phoenix-oss/llama-stack-mirror
0
0
Fork 1
mirror of https://github.com/meta-llama/llama-stack.git synced 2025-12-15 23:32:38 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions

Merge branch 'main' into dead_code_removal

Browse source
This commit is contained in:
Omar Abdelwahab 2025-10-06 13:21:36 -07:00 • committed by GitHub
commit 9886520b40
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
927 changed files with 171924 additions and 102933 deletions
Download patch file Download diff file Expand all files Collapse all files

66
llama_stack/apis/inference/inference.py
View file

@ -982,45 +982,6 @@ class InferenceProvider(Protocol):
model_store: ModelStore | None = None
async def chat_completion(
self,
model_id: str,
messages: list[Message],
sampling_params: SamplingParams | None = None,
tools: list[ToolDefinition] | None = None,
tool_choice: ToolChoice | None = ToolChoice.auto,
tool_prompt_format: ToolPromptFormat | None = None,
response_format: ResponseFormat | None = None,
stream: bool | None = False,
logprobs: LogProbConfig | None = None,
tool_config: ToolConfig | None = None,
) -> ChatCompletionResponse | AsyncIterator[ChatCompletionResponseStreamChunk]:
"""Generate a chat completion for the given messages using the specified model.
:param model_id: The identifier of the model to use. The model must be registered with Llama Stack and available via the /models endpoint.
:param messages: List of messages in the conversation.
:param sampling_params: Parameters to control the sampling strategy.
:param tools: (Optional) List of tool definitions available to the model.
:param tool_choice: (Optional) Whether tool use is required or automatic. Defaults to ToolChoice.auto.
.. deprecated::
Use tool_config instead.
:param tool_prompt_format: (Optional) Instructs the model how to format tool calls. By default, Llama Stack will attempt to use a format that is best adapted to the model.
- `ToolPromptFormat.json`: The tool calls are formatted as a JSON object.
- `ToolPromptFormat.function_tag`: The tool calls are enclosed in a <function=function_name> tag.
- `ToolPromptFormat.python_list`: The tool calls are output as Python syntax -- a list of function calls.
.. deprecated::
Use tool_config instead.
:param response_format: (Optional) Grammar specification for guided (structured) decoding. There are two options:
- `ResponseFormat.json_schema`: The grammar is a JSON schema. Most providers support this format.
- `ResponseFormat.grammar`: The grammar is a BNF grammar. This format is more flexible, but not all providers support it.
:param stream: (Optional) If True, generate an SSE event stream of the response. Defaults to False.
:param logprobs: (Optional) If specified, log probabilities for each token position will be returned.
:param tool_config: (Optional) Configuration for tool use.
:returns: If stream=False, returns a ChatCompletionResponse with the full completion.
If stream=True, returns an SSE event stream of ChatCompletionResponseStreamChunk.
"""
...
@webmethod(route="/inference/rerank", method="POST", level=LLAMA_STACK_API_V1ALPHA)
async def rerank(
self,
@ -1081,7 +1042,9 @@ class InferenceProvider(Protocol):
# for fill-in-the-middle type completion
suffix: str | None = None,
) -> OpenAICompletion:
"""Generate an OpenAI-compatible completion for the given prompt using the specified model.
"""Create completion.
Generate an OpenAI-compatible completion for the given prompt using the specified model.
:param model: The identifier of the model to use. The model must be registered with Llama Stack and available via the /models endpoint.
:param prompt: The prompt to generate a completion for.
@ -1138,7 +1101,9 @@ class InferenceProvider(Protocol):
top_p: float | None = None,
user: str | None = None,
) -> OpenAIChatCompletion | AsyncIterator[OpenAIChatCompletionChunk]:
"""Generate an OpenAI-compatible chat completion for the given messages using the specified model.
"""Create chat completions.
Generate an OpenAI-compatible chat completion for the given messages using the specified model.
:param model: The identifier of the model to use. The model must be registered with Llama Stack and available via the /models endpoint.
:param messages: List of messages in the conversation.
@ -1182,7 +1147,9 @@ class InferenceProvider(Protocol):
dimensions: int | None = None,
user: str | None = None,
) -> OpenAIEmbeddingsResponse:
"""Generate OpenAI-compatible embeddings for the given input using the specified model.
"""Create embeddings.
Generate OpenAI-compatible embeddings for the given input using the specified model.
:param model: The identifier of the model to use. The model must be an embedding model registered with Llama Stack and available via the /models endpoint.
:param input: Input text to embed, encoded as a string or array of strings. To embed multiple inputs in a single request, pass an array of strings.
@ -1195,7 +1162,9 @@ class InferenceProvider(Protocol):
class Inference(InferenceProvider):
"""Llama Stack Inference API for generating completions, chat completions, and embeddings.
"""Inference
Llama Stack Inference API for generating completions, chat completions, and embeddings.
This API provides the raw interface to the underlying models. Two kinds of models are supported:
- LLM models: these models generate "raw" and "chat" (conversational) completions.
@ -1216,7 +1185,7 @@ class Inference(InferenceProvider):
model: str | None = None,
order: Order | None = Order.desc,
) -> ListOpenAIChatCompletionResponse:
"""List all chat completions.
"""List chat completions.
:param after: The ID of the last chat completion to return.
:param limit: The maximum number of chat completions to return.
@ -1237,10 +1206,11 @@ class Inference(InferenceProvider):
method="GET",
level=LLAMA_STACK_API_V1,
)
async def get_chat_completion(
self, completion_id: str
) -> OpenAICompletionWithInputMessages:
"""Describe a chat completion by its ID.
@webmethod(route="/chat/completions/{completion_id}", method="GET", level=LLAMA_STACK_API_V1)
async def get_chat_completion(self, completion_id: str) -> OpenAICompletionWithInputMessages:
"""Get chat completion.
Describe a chat completion by its ID.
:param completion_id: ID of the chat completion.
:returns: A OpenAICompletionWithInputMessages.