chore: cleanup docs, extra model_config declaration (#1033)

Fixes #988

Remove `tunedModels/` prefix since it is no longer supported in GenAI

Ensure `models/` prefix for LLM and Embedding models (whilst ensuring it
is stripped for LangSmith meta)

Author: mdrxy

libs/community/langchain_google_community/bq_storage_vectorstores/_base.py

@@ -79,10 +79,6 @@ class BaseBigQueryVectorStore(VectorStore, BaseModel, ABC):
     _logger: Any = None
     _full_table_id: Optional[str] = None
 
-    model_config = ConfigDict(
-        arbitrary_types_allowed=True,
-    )
-
     @abstractmethod
     def sync_data(self) -> None:
         ...

libs/genai/langchain_google_genai/__init__.py

@@ -22,7 +22,7 @@
 .. code-block:: python
     pip install -U langchain-google-genai
 
-## Using Chat Models
+**Using Chat Models**
 
 After setting up your environment with the required API key, you can interact with the Google Gemini models.
 
@@ -33,7 +33,7 @@
     llm = ChatGoogleGenerativeAI(model="gemini-pro")
     llm.invoke("Sing a ballad of LangChain.")
 
-## Using LLMs
+**Using LLMs**
 
 The package also supports generating text with Google's models.
 
@@ -44,7 +44,7 @@
     llm = GoogleGenerativeAI(model="gemini-pro")
     llm.invoke("Once upon a time, a library called LangChain")
 
-## Embedding Generation
+**Embedding Generation**
 
 The package also supports creating embeddings with Google's models, useful for textual similarity and other NLP applications.
 

libs/genai/langchain_google_genai/chat_models.py

@@ -1232,9 +1232,7 @@ def validate_environment(self) -> Self:
         if self.top_k is not None and self.top_k <= 0:
             raise ValueError("top_k must be positive")
 
-        if not any(
-            self.model.startswith(prefix) for prefix in ("models/", "tunedModels/")
-        ):
+        if not any(self.model.startswith(prefix) for prefix in ("models/",)):
             self.model = f"models/{self.model}"
 
         additional_headers = self.additional_headers or {}

libs/genai/langchain_google_genai/embeddings.py

@@ -97,6 +97,9 @@ def validate_environment(self) -> Self:
             google_api_key = self.google_api_key
         client_info = get_client_info("GoogleGenerativeAIEmbeddings")
 
+        if not any(self.model.startswith(prefix) for prefix in ("models/",)):
+            self.model = f"models/{self.model}"
+
         self.client = build_generative_service(
             credentials=self.credentials,
             api_key=google_api_key,

libs/genai/langchain_google_genai/llms.py

@@ -63,6 +63,9 @@ def __init__(self, **kwargs: Any) -> None:
     def validate_environment(self) -> Self:
         """Validates params and passes them to google-generativeai package."""
 
+        if not any(self.model.startswith(prefix) for prefix in ("models/",)):
+            self.model = f"models/{self.model}"
+
         self.client = ChatGoogleGenerativeAI(
             api_key=self.google_api_key,
             credentials=self.credentials,
@@ -86,6 +89,15 @@ def _get_ls_params(
         """Get standard params for tracing."""
         ls_params = super()._get_ls_params(stop=stop, **kwargs)
         ls_params["ls_provider"] = "google_genai"
+
+        models_prefix = "models/"
+        ls_model_name = (
+            self.model[len(models_prefix) :]
+            if self.model and self.model.startswith(models_prefix)
+            else self.model
+        )
+        ls_params["ls_model_name"] = ls_model_name
+
         if ls_max_tokens := kwargs.get("max_output_tokens", self.max_output_tokens):
             ls_params["ls_max_tokens"] = ls_max_tokens
         return ls_params

libs/genai/tests/unit_tests/test_llms.py

@@ -38,9 +38,9 @@ def test_tracing_params() -> None:
                 "HARM_CATEGORY_DANGEROUS_CONTENT": "BLOCK_LOW_AND_ABOVE"
             },  # Invalid arg
         )
-        assert llm.model == "gemini-pro"
+        assert llm.model == "models/gemini-pro"
         ls_params = llm._get_ls_params()
-        assert ls_params["ls_model_name"] == "gemini-pro"
+        assert ls_params.get("ls_model_name") == "gemini-pro"
         mock_warning.assert_called_once()
         call_args = mock_warning.call_args[0][0]
         assert "Unexpected argument 'safety_setting'" in call_args