fix bug

metagpt/actions/write_code.py

@@ -132,7 +132,7 @@ class WriteCode(Action):
         code = await self.write_code(prompt)
         if not coding_context.code_doc:
             # avoid root_path pydantic ValidationError if use WriteCode alone
-            root_path = self.i_context.src_workspace if self.i_context.src_workspace else ""
+            root_path = self.context.src_workspace if self.context.src_workspace else ""
             coding_context.code_doc = Document(filename=coding_context.filename, root_path=str(root_path))
         coding_context.code_doc.content = code
         return coding_context

metagpt/config2.py

@@ -121,12 +121,10 @@ class Config(CLIParams, YamlModel):
             return llm[0]
         return None
 
-    def get_llm_config(self, name: Optional[str] = None, provider: LLMType = LLMType.OPENAI) -> LLMConfig:
+    def get_llm_config(self, name: Optional[str] = None, provider: LLMType = None) -> LLMConfig:
         """Return a LLMConfig instance"""
         if provider:
             llm_configs = self.get_llm_configs_by_type(provider)
-            if name:
-                llm_configs = [c for c in llm_configs if c.name == name]
 
             if len(llm_configs) == 0:
                 raise ValueError(f"Cannot find llm config with name {name} and provider {provider}")

metagpt/context.py

@@ -77,7 +77,7 @@ class Context(BaseModel):
     #     self._llm = None
     #     return self._llm
 
-    def llm(self, name: Optional[str] = None, provider: LLMType = LLMType.OPENAI) -> BaseLLM:
+    def llm(self, name: Optional[str] = None, provider: LLMType = None) -> BaseLLM:
         """Return a LLM instance, fixme: support cache"""
         # if self._llm is None:
         self._llm = create_llm_instance(self.config.get_llm_config(name, provider))
@@ -85,6 +85,14 @@ class Context(BaseModel):
             self._llm.cost_manager = self.cost_manager
         return self._llm
 
+    def llm_with_cost_manager_from_llm_config(self, llm_config: LLMConfig) -> BaseLLM:
+        """Return a LLM instance, fixme: support cache"""
+        # if self._llm is None:
+        llm = create_llm_instance(llm_config)
+        if llm.cost_manager is None:
+            llm.cost_manager = self.cost_manager
+        return llm
+
 
 class ContextMixin(BaseModel):
     """Mixin class for context and config"""
@@ -132,7 +140,7 @@ class ContextMixin(BaseModel):
         """Set llm"""
         self.set("_llm", llm, override)
 
-    def use_llm(self, name: Optional[str] = None, provider: LLMType = LLMType.OPENAI) -> BaseLLM:
+    def use_llm(self, name: Optional[str] = None, provider: LLMType = None) -> BaseLLM:
         """Use a LLM instance"""
         self._llm_config = self.config.get_llm_config(name, provider)
         self._llm = None
@@ -165,9 +173,9 @@ class ContextMixin(BaseModel):
     @property
     def llm(self) -> BaseLLM:
         """Role llm: role llm > context llm"""
-        # print(f"class:{self.__class__.__name__}({self.name}), llm: {self._llm}, llm_config: {self._llm_config}")
+        print(f"class:{self.__class__.__name__}({self.name}), llm: {self._llm}, llm_config: {self._llm_config}")
         if self._llm_config and not self._llm:
-            self._llm = self.context.llm(self._llm_config.name, self._llm_config.provider)
+            self._llm = self.context.llm_with_cost_manager_from_llm_config(self._llm_config)
         return self._llm or self.context.llm()
 
     @llm.setter

tests/metagpt/actions/test_write_code.py

@@ -19,8 +19,8 @@ from metagpt.const import (
     TEST_OUTPUTS_FILE_REPO,
 )
 from metagpt.context import CONTEXT
+from metagpt.llm import LLM
 from metagpt.logs import logger
-from metagpt.provider.openai_api import OpenAILLM as LLM
 from metagpt.schema import CodingContext, Document
 from metagpt.utils.common import aread
 from tests.metagpt.actions.mock_markdown import TASKS_2, WRITE_CODE_PROMPT_SAMPLE
@@ -32,7 +32,7 @@ async def test_write_code():
         filename="task_filename.py", design_doc=Document(content="设计一个名为'add'的函数，该函数接受两个整数作为输入，并返回它们的和。")
     )
     doc = Document(content=ccontext.model_dump_json())
-    write_code = WriteCode(context=doc)
+    write_code = WriteCode(i_context=doc)
 
     code = await write_code.run()
     logger.info(code.model_dump_json())
@@ -86,7 +86,7 @@ async def test_write_code_deps():
     )
     coding_doc = Document(root_path="snake1", filename="game.py", content=ccontext.json())
 
-    action = WriteCode(context=coding_doc)
+    action = WriteCode(i_context=coding_doc)
     rsp = await action.run()
     assert rsp
     assert rsp.code_doc.content