chore: default to pytest asyncio-mode=auto (#2730)

# What does this PR do?

previously, developers who ran `./scripts/unit-tests.sh` would get
`asyncio-mode=auto`, which meant `@pytest.mark.asyncio` and
`@pytest_asyncio.fixture` were redundent. developers who ran `pytest`
directly would get pytest's default (strict mode), would run into errors
leading them to add `@pytest.mark.asyncio` / `@pytest_asyncio.fixture`
to their code.

with this change -
- `asyncio_mode=auto` is included in `pyproject.toml` making behavior
consistent for all invocations of pytest
- removes all redundant `@pytest_asyncio.fixture` and
`@pytest.mark.asyncio`
 - for good measure, requires `pytest>=8.4` and `pytest-asyncio>=1.0`

## Test Plan

- `./scripts/unit-tests.sh`
- `uv run pytest tests/unit`

tests/unit/server/test_access_control.py

@@ -7,7 +7,6 @@
 from unittest.mock import MagicMock, Mock, patch
 
 import pytest
-import pytest_asyncio
 import yaml
 from pydantic import TypeAdapter, ValidationError
 
@@ -27,7 +26,7 @@ def _return_model(model):
     return model
 
 
-@pytest_asyncio.fixture
+@pytest.fixture
 async def test_setup(cached_disk_dist_registry):
     mock_inference = Mock()
     mock_inference.__provider_spec__ = MagicMock()
@@ -41,7 +40,6 @@ async def test_setup(cached_disk_dist_registry):
     yield cached_disk_dist_registry, routing_table
 
 
-@pytest.mark.asyncio
 @patch("llama_stack.distribution.routing_tables.common.get_authenticated_user")
 async def test_access_control_with_cache(mock_get_authenticated_user, test_setup):
     registry, routing_table = test_setup
@@ -106,7 +104,6 @@ async def test_access_control_with_cache(mock_get_authenticated_user, test_setup
         await routing_table.get_model("model-admin")
 
 
-@pytest.mark.asyncio
 @patch("llama_stack.distribution.routing_tables.common.get_authenticated_user")
 async def test_access_control_and_updates(mock_get_authenticated_user, test_setup):
     registry, routing_table = test_setup
@@ -145,7 +142,6 @@ async def test_access_control_and_updates(mock_get_authenticated_user, test_setu
     assert model.identifier == "model-updates"
 
 
-@pytest.mark.asyncio
 @patch("llama_stack.distribution.routing_tables.common.get_authenticated_user")
 async def test_access_control_empty_attributes(mock_get_authenticated_user, test_setup):
     registry, routing_table = test_setup
@@ -170,7 +166,6 @@ async def test_access_control_empty_attributes(mock_get_authenticated_user, test
     assert "model-empty-attrs" in model_ids
 
 
-@pytest.mark.asyncio
 @patch("llama_stack.distribution.routing_tables.common.get_authenticated_user")
 async def test_no_user_attributes(mock_get_authenticated_user, test_setup):
     registry, routing_table = test_setup
@@ -201,7 +196,6 @@ async def test_no_user_attributes(mock_get_authenticated_user, test_setup):
     assert all_models.data[0].identifier == "model-public-2"
 
 
-@pytest.mark.asyncio
 @patch("llama_stack.distribution.routing_tables.common.get_authenticated_user")
 async def test_automatic_access_attributes(mock_get_authenticated_user, test_setup):
     """Test that newly created resources inherit access attributes from their creator."""
@@ -246,7 +240,7 @@ async def test_automatic_access_attributes(mock_get_authenticated_user, test_set
     assert model.identifier == "auto-access-model"
 
 
-@pytest_asyncio.fixture
+@pytest.fixture
 async def test_setup_with_access_policy(cached_disk_dist_registry):
     mock_inference = Mock()
     mock_inference.__provider_spec__ = MagicMock()
@@ -281,7 +275,6 @@ async def test_setup_with_access_policy(cached_disk_dist_registry):
     yield routing_table
 
 
-@pytest.mark.asyncio
 @patch("llama_stack.distribution.routing_tables.common.get_authenticated_user")
 async def test_access_policy(mock_get_authenticated_user, test_setup_with_access_policy):
     routing_table = test_setup_with_access_policy

tests/unit/server/test_auth.py

@@ -202,7 +202,6 @@ def test_http_auth_request_payload(http_client, valid_api_key, mock_auth_endpoin
         assert "param2" in payload["request"]["params"]
 
 
-@pytest.mark.asyncio
 async def test_http_middleware_with_access_attributes(mock_http_middleware, mock_scope):
     """Test HTTP middleware behavior with access attributes"""
     middleware, mock_app = mock_http_middleware

tests/unit/server/test_resolver.py

@@ -9,7 +9,6 @@ import sys
 from typing import Any, Protocol
 from unittest.mock import AsyncMock, MagicMock
 
-import pytest
 from pydantic import BaseModel, Field
 
 from llama_stack.apis.inference import Inference
@@ -66,7 +65,6 @@ class SampleImpl:
         pass
 
 
-@pytest.mark.asyncio
 async def test_resolve_impls_basic():
     # Create a real provider spec
     provider_spec = InlineProviderSpec(

tests/unit/server/test_sse.py

@@ -7,13 +7,10 @@
 import asyncio
 from unittest.mock import AsyncMock, MagicMock
 
-import pytest
-
 from llama_stack.apis.common.responses import PaginatedResponse
 from llama_stack.distribution.server.server import create_dynamic_typed_route, create_sse_event, sse_generator
 
 
-@pytest.mark.asyncio
 async def test_sse_generator_basic():
     # An AsyncIterator wrapped in an Awaitable, just like our web methods
     async def async_event_gen():
@@ -35,7 +32,6 @@ async def test_sse_generator_basic():
     assert seen_events[1] == create_sse_event("Test event 2")
 
 
-@pytest.mark.asyncio
 async def test_sse_generator_client_disconnected():
     # An AsyncIterator wrapped in an Awaitable, just like our web methods
     async def async_event_gen():
@@ -58,7 +54,6 @@ async def test_sse_generator_client_disconnected():
     assert seen_events[0] == create_sse_event("Test event 1")
 
 
-@pytest.mark.asyncio
 async def test_sse_generator_client_disconnected_before_response_starts():
     # Disconnect before the response starts
     async def async_event_gen():
@@ -75,7 +70,6 @@ async def test_sse_generator_client_disconnected_before_response_starts():
     assert len(seen_events) == 0
 
 
-@pytest.mark.asyncio
 async def test_sse_generator_error_before_response_starts():
     # Raise an error before the response starts
     async def async_event_gen():
@@ -93,7 +87,6 @@ async def test_sse_generator_error_before_response_starts():
     assert 'data: {"error":' in seen_events[0]
 
 
-@pytest.mark.asyncio
 async def test_paginated_response_url_setting():
     """Test that PaginatedResponse gets url set to route path."""