llama-stack-api 0.6.1__tar.gz → 0.7.1__tar.gz

{llama_stack_api-0.6.1 → llama_stack_api-0.7.1}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: llama-stack-api
-Version: 0.6.1
+Version: 0.7.1
 Summary: API and Provider specifications for Llama Stack - lightweight package with protocol definitions and provider specs
 Author-email: Meta Llama <llama-oss@meta.com>
 License: MIT
@@ -66,8 +66,13 @@ uv pip install llama-stack-api
 ## Dependencies
 
 Minimal dependencies:
+
+- `openai>=2.5.0` - For OpenAI-compatible types
+- `fastapi>=0.115.0,<1.0` - For FastAPI route definitions
 - `pydantic>=2.11.9` - For data validation and serialization
 - `jsonschema` - For JSON schema utilities
+- `opentelemetry-sdk>=1.30.0` - For telemetry
+- `opentelemetry-exporter-otlp-proto-http>=1.30.0` - For OTLP export
 
 ## Versioning
 
@@ -77,19 +82,24 @@ This package follows semantic versioning independently from the main `llama-stac
 - **Minor versions** (0.x.0): New APIs, backward-compatible changes
 - **Major versions** (x.0.0): Breaking changes to existing APIs
 
-Current version: **0.4.0.dev0**
+The version is determined dynamically via `setuptools-scm` at build time.
 
 ## Usage Example
 
 ```python
-from llama_stack_api.inference import Inference, ChatCompletionRequest
-from llama_stack_api.providers.datatypes import ProviderSpec, InlineProviderSpec
-from llama_stack_api.datatypes import Api
+from llama_stack_api import (
+    Api,
+    Inference,
+    InlineProviderSpec,
+    OpenAIChatCompletionRequestWithExtraBody,
+)
 
 
 # Use protocol definitions for type checking
 class MyInferenceProvider(Inference):
-    async def chat_completion(self, request: ChatCompletionRequest):
+    async def openai_chat_completion(
+        self, request: OpenAIChatCompletionRequestWithExtraBody
+    ):
         # Your implementation
         pass
 
@@ -107,6 +117,7 @@ my_provider_spec = InlineProviderSpec(
 ## Relationship to llama-stack
 
 The main `llama-stack` package depends on `llama-stack-api` and provides:
+
 - Full server implementation
 - Built-in provider implementations
 - CLI tools for running and managing stacks

{llama_stack_api-0.6.1 → llama_stack_api-0.7.1}/README.md

@@ -42,8 +42,13 @@ uv pip install llama-stack-api
 ## Dependencies
 
 Minimal dependencies:
+
+- `openai>=2.5.0` - For OpenAI-compatible types
+- `fastapi>=0.115.0,<1.0` - For FastAPI route definitions
 - `pydantic>=2.11.9` - For data validation and serialization
 - `jsonschema` - For JSON schema utilities
+- `opentelemetry-sdk>=1.30.0` - For telemetry
+- `opentelemetry-exporter-otlp-proto-http>=1.30.0` - For OTLP export
 
 ## Versioning
 
@@ -53,19 +58,24 @@ This package follows semantic versioning independently from the main `llama-stac
 - **Minor versions** (0.x.0): New APIs, backward-compatible changes
 - **Major versions** (x.0.0): Breaking changes to existing APIs
 
-Current version: **0.4.0.dev0**
+The version is determined dynamically via `setuptools-scm` at build time.
 
 ## Usage Example
 
 ```python
-from llama_stack_api.inference import Inference, ChatCompletionRequest
-from llama_stack_api.providers.datatypes import ProviderSpec, InlineProviderSpec
-from llama_stack_api.datatypes import Api
+from llama_stack_api import (
+    Api,
+    Inference,
+    InlineProviderSpec,
+    OpenAIChatCompletionRequestWithExtraBody,
+)
 
 
 # Use protocol definitions for type checking
 class MyInferenceProvider(Inference):
-    async def chat_completion(self, request: ChatCompletionRequest):
+    async def openai_chat_completion(
+        self, request: OpenAIChatCompletionRequestWithExtraBody
+    ):
         # Your implementation
         pass
 
@@ -83,6 +93,7 @@ my_provider_spec = InlineProviderSpec(
 ## Relationship to llama-stack
 
