Add instructions for token limit parameters fro OpenAI models

Add instructions for token limit parameters fro OpenAI models

Author: paxiaatucsdedu

libs/oci/README.md

@@ -30,7 +30,14 @@ This repository includes two main integration categories:
 ```python
 from langchain_oci import ChatOCIGenAI
 
-llm = ChatOCIGenAI()
+llm = ChatOCIGenAI(
+        model_id="MY_MODEL_ID",
+        service_endpoint="MY_SERVICE_ENDPOINT",
+        compartment_id="MY_COMPARTMENT_ID",
+        model_kwargs={"max_tokens": 1024}, # Use max_completion_tokens instead of max_tokens for OpenAI models
+        auth_profile="MY_AUTH_PROFILE",
+        is_stream=True,
+        auth_type="SECURITY_TOKEN"
 llm.invoke("Sing a ballad of LangChain.")
 ```
 

libs/oci/langchain_oci/chat_models/oci_generative_ai.py

@@ -1058,6 +1058,15 @@ def _prepare_request(
         if stop is not None:
             _model_kwargs[self._provider.stop_sequence_key] = stop
 
+        # Warn if using max_tokens with OpenAI models
+        if self.model_id and self.model_id.startswith("openai.") and "max_tokens" in _model_kwargs:
+            import warnings
+            warnings.warn(
+                f"OpenAI models require 'max_completion_tokens' instead of 'max_tokens'.",
+                UserWarning,
+                stacklevel=2
+            )
+
         chat_params = {**_model_kwargs, **kwargs, **oci_params}
 
         if not self.model_id: