feat: create HTTP DELETE API endpoints to unregister ScoringFn and Benchmark resources in Llama Stack (#3371)

# What does this PR do?
<!-- Provide a short summary of what this PR does and why. Link to
relevant issues if applicable. -->
This PR provides functionality for users to unregister ScoringFn and
Benchmark resources for `scoring` and `eval` APIs.

<!-- If resolving an issue, uncomment and update the line below -->
<!-- Closes #[issue-number] -->
Closes #3051 

## Test Plan
<!-- Describe the tests you ran to verify your changes with result
summaries. *Provide clear instructions so the plan can be easily
re-executed.* -->
Updated integration and unit tests via CI workflow

tests/unit/distribution/routers/test_routing_tables.py

@@ -105,6 +105,9 @@ class ScoringFunctionsImpl(Impl):
     async def register_scoring_function(self, scoring_fn):
         return scoring_fn
 
+    async def unregister_scoring_function(self, scoring_fn_id: str):
+        return scoring_fn_id
+
 
 class BenchmarksImpl(Impl):
     def __init__(self):
@@ -113,6 +116,9 @@ class BenchmarksImpl(Impl):
     async def register_benchmark(self, benchmark):
         return benchmark
 
+    async def unregister_benchmark(self, benchmark_id: str):
+        return benchmark_id
+
 
 class ToolGroupsImpl(Impl):
     def __init__(self):
@@ -330,6 +336,13 @@ async def test_scoring_functions_routing_table(cached_disk_dist_registry):
     assert "test-scoring-fn" in scoring_fn_ids
     assert "test-scoring-fn-2" in scoring_fn_ids
 
+    # Unregister scoring functions and verify listing
+    for i in range(len(scoring_functions.data)):
+        await table.unregister_scoring_function(scoring_functions.data[i].scoring_fn_id)
+
+    scoring_functions_list_after_deletion = await table.list_scoring_functions()
+    assert len(scoring_functions_list_after_deletion.data) == 0
+
 
 async def test_benchmarks_routing_table(cached_disk_dist_registry):
     table = BenchmarksRoutingTable({"test_provider": BenchmarksImpl()}, cached_disk_dist_registry, {})
@@ -347,6 +360,15 @@ async def test_benchmarks_routing_table(cached_disk_dist_registry):
     benchmark_ids = {b.identifier for b in benchmarks.data}
     assert "test-benchmark" in benchmark_ids
 
+    # Unregister the benchmark and verify removal
+    await table.unregister_benchmark(benchmark_id="test-benchmark")
+    benchmarks_after = await table.list_benchmarks()
+    assert len(benchmarks_after.data) == 0
+
+    # Unregistering a non-existent benchmark should raise a clear error
+    with pytest.raises(ValueError, match="Benchmark 'dummy_benchmark' not found"):
+        await table.unregister_benchmark(benchmark_id="dummy_benchmark")
+
 
 async def test_tool_groups_routing_table(cached_disk_dist_registry):
     table = ToolGroupsRoutingTable({"test_provider": ToolGroupsImpl()}, cached_disk_dist_registry, {})

tests/unit/providers/nvidia/test_eval.py

@@ -52,14 +52,19 @@ class TestNVIDIAEvalImpl(unittest.TestCase):
         self.evaluator_post_patcher = patch(
             "llama_stack.providers.remote.eval.nvidia.eval.NVIDIAEvalImpl._evaluator_post"
         )
+        self.evaluator_delete_patcher = patch(
+            "llama_stack.providers.remote.eval.nvidia.eval.NVIDIAEvalImpl._evaluator_delete"
+        )
 
         self.mock_evaluator_get = self.evaluator_get_patcher.start()
         self.mock_evaluator_post = self.evaluator_post_patcher.start()
+        self.mock_evaluator_delete = self.evaluator_delete_patcher.start()
 
     def tearDown(self):
         """Clean up after each test."""
         self.evaluator_get_patcher.stop()
         self.evaluator_post_patcher.stop()
+        self.evaluator_delete_patcher.stop()
 
     def _assert_request_body(self, expected_json):
         """Helper method to verify request body in Evaluator POST request is correct"""
@@ -115,6 +120,13 @@ class TestNVIDIAEvalImpl(unittest.TestCase):
         self.mock_evaluator_post.assert_called_once()
         self._assert_request_body({"namespace": benchmark.provider_id, "name": benchmark.identifier, **eval_config})
 
+    def test_unregister_benchmark(self):
+        # Unregister the benchmark
+        self.run_async(self.eval_impl.unregister_benchmark(benchmark_id=MOCK_BENCHMARK_ID))
+
+        # Verify the Evaluator API was called correctly
+        self.mock_evaluator_delete.assert_called_once_with(f"/v1/evaluation/configs/nvidia/{MOCK_BENCHMARK_ID}")
+
     def test_run_eval(self):
         benchmark_config = BenchmarkConfig(
             eval_candidate=ModelCandidate(