_other_works()
except Exception as e:
tb_str = traceback.format_exc()
logger.error(f"agent {self.agent_id} on timer error:{e},{tb_str}")
continue
def token_len(self,text:str) -> int:
return ComputeKernel.llm_num_tokens_from_text(text,self.get_llm_model_name())
src/aios/agent/agent_base.py
+31 -14
View File
@@ -56,16 +56,6 @@ class AgentPrompt:
self.messages.extend(prompt.messages)
def get_prompt_token_len(self):
result = 0
if self.system_message:
result += len(self.system_message.get("content"))
for msg in self.messages:
result += len(msg.get("content"))
return result
def load_from_config(self,config:list) -> bool:
if isinstance(config,list) is not True:
logger.error("prompt is not list!")
@@ -245,8 +235,9 @@ class AgentTodo:
TODO_STATE_EXEC_FAILED = "exec_failed"
TDDO_STATE_CHECKFAILED = "check_failed"
TODO_STATE_CASNCEL = "cancel"
TODO_STATE_CANCEL = "cancel"
TODO_STATE_DONE = "done"
TODO_STATE_REVIEWED = "reviewed"
TODO_STATE_EXPIRED = "expired"
def __init__(self):
@@ -341,6 +332,23 @@ class AgentTodo:
result["retry_count"] = self.retry_count
return result
def to_prompt(self) -> AgentPrompt:
json_str = json.dumps(self.raw_obj)
return AgentPrompt(json_str)
def can_review(self) -> bool:
if self.state != AgentTodo.TODO_STATE_DONE:
return False
now = datetime.now().timestamp()
if self.last_review_time:
time_diff = now - self.last_review_time
if time_diff < 60*15:
logger.info(f"todo {self.title} is already reviewed, ignore")
return False
return True
def can_check(self)->bool:
if self.state != AgentTodo.TODO_STATE_WAITING_CHECK:
@@ -410,9 +418,18 @@ class BaseAIAgent(abc.ABC):
def get_max_token_size(self) -> int:
pass
@abstractmethod
async def _process_msg(self,msg:AgentMsg,workspace = None) -> AgentMsg:
pass
def token_len(self, text:str=None, prompt:AgentPrompt=None) -> int:
from .compute_kernel import ComputeKernel
if text:
return ComputeKernel.llm_num_tokens_from_text(text,self.get_llm_model_name())
elif prompt:
result = 0
if prompt.system_message:
result += ComputeKernel.llm_num_tokens_from_text(prompt.system_message.get("content"),self.get_llm_model_name())
for msg in prompt.messages:
result += ComputeKernel.llm_num_tokens_from_text(msg.get("content"),self.get_llm_model_name())
else:
return 0
@classmethod
def get_inner_functions(cls, env:Environment) -> (dict,int):
src/aios/agent/ai_function.py
+63 -4
View File
@@ -9,9 +9,6 @@ class ParameterDefine:
class AIFunction:
def __init__(self) -> None:
self.description : str = None
@abstractmethod
def get_name(self) -> str:
"""
@@ -24,7 +21,7 @@ class AIFunction:
"""
return a detailed description of what the function does
"""
return self.description
pass
@abstractmethod
def get_parameters(self) -> Dict:
@@ -112,6 +109,9 @@ class SimpleAIFunction(AIFunction):
def get_name(self) -> str:
return self.func_id
def get_description(self) -> str:
return self.description
def get_parameters(self) -> Dict:
if self.parameters is not None:
result = {}
@@ -142,3 +142,62 @@ class SimpleAIFunction(AIFunction):
def is_ready_only(self) -> bool:
return False
class AIOperation:
@abstractmethod
def get_name(self) -> str:
"""
return the name of the operation (should be snake case)
"""
pass
@abstractmethod
def get_description(self) -> str:
"""
return a detailed description of what the operation does
"""
pass
@abstractmethod
async def execute(self, params: dict) -> str:
"""
Execute the function and return a JSON serializable dict.
The parameters are passed in the form of kwargs
"""
pass
class SimpleAIOperation(AIOperation):
def __init__(self,op:str,description:str,func_handler:Coroutine) -> None:
self.op = op
self.description = description
self.func_handler = func_handler
def get_name(self) -> str:
return self.op
def get_description(self) -> str:
return self.description
async def execute(self, params: Dict) -> str:
if self.func_handler is None:
return "error: function not implemented"
return await self.func_handler(**params)
class AIFunctionOperation(AIOperation):
def __init__(self, func: AIFunction) -> None:
self.func = func
super().__init__()
@abstractmethod
def get_name(self) -> str:
return self.func.get_name()
@abstractmethod
def get_description(self) -> str:
return self.func.get_description()
@abstractmethod
async def execute(self, params: dict) -> str:
self.func.execute(**params)