Fix test output xfail ut

src/guidellm/benchmark/benchmarker.py

@@ -13,7 +13,7 @@
 import uuid
 from abc import ABC
 from collections.abc import AsyncIterator, Iterable
-from typing import Generic
+from typing import Any, Generic
 
 from guidellm.benchmark.profile import Profile
 from guidellm.benchmark.progress import BenchmarkerProgress

src/guidellm/data/deserializers/deserializer.py

@@ -107,8 +107,10 @@ def _deserialize_with_registered_deserializers(
 
         if len(errors) > 0:
             err_msgs = ""
+
             def sort_key(item):
                 return (isinstance(item[1], DataNotSupportedError), item[0])
+
             for key, err in sorted(errors.items(), key=sort_key):
                 err_msgs += f"\n  - Deserializer '{key}': ({type(err).__name__}) {err}"
             raise ValueError(
@@ -141,4 +143,3 @@ def _deserialize_with_specified_deserializer(
             random_seed=random_seed,
             **data_kwargs,
         )
-

src/guidellm/data/loaders.py

@@ -17,7 +17,6 @@
 __all__ = ["DataLoader", "DatasetsIterator"]
 
 
-
 class DatasetsIterator(TorchIterableDataset):
     def __init__(
         self,

src/guidellm/data/preprocessors/formatters.py

@@ -56,9 +56,7 @@ def __init__(
         self.stream: bool = stream
         self.max_tokens: int | None = max_tokens or max_completion_tokens
 
-    def __call__(
-        self, columns: dict[str, list[Any]]
-    ) -> GenerationRequest:
+    def __call__(self, columns: dict[str, list[Any]]) -> GenerationRequest:
         """
         :param columns: A dict of GenerativeDatasetColumnType to Any
         """
@@ -396,9 +394,7 @@ def __call__(  # noqa: C901
 class GenerativeAudioTranslationRequestFormatter(
     GenerativeAudioTranscriptionRequestFormatter
 ):
-    def __call__(
-        self, columns: dict[str, list[Any]]
-    ) -> GenerationRequest:
+    def __call__(self, columns: dict[str, list[Any]]) -> GenerationRequest:
         result = super().__call__(columns)
         result.request_type = "audio_translations"
         return result

src/guidellm/data/preprocessors/mappers.py

@@ -167,9 +167,7 @@ def __init__(
             dict[GenerativeDatasetColumnType, list[tuple[int, str]]] | None
         )
 
-    def __call__(
-        self, row: dict[str, Any]
-    ) -> dict[str, list[Any]]:
+    def __call__(self, row: dict[str, Any]) -> dict[str, list[Any]]:
         if self.datasets_column_mappings is None:
             raise ValueError("DefaultGenerativeColumnMapper not setup with data.")
 

src/guidellm/data/preprocessors/preprocessor.py

@@ -12,8 +12,7 @@
 
 @runtime_checkable
 class DatasetPreprocessor(Protocol):
-    def __call__(self, item: dict[str, Any]) -> (
-        GenerationRequest | dict[str, Any]): ...
+    def __call__(self, item: dict[str, Any]) -> GenerationRequest | dict[str, Any]: ...
 
 
 @runtime_checkable

src/guidellm/preprocess/dataset.py

@@ -238,7 +238,7 @@ def process_dataset(
     prompt_tokens: str | Path,
     output_tokens: str | Path,
     processor_args: dict[str, Any] | None = None,
-    data_args: dict[str, Any] | None = None,
+    data_args: dict[str, Any] | None = None,  # noqa: ARG001
     short_prompt_strategy: ShortPromptStrategy = ShortPromptStrategy.IGNORE,
     pad_char: str | None = None,
     concat_delimiter: str | None = None,

src/guidellm/scheduler/strategies.py

@@ -506,8 +506,10 @@ def init_processes_start(self, start_time: float):
         if self._processes_lock is None:
             raise RuntimeError("_processes_lock is None in init_processes_start")
         if self._offset is None:
-            raise RuntimeError("_offset is None in init_processes_start; was "
-                               "init_processes_timings not called?")
+            raise RuntimeError(
+                "_offset is None in init_processes_start; was "
+                "init_processes_timings not called?"
+            )
         with self._processes_lock:
             self._offset.value = start_time
 
@@ -527,11 +529,15 @@ async def next_request_time(self, offset: int) -> float:
         next_delay = self._random.expovariate(self.rate)
 
         if self._processes_lock is None:
-            raise RuntimeError("_processes_lock is None in next_request_time; was "
-                               "init_processes_timings not called?")
+            raise RuntimeError(
+                "_processes_lock is None in next_request_time; was "
+                "init_processes_timings not called?"
+            )
         if self._offset is None:
-            raise RuntimeError("_offset is None in next_request_time; was "
-                               "init_processes_timings not called?")
+            raise RuntimeError(
+                "_offset is None in next_request_time; was "
+                "init_processes_timings not called?"
+            )
         with self._processes_lock:
             self._offset.value += next_delay
 

src/guidellm/scheduler/worker.py

@@ -363,7 +363,6 @@ async def _process_next_request(self, target_start: float):
             async for resp, info in self.backend.resolve(  # type: ignore[attr-defined]
                 request, request_info, None
             ):
-
                 response = resp
                 request_info = info
                 if request_info is None:
@@ -407,10 +406,7 @@ async def _dequeue_next_request(
         return request, request_info
 
     async def _schedule_request(
-        self,
-        request: RequestT,
-        request_info: RequestInfo,
-        target_start: float
+        self, request: RequestT, request_info: RequestInfo, target_start: float
     ):
         current_time = time.time()
         request_info.timings.scheduled_at = current_time

src/guidellm/utils/cli.py

@@ -31,6 +31,7 @@ def parse_list_floats(ctx, param, value):  # noqa: ARG001
             f"of floats/ints. Error: {e}"
         ) from e
 
+
 def parse_json(ctx, param, value):  # noqa: ARG001
     if value is None or value == [None]:
         return None

tests/unit/benchmark/test_output.py

@@ -91,7 +91,6 @@ def test_file_yaml():
     mock_path.unlink()
 
 
-@pytest.mark.xfail(reason="old and broken", run=False)
 @pytest.mark.asyncio
 async def test_file_csv():
     args = BenchmarkGenerativeTextArgs(target="http://localhost:8000", data=["test"])

tests/unit/mock_benchmark.py

@@ -105,9 +105,7 @@ def mock_generative_benchmark() -> GenerativeBenchmark:
         benchmarker=BenchmarkerDict(
             profile=SynchronousProfile.create("synchronous", rate=None),
             requests={
-                "attributes": {
-                    "data": "prompt_tokens=256,output_tokens=128",
-                },
+                "data": "prompt_tokens=256,output_tokens=128",
             },
             backend={},
             environment={},