 The main `llama-stack` package depends on `llama-stack-api` and provides:
+
 - Full server implementation
 - Built-in provider implementations
 - CLI tools for running and managing stacks

{llama_stack_api-0.6.1 → llama_stack_api-0.7.1}/__init__.py

@@ -15,18 +15,23 @@ server implementation.
 All imports from this package MUST use the form:
     from llama_stack_api import <symbol>
 
-Sub-module imports (e.g., from llama_stack_api.agents import Agents) are NOT supported
+Sub-module imports (e.g., from llama_stack_api.responses import Responses) are NOT supported
 and considered a code smell. All exported symbols are explicitly listed in __all__.
 """
 
-__version__ = "0.4.0.dev0"
+try:
+    from importlib.metadata import PackageNotFoundError
+    from importlib.metadata import version as _metadata_version
+
+    __version__ = _metadata_version("llama-stack-api")
+except PackageNotFoundError:
+    __version__ = "0.0.0.dev0"
 
 # Import submodules for those who need them
 from .schema_utils import (  # noqa: I001
     CallableT,
     ExtraBodyField,
     SchemaInfo,
-    WebMethod,
     clear_dynamic_schema_types,
     get_registered_schema_info,
     iter_dynamic_schema_types,
@@ -35,7 +40,6 @@ from .schema_utils import (  # noqa: I001
     json_schema_type,
     register_dynamic_schema_type,
     register_schema,
-    webmethod,
 )
 from .admin import (
     Admin,
@@ -51,8 +55,9 @@ from .admin import (
 )
 
