refactor: enforce top-level imports for llama-stack-api

Enforce that all imports from llama-stack-api use the form:

from llama_stack_api import <symbol>

 This prevents external code from accessing internal package structure
 (e.g., llama_stack_api.agents, llama_stack_api.common.*) and establishes
 a clear public API boundary.

 Changes:
 - Export 400+ symbols from llama_stack_api/__init__.py
 - Include all API types, common utilities, and strong_typing helpers
 - Update files across src/llama_stack, docs/, tests/, scripts/
 - Convert all submodule imports to top-level imports
 - ensure docs use the proper importing structure

 Addresses PR review feedback requiring explicit __all__ definition to
 prevent "peeking inside" the API package.

Signed-off-by: Charlie Doern <cdoern@redhat.com>

tests/unit/providers/nvidia/test_datastore.py

@@ -8,8 +8,7 @@ import os
 from unittest.mock import patch
 
 import pytest
-from llama_stack_api.datasets import Dataset, DatasetPurpose, URIDataSource
-from llama_stack_api.resource import ResourceType
+from llama_stack_api import Dataset, DatasetPurpose, ResourceType, URIDataSource
 
 from llama_stack.providers.remote.datasetio.nvidia.config import NvidiaDatasetIOConfig
 from llama_stack.providers.remote.datasetio.nvidia.datasetio import NvidiaDatasetIOAdapter

tests/unit/providers/nvidia/test_eval.py

@@ -8,11 +8,17 @@ import os
 from unittest.mock import MagicMock, patch
 
 import pytest
-from llama_stack_api.benchmarks import Benchmark
-from llama_stack_api.common.job_types import Job, JobStatus
-from llama_stack_api.eval import BenchmarkConfig, EvaluateResponse, ModelCandidate, SamplingParams
-from llama_stack_api.inference import TopPSamplingStrategy
-from llama_stack_api.resource import ResourceType
+from llama_stack_api import (
+    Benchmark,
+    BenchmarkConfig,
+    EvaluateResponse,
+    Job,
+    JobStatus,
+    ModelCandidate,
+    ResourceType,
+    SamplingParams,
+    TopPSamplingStrategy,
+)
 
 from llama_stack.models.llama.sku_types import CoreModelId
 from llama_stack.providers.remote.eval.nvidia.config import NVIDIAEvalConfig

tests/unit/providers/nvidia/test_parameters.py

@@ -9,7 +9,7 @@ import warnings
 from unittest.mock import patch
 
 import pytest
-from llama_stack_api.post_training import (
+from llama_stack_api import (
     DataConfig,
     DatasetFormat,
     EfficiencyConfig,

tests/unit/providers/nvidia/test_rerank_inference.py

@@ -8,7 +8,7 @@ from unittest.mock import AsyncMock, MagicMock, patch
 
 import aiohttp
 import pytest
-from llama_stack_api.models import ModelType
+from llama_stack_api import ModelType
 
 from llama_stack.providers.remote.inference.nvidia.config import NVIDIAConfig
 from llama_stack.providers.remote.inference.nvidia.nvidia import NVIDIAInferenceAdapter

tests/unit/providers/nvidia/test_safety.py

@@ -9,13 +9,14 @@ from typing import Any
 from unittest.mock import AsyncMock, MagicMock, patch
 
 import pytest
-from llama_stack_api.inference import (
+from llama_stack_api import (
     OpenAIAssistantMessageParam,
     OpenAIUserMessageParam,
+    ResourceType,
+    RunShieldResponse,
+    Shield,
+    ViolationLevel,
 )
-from llama_stack_api.resource import ResourceType
-from llama_stack_api.safety import RunShieldResponse, ViolationLevel
-from llama_stack_api.shields import Shield
 
 from llama_stack.providers.remote.safety.nvidia.config import NVIDIASafetyConfig
 from llama_stack.providers.remote.safety.nvidia.nvidia import NVIDIASafetyAdapter

tests/unit/providers/nvidia/test_supervised_fine_tuning.py

@@ -9,7 +9,7 @@ import warnings
 from unittest.mock import patch
 
 import pytest
-from llama_stack_api.post_training import (
+from llama_stack_api import (
     DataConfig,
     DatasetFormat,
     LoraFinetuningConfig,