fix role._rc init

metagpt/environment.py

@@ -31,6 +31,7 @@ class Environment(BaseModel):
         arbitrary_types_allowed = True
 
     def __init__(self, **kwargs):
+        roles = []
         for role_key, role in kwargs.get("roles", {}).items():
             current_role = kwargs["roles"][role_key]
             if isinstance(current_role, dict):
@@ -41,8 +42,11 @@ class Environment(BaseModel):
                         current_role = subclass(**current_role)
                         break
                 kwargs["roles"][role_key] = current_role
+                roles.append(current_role)
         super().__init__(**kwargs)
 
+        self.add_roles(roles)  # add_roles again to init the Role.set_env
+
     def serialize(self, stg_path: Path):
         roles_path = stg_path.joinpath("roles.json")
         roles_info = []

metagpt/roles/role.py

@@ -88,13 +88,14 @@ class RoleSetting(BaseModel):
 
 class RoleContext(BaseModel):
     """Role Runtime Context"""
+    # # env exclude=True to avoid `RecursionError: maximum recursion depth exceeded in comparison`
     env: "Environment" = Field(default=None, exclude=True)
     memory: Memory = Field(default_factory=Memory)
-    long_term_memory: LongTermMemory = Field(default_factory=LongTermMemory)
+    long_term_memory: LongTermMemory = Field(default_factory=LongTermMemory, exclude=True)  # TODO not used now
     state: int = Field(default=-1)  # -1 indicates initial or termination state where todo is None
     todo: Action = Field(default=None)
     watch: set[Type[Action]] = Field(default_factory=set)
-    news: list[Type[Message]] = Field(default=[])
+    news: list[Type[Message]] = Field(default=[], exclude=True)  # TODO not used
     react_mode: RoleReactMode = RoleReactMode.REACT # see `Role._set_react_mode` for definitions of the following two attributes
     max_react_loop: int = 1
     
@@ -128,12 +129,12 @@ class Role(BaseModel):
     desc: str = ""
     is_human: bool = False
 
-    _llm: BaseGPTAPI = Field(default_factory=LLM, exclude=True)
+    _llm: BaseGPTAPI = Field(default_factory=LLM)
     _setting: RoleSetting = Field(default_factory=RoleSetting, alias=True)
     _role_id: str = ""
     _states: list[str] = Field(default=[])
     _actions: list[Action] = Field(default=[])
-    _rc: RoleContext = Field(default=RoleContext, exclude=True)
+    _rc: RoleContext = Field(default=RoleContext)
 
     # builtin variables
     recovered: bool = False  # to tag if a recovered role
@@ -179,7 +180,7 @@ class Role(BaseModel):
                     setting = RoleSetting(**kwargs[key])
                     object.__setattr__(self, "_setting", setting)
                 elif key == "_rc":
-                    _rc = RoleContext()
+                    _rc = RoleContext(**kwargs["_rc"])
                     object.__setattr__(self, "_rc", _rc)
             else:
                 if key == "_rc":

tests/metagpt/serialize_deserialize/test_team.py

@@ -39,7 +39,7 @@ def test_team_deserialize():
     assert new_company.environment.get_role(arch.profile) is not None
 
 
-def test_team_serdeser():
+def test_team_serdeser_save():
     company = Team()
     company.hire([RoleC()])
 
@@ -60,12 +60,19 @@ async def test_team_recover():
     shutil.rmtree(stg_path, ignore_errors=True)
 
     company = Team()
-    company.hire([RoleC()])
+    role_c = RoleC()
+    company.hire([role_c])
     company.start_project(idea)
     await company.run(n_round=4)
 
     ser_data = company.dict()
     new_company = Team(**ser_data)
+
+    new_role_c = new_company.environment.get_role(role_c.profile)
+    assert new_role_c._rc.memory == role_c._rc.memory
+    assert new_role_c._rc.env != role_c._rc.env  # due to Action raise, role's memory has been changed.
+    assert new_role_c._rc.env.memory == role_c._rc.env.memory
+
     assert new_company.environment.memory.count() == 1
     assert type(list(new_company.environment.roles.values())[0]._actions[0]) == ActionOK
 
@@ -80,11 +87,17 @@ async def test_team_recover_save():
     shutil.rmtree(stg_path, ignore_errors=True)
 
     company = Team()
-    company.hire([RoleC()])
+    role_c = RoleC()
+    company.hire([role_c])
     company.start_project(idea)
     await company.run(n_round=4)
 
     new_company = Team.recover(stg_path)
+    new_role_c = new_company.environment.get_role(role_c.profile)
+    assert new_role_c._rc.memory == role_c._rc.memory
+    assert new_role_c._rc.env != role_c._rc.env  # due to Action raise, role's memory has been changed.
+    assert new_role_c._rc.env.memory == role_c._rc.env.memory
+
     new_company.start_project(idea)
     await new_company.run(n_round=4)