 # Import all public API symbols
-from .agents import (
-    Agents,
+from .responses import (
+    Responses,
+    CancelResponseRequest,
     CreateResponseRequest,
     DeleteResponseRequest,
     ListResponseInputItemsRequest,
@@ -61,6 +66,7 @@ from .agents import (
     ResponseGuardrailSpec,
     ResponseItemInclude,
     ResponseTruncation,
+    ResponseStreamOptions,
     RetrieveResponseRequest,
 )
 from .batches import (
@@ -117,7 +123,6 @@ from .common.errors import (
 )
 from .common.job_types import Job, JobStatus
 from .common.responses import Order, PaginatedResponse
-from .common.training_types import Checkpoint, PostTrainingMetric
 from .common.type_system import (
     ChatCompletionInputType,
     CompletionInputType,
@@ -157,6 +162,7 @@ from .conversations import (
     UpdateConversationRequest,
 )
 from .datasetio import (
+    AppendRowsParams,
     AppendRowsRequest,
     DatasetIO,
     DatasetStore,
@@ -250,6 +256,8 @@ from .inference import (
     ModelStore,
     OpenAIAssistantMessageParam,
     OpenAIChatCompletion,
+    OpenAIChatCompletionChunkWithReasoning,
+    OpenAIChatCompletionWithReasoning,
     OpenAIChatCompletionChunk,
     OpenAIChatCompletionContentPartImageParam,
     OpenAIChatCompletionContentPartParam,
@@ -257,6 +265,8 @@ from .inference import (
     OpenAIChatCompletionMessageContent,
     OpenAIChatCompletionRequestWithExtraBody,
     OpenAIChatCompletionTextOnlyMessageContent,
+    OpenAIChatCompletionCustomToolCall,
+    OpenAIChatCompletionCustomToolCallFunction,
     OpenAIChatCompletionToolCall,
     OpenAIChatCompletionToolCallFunction,
     OpenAIChatCompletionToolChoice,
@@ -417,6 +427,9 @@ from .openai_responses import (
     OpenAIResponseOutputMessageFunctionToolCall,
     OpenAIResponseOutputMessageMCPCall,
     OpenAIResponseOutputMessageMCPListTools,
+    OpenAIResponseOutputMessageReasoningContent,
+    OpenAIResponseOutputMessageReasoningItem,
+    OpenAIResponseOutputMessageReasoningSummary,
     OpenAIResponseOutputMessageWebSearchToolCall,
     OpenAIResponsePrompt,
     OpenAIResponseReasoning,
@@ -429,32 +442,6 @@ from .openai_responses import (
     OpenAIResponseUsageOutputTokensDetails,
     WebSearchToolTypes,
 )
-from .post_training import (
-    AlgorithmConfig,
-    CancelTrainingJobRequest,
-    DataConfig,
-    DatasetFormat,
-    DPOAlignmentConfig,
-    DPOLossType,
-    EfficiencyConfig,
-    GetTrainingJobArtifactsRequest,
-    GetTrainingJobStatusRequest,
-    ListPostTrainingJobsResponse,
-    LoraFinetuningConfig,
-    OptimizerConfig,
-    OptimizerType,
-    PostTraining,
-    PostTrainingJob,
-    PostTrainingJobArtifactsResponse,
-    PostTrainingJobLogStream,
-    PostTrainingJobStatusResponse,
-    PostTrainingRLHFRequest,
-    PreferenceOptimizeRequest,
-    QATFinetuningConfig,
-    RLHFAlgorithm,
-    SupervisedFineTuneRequest,
-    TrainingConfig,
-)
 from .prompts import (
     CreatePromptRequest,
     DeletePromptRequest,
@@ -535,6 +522,7 @@ from .shields import (
 from .tools import (
     ListToolDefsResponse,
     ListToolGroupsResponse,
+    ListToolsRequest,
     SpecialToolGroup,
     ToolDef,
     ToolGroup,
@@ -609,15 +597,15 @@ __all__ = [
     "LLAMA_STACK_API_V1ALPHA",
     "LLAMA_STACK_API_V1BETA",
     # API Symbols
-    "Agents",
+    "Responses",
     "AggregationFunctionType",
-    # Agents Request Models
+    # Responses Request Models
+    "CancelResponseRequest",
     "CreateResponseRequest",
     "DeleteResponseRequest",
     "ListResponseInputItemsRequest",
     "ListResponsesRequest",
     "RetrieveResponseRequest",
-    "AlgorithmConfig",
     "AllowedToolsFilter",
     "Api",
     "ApiFilter",
@@ -627,7 +615,6 @@ __all__ = [
     "BatchNotFoundError",
     "BatchObject",
     "CancelBatchRequest",
-    "CancelTrainingJobRequest",
     "CreateBatchRequest",
     "ListBatchesRequest",
     "Benchmark",
@@ -639,7 +626,6 @@ __all__ = [
     "CallableT",
     "ChatCompletionInputType",
     "ChatCompletionResponseEventType",
-    "Checkpoint",
     "Chunk",
     "ChunkForDeletion",
     "ChunkMetadata",
@@ -680,18 +666,15 @@ __all__ = [
     "ListItemsRequest",
     "RetrieveItemRequest",
     "UpdateConversationRequest",
-    "DPOAlignmentConfig",
-    "DPOLossType",
-    "DataConfig",
     "DataSource",
