x

libs/partners/together/tests/integration_tests/test_chat_models_standard.py

@@ -4,12 +4,17 @@
 
 import pytest
 from langchain_core.language_models import BaseChatModel
-from langchain_standard_tests.integration_tests import (  # type: ignore[import-not-found]
-    ChatModelIntegrationTests,  # type: ignore[import-not-found]
-)
+from langchain_core.rate_limiters import InMemoryRateLimiter
+from langchain_standard_tests.integration_tests import ChatModelIntegrationTests
 
 from langchain_together import ChatTogether
 
+# Initialize the rate limiter in global scope, so it can be re-used
+# across tests.
+rate_limiter = InMemoryRateLimiter(
+    requests_per_second=0.5,
+)
+
 
 class TestTogetherStandard(ChatModelIntegrationTests):
     @property
@@ -18,7 +23,10 @@ def chat_model_class(self) -> Type[BaseChatModel]:
 
     @property
     def chat_model_params(self) -> dict:
-        return {"model": "mistralai/Mistral-7B-Instruct-v0.1"}
+        return {
+            "model": "mistralai/Mistral-7B-Instruct-v0.1",
+            "rate_limiter": rate_limiter,
+        }
 
     @pytest.mark.xfail(reason=("May not call a tool."))
     def test_tool_calling_with_no_arguments(self, model: BaseChatModel) -> None: