fix(tests): fix repeating chunk and audio usage streaming tests (#23061)

- Replace ModelResponse(stream=True) with ModelResponseStream in
  test_unit_test_custom_stream_wrapper_repeating_chunk — stream=True
  stores delta as a plain dict causing AttributeError in CustomStreamWrapper
- Accept MidStreamFallbackError alongside InternalServerError in the
  repeating-chunk safety check assertion
- Add @pytest.mark.flaky(retries=3) to the live OpenAI audio output
  usage test

tests/local_testing/test_stream_chunk_builder.py

@@ -636,6 +636,7 @@ def test_stream_chunk_builder_openai_prompt_caching():
             assert response_usage_value == v
 
 
+@pytest.mark.flaky(retries=3, delay=2)
 def test_stream_chunk_builder_openai_audio_output_usage():
     from pydantic import BaseModel
     from openai import OpenAI

tests/local_testing/test_streaming.py

@@ -3075,22 +3075,18 @@ def test_unit_test_custom_stream_wrapper_repeating_chunk(
     """
     litellm.set_verbose = False
     chunks = [
-        litellm.ModelResponse(
-            **{
-                "id": "chatcmpl-123",
-                "object": "chat.completion.chunk",
-                "created": 1694268190,
-                "model": "gpt-3.5-turbo-0125",
-                "system_fingerprint": "fp_44709d6fcb",
-                "choices": [
-                    {
-                        "index": 0,
-                        "delta": {"content": chunk_value},
-                        "finish_reason": "stop",
-                    }
-                ],
-            },
-            stream=True,
+        litellm.ModelResponseStream(
+            id="chatcmpl-123",
+            created=1694268190,
+            model="gpt-3.5-turbo-0125",
+            system_fingerprint="fp_44709d6fcb",
+            choices=[
+                {
+                    "index": 0,
+                    "delta": {"content": chunk_value},
+                    "finish_reason": "stop",
+                }
+            ],
         )
     ] * loop_amount
     completion_stream = ModelResponseListIterator(model_responses=chunks)
@@ -3113,7 +3109,7 @@ def test_unit_test_custom_stream_wrapper_repeating_chunk(
     print(f"expected_chunk_fail: {expected_chunk_fail}")
 
     if (loop_amount > litellm.REPEATED_STREAMING_CHUNK_LIMIT) and expected_chunk_fail:
-        with pytest.raises(litellm.InternalServerError):
+        with pytest.raises((litellm.InternalServerError, litellm.exceptions.MidStreamFallbackError)):
             for chunk in response:
                 continue
     else: