Fix some testing

src/c++/perf_analyzer/genai-pa/genai_pa/parser.py

@@ -160,6 +160,7 @@ def _add_profile_args(parser):
         "For example, if the profile export file is profile_export.json, the GenAi-PA file will be "
         "exported to profile_export_genai_pa.csv.",
     )
+
     load_management_group.add_argument(
         "--request-rate",
         type=float,
@@ -185,7 +186,7 @@ def _add_profile_args(parser):
         help="Indicates the allowed variation in "
         "latency measurements when determining if a result is stable. The "
         "measurement is considered as stable if the ratio of max / min "
-        "from the recent 3 measurements is within (stability percentage)% "
+        "from the recent 3 measurements is within (stability percentage) "
         "in terms of both infer per second and latency.",
     )
 

src/c++/perf_analyzer/genai-pa/tests/test_cli.py

@@ -59,15 +59,12 @@ def test_help_arguments_output_and_exit(self, arg, expected_output, capsys):
         [
             (["-b", "2"], {"batch_size": 2}),
             (["--batch-size", "2"], {"batch_size": 2}),
-            (["--max-threads", "4"], {"max_threads": 4}),
             (
                 ["--profile-export-file", "text.txt"],
                 {"profile_export_file": Path("text.txt")},
             ),
             (["--service-kind", "triton"], {"service_kind": "triton"}),
             (["--service-kind", "openai"], {"service_kind": "openai"}),
-            # TODO: Remove streaming from implementation. It is invalid with HTTP.
-            # (["--streaming"], {"streaming": True}),
             (["--version"], {"version": True}),
             (["-u", "test_url"], {"u": "test_url"}),
             (["--url", "test_url"], {"u": "test_url"}),

src/c++/perf_analyzer/genai-pa/tests/test_llm_metrics.py

@@ -199,7 +199,7 @@ def test_triton_llm_profile_data(self, prepare_triton_profile_data) -> None:
         )
 
         # experiment 1 statistics
-        stat = pd.get_statistics(infer_mode="concurrency", load_level=10)
+        stat = pd.get_statistics(infer_mode="concurrency", load_level="10")
 
         assert stat.avg_time_to_first_token == 2
         assert stat.avg_inter_token_latency == 2.25
@@ -227,7 +227,7 @@ def test_triton_llm_profile_data(self, prepare_triton_profile_data) -> None:
         assert stat.std_num_output_token == np.std([3, 5])
 
         # experiment 2 statistics
-        stat = pd.get_statistics(infer_mode="request_rate", load_level=2.0)
+        stat = pd.get_statistics(infer_mode="request_rate", load_level="2.0")
 
         assert stat.avg_time_to_first_token == 2.5
         assert stat.avg_inter_token_latency == 3
@@ -256,7 +256,7 @@ def test_triton_llm_profile_data(self, prepare_triton_profile_data) -> None:
 
         # check non-existing profile data
         with pytest.raises(KeyError):
-            pd.get_statistics(infer_mode="concurrency", load_level=30)
+            pd.get_statistics(infer_mode="concurrency", load_level="30")
 
     def test_openai_llm_profile_data(self, prepare_openai_profile_data) -> None:
         """Collect LLM metrics from profile export data and check values.
@@ -280,7 +280,7 @@ def test_openai_llm_profile_data(self, prepare_openai_profile_data) -> None:
         )
 
         # experiment 1 statistics
-        stat = pd.get_statistics(infer_mode="concurrency", load_level=10)
+        stat = pd.get_statistics(infer_mode="concurrency", load_level="10")
 
         assert stat.avg_time_to_first_token == 2
         assert stat.avg_inter_token_latency == 2.4
@@ -309,7 +309,7 @@ def test_openai_llm_profile_data(self, prepare_openai_profile_data) -> None:
 
         # check non-existing profile data
         with pytest.raises(KeyError):
-            pd.get_statistics(infer_mode="concurrency", load_level=40)
+            pd.get_statistics(infer_mode="concurrency", load_level="40")
 
     def test_llm_metrics_get_base_name(self) -> None:
         """Test get_base_name method in LLMMetrics class."""