groq[patch]: update standard tests (#27744)

- Add xfail on integration test (fails [> 50% of the
time](https://github.com/langchain-ai/langchain/actions/workflows/scheduled_test.yml));
- Remove xfail on passing unit test.

libs/partners/groq/tests/integration_tests/test_standard.py

@@ -49,3 +49,12 @@ def test_tool_calling_with_no_arguments(self, model: BaseChatModel) -> None:
     )
     def test_tool_message_histories_string_content(self, model: BaseChatModel) -> None:
         super().test_tool_message_histories_string_content(model)
+
+    @pytest.mark.xfail(
+        reason=(
+            "Sometimes fails with 'Failed to call a function. "
+            "Please adjust your prompt.'"
+        )
+    )
+    def test_bind_runnables_as_tools(self, model: BaseChatModel) -> None:
+        super().test_bind_runnables_as_tools(model)

libs/partners/groq/tests/unit_tests/test_standard.py

@@ -2,7 +2,6 @@
 
 from typing import Type
 
-import pytest
 from langchain_core.language_models import BaseChatModel
 from langchain_standard_tests.unit_tests.chat_models import (
     ChatModelUnitTests,
@@ -15,7 +14,3 @@ class TestGroqStandard(ChatModelUnitTests):
     @property
     def chat_model_class(self) -> Type[BaseChatModel]:
         return ChatGroq
-
-    @pytest.mark.xfail(reason="Groq does not support tool_choice='any'")
-    def test_bind_tool_pydantic(self, model: BaseChatModel) -> None:
-        super().test_bind_tool_pydantic(model)