fix(evaluation): initialize sample_input in prompty format_llm_response (#47005)

Agent-Logs-Url: https://github.com/Azure/azure-sdk-for-python/sessions/40ee70eb-f614-4ef1-a0a5-472639f541a8

Co-authored-by: copilot-swe-agent[bot] <198982749+Copilot@users.noreply.github.com>
Co-authored-by: posaninagendra <2713981+posaninagendra@users.noreply.github.com>
Co-authored-by: Nagendra Posani <naposani@microsoft.com>

Author: Copilot

sdk/evaluation/azure-ai-evaluation/CHANGELOG.md

@@ -27,6 +27,7 @@
 - Fixed `_get_metric_result` prefix matching where shorter metric names (e.g., `xpia`) could match before longer, more-specific ones (e.g., `xpia_manipulated_content`). Now sorts by length descending for correct longest-prefix matching.
 - Fixed non-dict `_properties` values from evaluators causing downstream issues. Values that are not dicts are now logged and dropped gracefully.
 - Fixed filename length error in `_inline_image` by catching OSError/ValueError during local path resolution and fall back to returning a text chunk instead of throwing.
+- Fixed `format_llm_response` raising `UnboundLocalError` when `inputs` was not provided by ensuring `sample_input` is always initialized.
 
 ### Other Changes
 

sdk/evaluation/azure-ai-evaluation/azure/ai/evaluation/_legacy/prompty/_utils.py

@@ -609,6 +609,7 @@ async def format_stream(llm_response: AsyncStream[ChatCompletionChunk]) -> Async
         )
         sample_output = json.dumps(sample_output_list)
         input_str = f"{json.dumps(inputs)}" if inputs else ""
+        sample_input = ""
         if inputs and len(inputs) > 0:
             sample_input_json = []
             msg = ChatCompletionUserMessageParam(

sdk/evaluation/azure-ai-evaluation/tests/unittests/test_prompty_format_llm_response.py

@@ -0,0 +1,52 @@
+# ---------------------------------------------------------
+# Copyright (c) Microsoft Corporation. All rights reserved.
+# ---------------------------------------------------------
+
+import asyncio
+from types import SimpleNamespace
+
+import pytest
+
+from azure.ai.evaluation._legacy.prompty._utils import format_llm_response
+
+
+class _FakeResponse:
+    def __init__(self):
+        self.usage = SimpleNamespace(prompt_tokens=1, completion_tokens=2, total_tokens=3)
+        self.choices = [
+            SimpleNamespace(
+                finish_reason="stop",
+                message=SimpleNamespace(role="assistant", content="test-output"),
+            )
+        ]
+        self.model = "test-model"
+
+    def model_dump(self):
+        return {
+            "choices": [
+                {
+                    "message": {
+                        "role": "assistant",
+                        "content": "test-output",
+                    }
+                }
+            ]
+        }
+
+
+@pytest.mark.unittest
+def test_format_llm_response_with_no_inputs_sets_empty_sample_input():
+    response = _FakeResponse()
+
+    result = asyncio.run(
+        format_llm_response(
+            response=response,
+            is_first_choice=True,
+            response_format={"type": "text"},
+            outputs=None,
+            inputs=None,
+        )
+    )
+
+    assert result["llm_output"] == "test-output"
+    assert result["sample_input"] == ""