feat: split API and provider specs into separate llama-stack-api pkg (#3895)

# What does this PR do?

Extract API definitions and provider specifications into a standalone
llama-stack-api package that can be published to PyPI independently of
the main llama-stack server.


see: https://github.com/llamastack/llama-stack/pull/2978 and
https://github.com/llamastack/llama-stack/pull/2978#issuecomment-3145115942

Motivation

External providers currently import from llama-stack, which overrides
the installed version and causes dependency conflicts. This separation
allows external providers to:

- Install only the type definitions they need without server
dependencies
- Avoid version conflicts with the installed llama-stack package
- Be versioned and released independently

This enables us to re-enable external provider module tests that were
previously blocked by these import conflicts.

Changes

- Created llama-stack-api package with minimal dependencies (pydantic,
jsonschema)
- Moved APIs, providers datatypes, strong_typing, and schema_utils
- Updated all imports from llama_stack.* to llama_stack_api.*
- Configured local editable install for development workflow
- Updated linting and type-checking configuration for both packages

Next Steps

- Publish llama-stack-api to PyPI
- Update external provider dependencies
- Re-enable external provider module tests


Pre-cursor PRs to this one:

- #4093 
- #3954 
- #4064 

These PRs moved key pieces _out_ of the Api pkg, limiting the scope of
change here.


relates to #3237 

## Test Plan

Package builds successfully and can be imported independently. All
pre-commit hooks pass with expected exclusions maintained.

---------

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

tests/unit/providers/nvidia/test_datastore.py

@@ -8,9 +8,8 @@ import os
 from unittest.mock import patch
 
 import pytest
+from llama_stack_api import Dataset, DatasetPurpose, ResourceType, URIDataSource
 
-from llama_stack.apis.datasets import Dataset, DatasetPurpose, URIDataSource
-from llama_stack.apis.resource import ResourceType
 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

@@ -8,12 +8,18 @@ import os
 from unittest.mock import MagicMock, patch
 
 import pytest
+from llama_stack_api import (
+    Benchmark,
+    BenchmarkConfig,
+    EvaluateResponse,
+    Job,
+    JobStatus,
+    ModelCandidate,
+    ResourceType,
+    SamplingParams,
+    TopPSamplingStrategy,
+)
 
-from llama_stack.apis.benchmarks import Benchmark
-from llama_stack.apis.common.job_types import Job, JobStatus
-from llama_stack.apis.eval.eval import BenchmarkConfig, EvaluateResponse, ModelCandidate, SamplingParams
-from llama_stack.apis.inference.inference import TopPSamplingStrategy
-from llama_stack.apis.resource import ResourceType
 from llama_stack.models.llama.sku_types import CoreModelId
 from llama_stack.providers.remote.eval.nvidia.config import NVIDIAEvalConfig
 from llama_stack.providers.remote.eval.nvidia.eval import NVIDIAEvalImpl

tests/unit/providers/nvidia/test_parameters.py

@@ -9,8 +9,7 @@ import warnings
 from unittest.mock import patch
 
 import pytest
-
-from llama_stack.apis.post_training.post_training import (
+from llama_stack_api import (
     DataConfig,
     DatasetFormat,
     EfficiencyConfig,
@@ -19,6 +18,7 @@ from llama_stack.apis.post_training.post_training import (
     OptimizerType,
     TrainingConfig,
 )
+
 from llama_stack.core.library_client import convert_pydantic_to_json_value
 from llama_stack.providers.remote.post_training.nvidia.post_training import (
     NvidiaPostTrainingAdapter,

tests/unit/providers/nvidia/test_rerank_inference.py

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

tests/unit/providers/nvidia/test_safety.py

@@ -9,14 +9,15 @@ from typing import Any
 from unittest.mock import AsyncMock, MagicMock, patch
 
 import pytest
-
-from llama_stack.apis.inference import (
+from llama_stack_api import (
     OpenAIAssistantMessageParam,
     OpenAIUserMessageParam,
+    ResourceType,
+    RunShieldResponse,
+    Shield,
+    ViolationLevel,
 )
-from llama_stack.apis.resource import ResourceType
-from llama_stack.apis.safety import RunShieldResponse, ViolationLevel
-from llama_stack.apis.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

@@ -9,8 +9,7 @@ import warnings
 from unittest.mock import patch
 
 import pytest
-
-from llama_stack.apis.post_training.post_training import (
+from llama_stack_api import (
     DataConfig,
     DatasetFormat,
     LoraFinetuningConfig,
@@ -19,6 +18,7 @@ from llama_stack.apis.post_training.post_training import (
     QATFinetuningConfig,
     TrainingConfig,
 )
+
 from llama_stack.core.library_client import convert_pydantic_to_json_value
 from llama_stack.providers.remote.post_training.nvidia.post_training import (
     ListNvidiaPostTrainingJobs,