Litellm dev 01 25 2025 p2 (#8003)

* fix(base_utils.py): supported nested json schema passed in for anthropic calls

* refactor(base_utils.py): refactor ref parsing to prevent infinite loop

* test(test_openai_endpoints.py): refactor anthropic test to use bedrock

* fix(langfuse_prompt_management.py): add unit test for sync langfuse calls

Resolves https://github.com/BerriAI/litellm/issues/7938#issuecomment-2613293757

tests/llm_translation/base_llm_unit_tests.py

@@ -312,6 +312,56 @@ class BaseLLMChatTest(ABC):
         except litellm.InternalServerError:
             pytest.skip("Model is overloaded")
 
+    @pytest.mark.flaky(retries=6, delay=1)
+    def test_json_response_nested_json_schema(self):
+        """
+        PROD Test: ensure nested json schema sent to proxy works as expected.
+        """
+        from pydantic import BaseModel
+        from litellm.utils import supports_response_schema
+        from litellm.llms.base_llm.base_utils import type_to_response_format_param
+
+        os.environ["LITELLM_LOCAL_MODEL_COST_MAP"] = "True"
+        litellm.model_cost = litellm.get_model_cost_map(url="")
+
+        class CalendarEvent(BaseModel):
+            name: str
+            date: str
+            participants: list[str]
+
+        class EventsList(BaseModel):
+            events: list[CalendarEvent]
+
+        response_format = type_to_response_format_param(EventsList)
+
+        messages = [
+            {"role": "user", "content": "List 5 important events in the XIX century"}
+        ]
+
+        base_completion_call_args = self.get_base_completion_call_args()
+        if not supports_response_schema(base_completion_call_args["model"], None):
+            pytest.skip(
+                f"Model={base_completion_call_args['model']} does not support response schema"
+            )
+
+        try:
+            res = self.completion_function(
+                **base_completion_call_args,
+                messages=messages,
+                response_format=response_format,
+                timeout=60,
+            )
+            assert res is not None
+
+            print(res.choices[0].message)
+
+            assert res.choices[0].message.content is not None
+            assert res.choices[0].message.tool_calls is None
+        except litellm.Timeout:
+            pytest.skip("Model took too long to respond")
+        except litellm.InternalServerError:
+            pytest.skip("Model is overloaded")
+
     @pytest.mark.flaky(retries=6, delay=1)
     def test_json_response_format_stream(self):
         """