[GuideLLM Refactor] Fixes for various failing unit tests (#394)

## Summary

<!--
Include a short paragraph of the changes introduced in this PR.
If this PR requires additional context or rationale, explain why
the changes are necessary.
-->
Fixes failing unit tests. Most were failing due to changes in
functionality but a couple were regressions.

---

- [x] "I certify that all code in this PR is my own, except as noted
below."

## Use of AI

- [x] Includes AI-assisted code completion
- [ ] Includes code generated by an AI application
- [x] Includes AI-generated tests (NOTE: AI written tests should have a
docstring that includes `## WRITTEN BY AI ##`)

Author: jaredoconnell

src/guidellm/benchmark/types.py

@@ -2,12 +2,13 @@
 
 from collections.abc import Iterable
 from pathlib import Path
-from typing import Any, TypeAliasType
+from typing import Any
 
 from datasets import Dataset, DatasetDict, IterableDataset, IterableDatasetDict
 from transformers import (  # type: ignore[import]
     PreTrainedTokenizerBase,
 )
+from typing_extensions import TypeAliasType
 
 from guidellm.benchmark.aggregator import (
     Aggregator,

src/guidellm/scheduler/objects.py

@@ -20,6 +20,7 @@
     Protocol,
     TypeVar,
     Union,
+    runtime_checkable,
 )
 
 from pydantic import Field, computed_field
@@ -232,6 +233,7 @@ def model_copy(self, **kwargs) -> ScheduledRequestInfo:  # type: ignore[override
         )
 
 
+@runtime_checkable
 class BackendInterface(Protocol, Generic[RequestT, ResponseT]):
     """
     Abstract interface for request processing backends.

src/guidellm/utils/pydantic_utils.py

@@ -275,6 +275,17 @@ class DatabaseConfig(BaseConfig):
 
     schema_discriminator: ClassVar[str] = "model_type"
 
+    def __new__(cls, *args, **kwargs):  # noqa: ARG004
+        """
+        Prevent direct instantiation of base classes that use this mixin.
+
+        Only allows instantiation of concrete subclasses, not the base class.
+        """
+        base_type = cls.__pydantic_schema_base_type__()
+        if cls is base_type:
+            raise TypeError(f"only children of '{cls.__name__}' may be instantiated")
+        return super().__new__(cls)
+
     @classmethod
     def register_decorator(
         cls, clazz: RegisterClassT, name: str | list[str] | None = None

src/guidellm/utils/statistics.py

@@ -275,18 +275,9 @@ def from_request_times(
         """
         if distribution_type == "concurrency":
             # convert to delta changes based on when requests were running
-            time_deltas: dict[float, int] = defaultdict(int)
-            for start, end in requests:
-                time_deltas[start] += 1
-                time_deltas[end] -= 1
-
-            # convert to the events over time measuring concurrency changes
-            events = []
-            active = 0
-
-            for time, delta in sorted(time_deltas.items()):
-                active += delta
-                events.append((time, active))
+            events = [(start, 1) for start, _ in requests] + [
+                (end, -1) for _, end in requests
+            ]
         elif distribution_type == "rate":
             # convert to events for when requests finished
             global_start = min(start for start, _ in requests) if requests else 0
@@ -313,6 +304,16 @@ def from_request_times(
             else:
                 flattened_events.append((time, val))
 
+        if distribution_type == "concurrency":
+            # convert to the events over time measuring concurrency changes
+            events_over_time: list[tuple[float, float]] = []
+            active = 0
+            for time, delta in flattened_events:
+                active += delta  # type: ignore [assignment]
+                events_over_time.append((time, active))
+
+            flattened_events = events_over_time
+
         # convert to value distribution function
         distribution: dict[float, float] = defaultdict(float)
 

tests/unit/backend/__init__.py renamed to tests/unit/backends/__init__.py

@@ -80,7 +80,7 @@ async def default_model(self) -> str | None:
     def test_class_signatures(self):
         """Test Backend inheritance and type relationships."""
         assert issubclass(Backend, RegistryMixin)
-        assert issubclass(Backend, BackendInterface)
+        assert isinstance(Backend, BackendInterface)
         assert hasattr(Backend, "create")
         assert hasattr(Backend, "register")
         assert hasattr(Backend, "get_registered_object")

tests/unit/backend/test_backend.py renamed to tests/unit/backends/test_backend.py

@@ -397,7 +397,6 @@ def valid_instances(self, request):
     def test_class_signatures(self):
         """Test GenerationRequestTimings inheritance and type relationships."""
         assert issubclass(GenerationRequestTimings, MeasuredRequestTimings)
-        assert issubclass(GenerationRequestTimings, StandardBaseModel)
         assert hasattr(GenerationRequestTimings, "model_dump")
         assert hasattr(GenerationRequestTimings, "model_validate")
 

tests/unit/backend/test_objects.py renamed to tests/unit/backends/test_objects.py

@@ -237,7 +237,7 @@ async def test_info(self):
             target="http://test", model="test-model", timeout=30.0
         )
 
-        info = backend.info()
+        info = backend.info
 
         assert info["target"] == "http://test"
         assert info["model"] == "test-model"
@@ -1074,7 +1074,7 @@ def test_get_chat_message_media_item_jpeg_file(self):
         mock_image = Mock(spec=Image.Image)
         mock_image.tobytes.return_value = b"fake_jpeg_data"
 
-        with patch("guidellm.backend.openai.Image.open", return_value=mock_image):
+        with patch("guidellm.backends.openai.Image.open", return_value=mock_image):
             result = backend._get_chat_message_media_item(mock_jpeg_path)
 
         expected_data = base64.b64encode(b"fake_jpeg_data").decode("utf-8")