     "Dataset",
-    "DatasetFormat",
     "DatasetIO",
     "DatasetInput",
     "DatasetPurpose",
     "DatasetNotFoundError",
     "DatasetStore",
     "DatasetType",
+    "AppendRowsParams",
     "AppendRowsRequest",
     "IterRowsRequest",
     "Datasets",
@@ -700,7 +683,6 @@ __all__ = [
     "DeleteFileRequest",
     "Docstring",
     "DynamicApiMeta",
-    "EfficiencyConfig",
     "EmbeddingTaskType",
     "EmbeddingsResponse",
     "Error",
@@ -790,7 +772,6 @@ __all__ = [
     "ListOpenAIFileResponse",
     "ListOpenAIResponseInputItem",
     "ListOpenAIResponseObject",
-    "ListPostTrainingJobsResponse",
     "ListPromptsResponse",
     "ListProvidersResponse",
     "ListRoutesRequest",
@@ -799,9 +780,9 @@ __all__ = [
     "ListShieldsResponse",
     "ListToolDefsResponse",
     "ListToolGroupsResponse",
+    "ListToolsRequest",
     "ListToolsResponse",
     "LogProbConfig",
-    "LoraFinetuningConfig",
     "MAX_PAGINATION_LIMIT",
     "MCPListToolsTool",
     "Metadata",
@@ -814,8 +795,6 @@ __all__ = [
     "ModelTypeError",
     "Models",
     "GetModelRequest",
-    "GetTrainingJobArtifactsRequest",
-    "GetTrainingJobStatusRequest",
     "RegisterModelRequest",
     "UnregisterModelRequest",
     "ModelsProtocolPrivate",
@@ -826,12 +805,16 @@ __all__ = [
     "OpenAIAssistantMessageParam",
     "OpenAIChatCompletion",
     "OpenAIChatCompletionChunk",
+    "OpenAIChatCompletionChunkWithReasoning",
+    "OpenAIChatCompletionWithReasoning",
     "OpenAIChatCompletionContentPartImageParam",
     "OpenAIChatCompletionContentPartParam",
     "OpenAIChatCompletionContentPartTextParam",
     "OpenAIChatCompletionMessageContent",
     "OpenAIChatCompletionRequestWithExtraBody",
     "OpenAIChatCompletionTextOnlyMessageContent",
+    "OpenAIChatCompletionCustomToolCall",
+    "OpenAIChatCompletionCustomToolCallFunction",
     "OpenAIChatCompletionToolCall",
     "OpenAIChatCompletionToolCallFunction",
     "OpenAIChatCompletionUsage",
@@ -962,6 +945,9 @@ __all__ = [
     "OpenAIResponseOutputMessageFunctionToolCall",
     "OpenAIResponseOutputMessageMCPCall",
     "OpenAIResponseOutputMessageMCPListTools",
+    "OpenAIResponseOutputMessageReasoningContent",
+    "OpenAIResponseOutputMessageReasoningItem",
+    "OpenAIResponseOutputMessageReasoningSummary",
     "OpenAIResponseOutputMessageWebSearchToolCall",
     "OpenAIResponsePrompt",
     "OpenAIResponseReasoning",
@@ -977,21 +963,11 @@ __all__ = [
     "OpenAIToolMessageParam",
     "OpenAITopLogProb",
     "OpenAIUserMessageParam",
-    "OptimizerConfig",
-    "OptimizerType",
     "PaginatedResponse",
     "ParamType",
     "parse_type",
-    "PostTraining",
     "ProcessFileRequest",
     "ProcessFileResponse",
-    "PostTrainingMetric",
-    "PostTrainingJob",
-    "PostTrainingJobArtifactsResponse",
-    "PostTrainingJobLogStream",
-    "PostTrainingJobStatusResponse",
-    "PostTrainingRLHFRequest",
-    "PreferenceOptimizeRequest",
     "Prompt",
     "Prompts",
     "CreatePromptRequest",
@@ -1006,7 +982,6 @@ __all__ = [
     "ProviderSpec",
     "Providers",
     "python_type_to_name",
-    "QATFinetuningConfig",
     "QuantizationConfig",
     "QuantizationType",
     "QueryChunksRequest",
@@ -1019,7 +994,6 @@ __all__ = [
     "RAGSearchMode",
     "register_dynamic_schema_type",
     "register_schema",
-    "RLHFAlgorithm",
     "RRFRanker",
     "Ranker",
     "RegexParserScoringFnParams",
@@ -1039,6 +1013,7 @@ __all__ = [
     "ResponseItemInclude",
     "ResponseTruncation",
     "ResponseNotFoundError",
+    "ResponseStreamOptions",
     "RetrieveFileContentRequest",
     "RetrieveFileRequest",
     "RouteInfo",
@@ -1085,7 +1060,6 @@ __all__ = [
     "SpecialToolGroup",
     "StrictJsonType",
     "StringType",
-    "SupervisedFineTuneRequest",
     "SystemMessage",
     "SystemMessageBehavior",
     "TextContentItem",
@@ -1105,7 +1079,6 @@ __all__ = [
     "ToolStore",
     "TopKSamplingStrategy",
     "TopPSamplingStrategy",
-    "TrainingConfig",
     "UnsupportedModelError",
     "unwrap_generic_list",
     "unwrap_optional_type",
@@ -1147,8 +1120,6 @@ __all__ = [
     "VectorStoresProtocolPrivate",
     "VersionInfo",
     "ViolationLevel",
-    "webmethod",
-    "WebMethod",
     "WebSearchToolTypes",
     "WeightedRanker",
     # Validators

