add alias for model (#4553)

Co-authored-by: Dev 2049 <dev.dev2049@gmail.com>

langchain/base_language.py

@@ -2,7 +2,7 @@
 from __future__ import annotations
 
 from abc import ABC, abstractmethod
-from typing import List, Optional, Sequence
+from typing import List, Optional, Sequence, Set
 
 from pydantic import BaseModel
 
@@ -68,3 +68,12 @@ class BaseLanguageModel(BaseModel, ABC):
     def get_num_tokens_from_messages(self, messages: List[BaseMessage]) -> int:
         """Get the number of tokens in the message."""
         return sum([self.get_num_tokens(get_buffer_string([m])) for m in messages])
+
+    @classmethod
+    def all_required_field_names(cls) -> Set:
+        all_required_field_names = set()
+        for field in cls.__fields__.values():
+            all_required_field_names.add(field.name)
+            if field.has_alias:
+                all_required_field_names.add(field.alias)
+        return all_required_field_names

langchain/chat_models/openai.py

@@ -112,7 +112,7 @@ class ChatOpenAI(BaseChatModel):
     """
 
     client: Any  #: :meta private:
-    model_name: str = "gpt-3.5-turbo"
+    model_name: str = Field(default="gpt-3.5-turbo", alias="model")
     """Model name to use."""
     temperature: float = 0.7
     """What sampling temperature to use."""
@@ -138,12 +138,12 @@ class ChatOpenAI(BaseChatModel):
         """Configuration for this pydantic object."""
 
         extra = Extra.ignore
+        allow_population_by_field_name = True
 
     @root_validator(pre=True)
     def build_extra(cls, values: Dict[str, Any]) -> Dict[str, Any]:
         """Build extra kwargs from additional params that were passed in."""
-        all_required_field_names = {field.alias for field in cls.__fields__.values()}
+        all_required_field_names = cls.all_required_field_names()
-
         extra = values.get("model_kwargs", {})
         for field_name in list(values):
             if field_name in extra:
@@ -156,8 +156,7 @@ class ChatOpenAI(BaseChatModel):
                 )
                 extra[field_name] = values.pop(field_name)
 
-        disallowed_model_kwargs = all_required_field_names | {"model"}
+        invalid_model_kwargs = all_required_field_names.intersection(extra.keys())
-        invalid_model_kwargs = disallowed_model_kwargs.intersection(extra.keys())
         if invalid_model_kwargs:
             raise ValueError(
                 f"Parameters {invalid_model_kwargs} should be specified explicitly. "

langchain/llms/openai.py

@@ -124,7 +124,7 @@ class BaseOpenAI(BaseLLM):
     """Wrapper around OpenAI large language models."""
 
     client: Any  #: :meta private:
-    model_name: str = "text-davinci-003"
+    model_name: str = Field("text-davinci-003", alias="model")
     """Model name to use."""
     temperature: float = 0.7
     """What sampling temperature to use."""
@@ -178,12 +178,12 @@ class BaseOpenAI(BaseLLM):
         """Configuration for this pydantic object."""
 
         extra = Extra.ignore
+        allow_population_by_field_name = True
 
     @root_validator(pre=True)
     def build_extra(cls, values: Dict[str, Any]) -> Dict[str, Any]:
         """Build extra kwargs from additional params that were passed in."""
-        all_required_field_names = {field.alias for field in cls.__fields__.values()}
+        all_required_field_names = cls.all_required_field_names()
-
         extra = values.get("model_kwargs", {})
         for field_name in list(values):
             if field_name in extra:
@@ -196,8 +196,7 @@ class BaseOpenAI(BaseLLM):
                 )
                 extra[field_name] = values.pop(field_name)
 
-        disallowed_model_kwargs = all_required_field_names | {"model"}
+        invalid_model_kwargs = all_required_field_names.intersection(extra.keys())
-        invalid_model_kwargs = disallowed_model_kwargs.intersection(extra.keys())
         if invalid_model_kwargs:
             raise ValueError(
                 f"Parameters {invalid_model_kwargs} should be specified explicitly. "

tests/integration_tests/chat_models/test_openai.py

@@ -25,6 +25,14 @@ def test_chat_openai() -> None:
     assert isinstance(response.content, str)
 
 
+def test_chat_openai_model() -> None:
+    """Test ChatOpenAI wrapper handles model_name."""
+    chat = ChatOpenAI(model="foo")
+    assert chat.model_name == "foo"
+    chat = ChatOpenAI(model_name="bar")
+    assert chat.model_name == "bar"
+
+
 def test_chat_openai_system_message() -> None:
     """Test ChatOpenAI wrapper with system message."""
     chat = ChatOpenAI(max_tokens=10)

tests/integration_tests/llms/test_openai.py

@@ -19,6 +19,13 @@ def test_openai_call() -> None:
     assert isinstance(output, str)
 
 
+def test_openai_model_param() -> None:
+    llm = OpenAI(model="foo")
+    assert llm.model_name == "foo"
+    llm = OpenAI(model_name="foo")
+    assert llm.model_name == "foo"
+
+
 def test_openai_extra_kwargs() -> None:
     """Test extra kwargs to openai."""
     # Check that foo is saved in extra_kwargs.