use settings for some defaults, avoid warnings

src/marvin/_mappings/chat_completion.py

@@ -6,7 +6,6 @@
     Callable,
     TypeVar,
     Union,
-    cast,
 )
 
 from pydantic import BaseModel, TypeAdapter, ValidationError
@@ -41,9 +40,6 @@ def chat_completion_to_model(
         data: dict[str, Any] = {}
         data[field_name] = json.loads(tool_arguments[0])
         return response_model.model_validate_json(json.dumps(data))
-    else:
-        data: dict[str, Any] = json.loads(tool_arguments[0])
-        return cast(T, data)
 
 
 def chat_completion_to_type(response_type: U, completion: "ChatCompletion") -> "U":

src/marvin/_mappings/types.py

@@ -26,6 +26,7 @@ def cast_type_to_model(
 
     return create_model(
         model_name,
+        __doc__=model_description,
         __config__=None,
         __base__=None,
         __module__=__name__,

src/marvin/components/ai_function.py

@@ -17,10 +17,12 @@
 from pydantic import BaseModel, ConfigDict, Field
 from typing_extensions import NotRequired, ParamSpec, Self, Unpack
 
+import marvin
 from marvin._mappings.chat_completion import chat_completion_to_model
 from marvin.client.openai import AsyncMarvinClient, MarvinClient
 from marvin.components.prompt.fn import PromptFunction
 from marvin.utilities.jinja import BaseEnvironment
+from marvin.utilities.logging import get_logger
 
 if TYPE_CHECKING:
     from openai.types.chat import ChatCompletion
@@ -44,17 +46,17 @@ class AIFunctionKwargs(TypedDict):
 
 
 class AIFunctionKwargsDefaults(BaseModel):
-    model_config = ConfigDict(arbitrary_types_allowed=True)
+    model_config = ConfigDict(arbitrary_types_allowed=True, protected_namespaces=())
     environment: Optional[BaseEnvironment] = None
     prompt: Optional[str] = None
     model_name: str = "FormatResponse"
     model_description: str = "Formats the response."
     field_name: str = "data"
     field_description: str = "The data to format."
-    model: Optional[str] = None
+    model: str = marvin.settings.openai.chat.completions.model
     client: Optional[Client] = None
     aclient: Optional[AsyncClient] = None
-    temperature: Optional[float] = None
+    temperature: Optional[float] = marvin.settings.openai.chat.completions.temperature
 
 
 class AIFunction(
@@ -89,6 +91,10 @@ class AIFunction(
     client: Client = Field(default_factory=lambda: MarvinClient().client)
     aclient: AsyncClient = Field(default_factory=lambda: AsyncMarvinClient().client)
 
+    @property
+    def logger(self):
+        return get_logger(self.__class__.__name__)
+
     def __call__(
         self, *args: P.args, **kwargs: P.kwargs
     ) -> Union[T, Coroutine[Any, Any, T]]:
@@ -101,6 +107,7 @@ def call(self, *args: P.args, **kwargs: P.kwargs) -> T:
         response: ChatCompletion = MarvinClient(client=self.client).chat(
             **prompt.serialize()
         )
+        self.logger.debug_kv("Calling", f"{self.fn.__name__}({args}, {kwargs})", "blue")
         return getattr(
             chat_completion_to_model(model, response, field_name=self.field_name),
             self.field_name,

src/marvin/settings.py

@@ -22,6 +22,7 @@
 
 class MarvinSettings(BaseSettings):
     def __setattr__(self, name: str, value: Any) -> None:
+        # wrap bare strings in SecretStr if the field is annotated with SecretStr
         field = self.model_fields.get(name)
         if field:
             annotation = field.annotation
@@ -46,6 +47,10 @@ class ChatCompletionSettings(MarvinSettings):
         description="The default chat model to use.", default="gpt-3.5-turbo"
     )
 
+    temperature: float = Field(
+        description="The default temperature to use.", default=0.1
+    )
+
     @property
     def encoder(self):
         import tiktoken