{llama_stack_api-0.6.1 → llama_stack_api-0.7.1}/benchmarks/api.py

@@ -18,6 +18,8 @@ from .models import (
 
 @runtime_checkable
 class Benchmarks(Protocol):
+    """Protocol for managing benchmark resources."""
+
     async def list_benchmarks(
         self,
         request: ListBenchmarksRequest,

{llama_stack_api-0.6.1 → llama_stack_api-0.7.1}/benchmarks/models.py

@@ -54,6 +54,8 @@ class UnregisterBenchmarkRequest(BaseModel):
 
 
 class CommonBenchmarkFields(BaseModel):
+    """Common fields shared across benchmark creation and retrieval."""
+
     dataset_id: str = Field(..., description="Identifier of the dataset to use for the benchmark evaluation.")
     scoring_functions: list[str] = Field(
         ..., description="List of scoring function identifiers to apply during evaluation."
@@ -83,6 +85,8 @@ class Benchmark(CommonBenchmarkFields, Resource):
 
 
 class BenchmarkInput(CommonBenchmarkFields, BaseModel):
+    """Input model for registering a new benchmark."""
+
     benchmark_id: str = Field(..., description="The ID of the benchmark.")
     provider_id: str | None = Field(default=None, description="The ID of the provider to use for the benchmark.")
     provider_benchmark_id: str | None = Field(

{llama_stack_api-0.6.1 → llama_stack_api-0.7.1}/common/content_types.py

@@ -4,7 +4,7 @@
 # This source code is licensed under the terms described in the LICENSE file in
 # the root directory of this source tree.
 
-from typing import Annotated, Literal
+from typing import Annotated, Any, Literal
 
 from pydantic import BaseModel, Field, model_validator
 
@@ -35,7 +35,7 @@ class _URLOrData(BaseModel):
 
     @model_validator(mode="before")
     @classmethod
-    def validator(cls, values):
+    def validator(cls, values: Any) -> dict[str, Any]:
         if isinstance(values, dict):
             return values
         return {"url": values}

{llama_stack_api-0.6.1 → llama_stack_api-0.7.1}/common/job_types.py

@@ -3,6 +3,7 @@
 #
 # This source code is licensed under the terms described in the LICENSE file in
 # the root directory of this source tree.
+
 from enum import Enum
 
 from pydantic import BaseModel

{llama_stack_api-0.6.1 → llama_stack_api-0.7.1}/connectors/fastapi_routes.py

@@ -16,7 +16,7 @@ from fastapi import APIRouter, Depends, Path, Query
 
 from llama_stack_api.router_utils import create_path_dependency, standard_responses
 from llama_stack_api.tools import ToolDef
-from llama_stack_api.version import LLAMA_STACK_API_V1ALPHA
+from llama_stack_api.version import LLAMA_STACK_API_V1BETA
 
 from .api import Connectors
 from .models import (
@@ -43,7 +43,7 @@ def create_router(impl: Connectors) -> APIRouter:
         APIRouter configured for the Connectors API
     """
     router = APIRouter(
-        prefix=f"/{LLAMA_STACK_API_V1ALPHA}",
+        prefix=f"/{LLAMA_STACK_API_V1BETA}",
         tags=["Connectors"],
         responses=standard_responses,
     )

{llama_stack_api-0.6.1 → llama_stack_api-0.7.1}/conversations/models.py

@@ -24,6 +24,7 @@ from llama_stack_api.openai_responses import (
     OpenAIResponseOutputMessageFunctionToolCall,
     OpenAIResponseOutputMessageMCPCall,
     OpenAIResponseOutputMessageMCPListTools,
+    OpenAIResponseOutputMessageReasoningItem,
     OpenAIResponseOutputMessageWebSearchToolCall,
 )
 from llama_stack_api.schema_utils import json_schema_type, register_schema
@@ -74,8 +75,7 @@ ConversationItem = Annotated[
     | OpenAIResponseMCPApprovalResponse
     | OpenAIResponseOutputMessageMCPCall
     | OpenAIResponseOutputMessageMCPListTools
-    | OpenAIResponseOutputMessageMCPCall
-    | OpenAIResponseOutputMessageMCPListTools,
+    | OpenAIResponseOutputMessageReasoningItem,
     Field(discriminator="type"),
 ]
 register_schema(ConversationItem, name="ConversationItem")

