Fix tests

Author: markVaykhansky

src/guidellm/benchmark/output.py

@@ -419,6 +419,7 @@ def benchmarks_args_str(self) -> str:
             {
                 "max_number": args.max_number,
                 "max_duration": args.max_duration,
+                "max_error_rate": args.max_error_rate,
                 "warmup_number": args.warmup_number,
                 "warmup_duration": args.warmup_duration,
                 "cooldown_number": args.cooldown_number,

src/guidellm/scheduler/scheduler.py

@@ -315,13 +315,15 @@ def _determine_total_requests_count(
             iter_length = len(self.request_loader)  # type: ignore[arg-type]
             if 0 < iter_length < end_number:
                 end_number = iter_length
-        except InfiniteDatasetError:  # noqa: BLE001, S110
+        except InfiniteDatasetError:
+            # Only when RPS is constant and duration is capped we can determine the total
+            # amount of requests that are supposed to be sent
             if scheduling_strategy.type_ == "constant" and max_duration is not None:
                 total_requests_in_max_duration = int(scheduling_strategy.rate * max_duration)
                 if total_requests_in_max_duration < end_number:
                     assert total_requests_in_max_duration > 0
                     end_number = total_requests_in_max_duration
-        except Exception:
+        except Exception:  # noqa: BLE001, S110
             pass
         return end_number
 

tests/unit/benchmark/test_output.py

@@ -113,7 +113,7 @@ def test_console_benchmarks_args_str():
     mock_benchmark = mock_generative_benchmark()
     console.benchmarks = [mock_benchmark]
     assert console.benchmarks_args_str == (
-        "max_number=None, max_duration=10.0, warmup_number=None, "
+        "max_number=None, max_duration=10.0, max_error_rate=0.05, warmup_number=None, "
         "warmup_duration=None, cooldown_number=None, cooldown_duration=None"
     )
 

tests/unit/mock_benchmark.py

@@ -221,6 +221,7 @@ def mock_generative_benchmark() -> GenerativeBenchmark:
             strategy=SynchronousStrategy(),
             max_number=None,
             max_duration=10.0,
+            max_error_rate=0.05,
             warmup_number=None,
             warmup_duration=None,
             cooldown_number=None,
@@ -245,6 +246,7 @@ def mock_generative_benchmark() -> GenerativeBenchmark:
             request_start_time_targeted_delay_avg=1.2827096836907523,
             request_time_delay_avg=0.0004316908972603934,
             request_time_avg=1.426228676523481,
+            error_rate=0.345346,
         ),
         worker=GenerativeRequestsWorkerDescription(
             backend_type="openai_http",