{llama_stack_api-0.6.1 → llama_stack_api-0.7.1}/datasetio/__init__.py

@@ -19,6 +19,7 @@ from .api import DatasetIO, DatasetStore
 
 # Import models for re-export
 from .models import (
+    AppendRowsParams,
     AppendRowsRequest,
     IterRowsRequest,
     PaginatedResponse,
@@ -27,6 +28,7 @@ from .models import (
 __all__ = [
     "DatasetIO",
     "DatasetStore",
+    "AppendRowsParams",
     "AppendRowsRequest",
     "IterRowsRequest",
     "PaginatedResponse",

{llama_stack_api-0.6.1 → llama_stack_api-0.7.1}/datasetio/api.py

@@ -16,13 +16,15 @@ from typing import Protocol, runtime_checkable
 from llama_stack_api.datasets import Dataset
 
 from .models import (
-    AppendRowsRequest,
+    AppendRowsParams,
     IterRowsRequest,
     PaginatedResponse,
 )
 
 
 class DatasetStore(Protocol):
+    """Protocol for storing and retrieving dataset definitions."""
+
     def get_dataset(self, dataset_id: str) -> Dataset: ...
 
 
@@ -39,4 +41,4 @@ class DatasetIO(Protocol):
 
     async def iterrows(self, request: IterRowsRequest) -> PaginatedResponse: ...
 
-    async def append_rows(self, request: AppendRowsRequest) -> None: ...
+    async def append_rows(self, params: AppendRowsParams) -> None: ...

{llama_stack_api-0.6.1 → llama_stack_api-0.7.1}/datasetio/fastapi_routes.py

@@ -20,6 +20,7 @@ from llama_stack_api.version import LLAMA_STACK_API_V1BETA
 
 from .api import DatasetIO
 from .models import (
+    AppendRowsParams,
     AppendRowsRequest,
     IterRowsRequest,
 )
@@ -82,13 +83,13 @@ The response includes:
     )
     async def append_rows(
         dataset_id: Annotated[str, Path(description="The ID of the dataset to append the rows to.")],
-        request: Annotated[AppendRowsRequest, Body(...)],
+        body: Annotated[AppendRowsRequest, Body(...)],
     ) -> None:
-        # Override the dataset_id from the path
-        request_with_id = AppendRowsRequest(
+        # Combine path parameter with request body
+        params = AppendRowsParams(
             dataset_id=dataset_id,
-            rows=request.rows,
+            rows=body.rows,
         )
-        return await impl.append_rows(request_with_id)
+        return await impl.append_rows(params)
 
     return router

{llama_stack_api-0.6.1 → llama_stack_api-0.7.1}/datasetio/models.py

@@ -35,7 +35,13 @@ class IterRowsRequest(BaseModel):
 
 @json_schema_type
 class AppendRowsRequest(BaseModel):
-    """Request model for appending rows to a dataset."""
+    """Request body for appending rows to a dataset."""
+
+    rows: list[dict[str, Any]] = Field(..., description="The rows to append to the dataset.")
+
+
+class AppendRowsParams(BaseModel):
+    """Internal parameters for appending rows to a dataset (includes dataset_id)."""
 
     dataset_id: str = Field(..., description="The ID of the dataset to append the rows to.")
     rows: list[dict[str, Any]] = Field(..., description="The rows to append to the dataset.")
@@ -43,6 +49,7 @@ class AppendRowsRequest(BaseModel):
 
 __all__ = [
     "AppendRowsRequest",
+    "AppendRowsParams",
     "IterRowsRequest",
     "PaginatedResponse",
 ]

{llama_stack_api-0.6.1 → llama_stack_api-0.7.1}/datasets/models.py

@@ -22,8 +22,6 @@ from llama_stack_api.schema_utils import json_schema_type, register_schema
 class DatasetPurpose(StrEnum):
     """Purpose of the dataset. Each purpose has a required input data schema."""
 
-    post_training_messages = "post-training/messages"
-    """The dataset contains messages used for post-training."""
     eval_question_answer = "eval/question-answer"
     """The dataset contains a question column and an answer column."""
     eval_messages_answer = "eval/messages-answer"