phoenix-oss/llama-stack-mirror
0
0
Fork 1
mirror of https://github.com/meta-llama/llama-stack.git synced 2025-10-06 12:37:33 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions

Merge branch 'main' into responses_object

Browse source
This commit is contained in:
Emilio Garcia 2025-08-20 11:33:41 -04:00 • committed by Emilio Garcia
commit 8fb17ba18e
67 changed files with 794 additions and 218 deletions
Download patch file Download diff file Expand all files Collapse all files

264
docs/_static/llama-stack-spec.html vendored
View file

@ -8585,6 +8585,21 @@
"title": "OpenAIResponseError",
"description": "Error details for failed OpenAI response requests."
},
"OpenAIResponseIncompleteDetails": {
"type": "object",
"properties": {
"reason": {
"type": "string",
"description": "Reason for the response being incomplete"
}
},
"additionalProperties": false,
"required": [
"reason"
],
"title": "OpenAIResponseIncompleteDetails",
"description": "Incomplete details for OpenAI responses."
},
"OpenAIResponseObject": {
"type": "object",
"properties": {
@ -8600,6 +8615,39 @@
"type": "string",
"description": "Unique identifier for this response"
},
"incomplete_details": {
"$ref": "#/components/schemas/OpenAIResponseIncompleteDetails",
"description": "(Optional) Incomplete details if the response is incomplete"
},
"instructions": {
"oneOf": [
{
"type": "string"
},
{
"type": "array",
"items": {
"type": "string"
}
}
],
"description": "(Optional) A system (or developer) message inserted into the model's context."
},
"max_output_tokens": {
"type": "integer",
"description": "(Optional) An upper bound for the number of tokens that can be generated for a response, including visible output tokens and reasoning tokens."
},
"max_tool_calls": {
"type": "integer",
"description": "(Optional) The maximum number of total calls to built-in tools that can be processed in a response."
},
"metadata": {
"type": "object",
"additionalProperties": {
"type": "string"
},
"description": "(Optional) Set of 16 key-value pairs that can be attached to an object. This can be useful for storing additional information about the object in a structured format, and querying for objects via API or the dashboard."
},
"model": {
"type": "string",
"description": "Model identifier used for generation"
@ -8626,6 +8674,26 @@
"type": "string",
"description": "(Optional) ID of the previous response in a conversation"
},
"prompt": {
"$ref": "#/components/schemas/OpenAIResponsePrompt",
"description": "(Optional) Reference to a prompt template and its variables."
},
"prompt_cache_key": {
"type": "string",
"description": "(Optional)Used to cache responses for similar requests to optimize your cache hit rates. Replaces the user field."
},
"reasoning": {
"$ref": "#/components/schemas/OpenAIResponseReasoning",
"description": "(Optional) Configuration options for reasoning models."
},
"safety_identifier": {
"type": "string",
"description": "(Optional) A stable identifier used to help detect users of your application that may be violating OpenAI's usage policies."
},
"service_tier": {
"type": "string",
"description": "(Optional) Specifies the processing type used for serving the request."
},
"status": {
"type": "string",
"description": "Current status of the response generation"
@ -8638,17 +8706,29 @@
"$ref": "#/components/schemas/OpenAIResponseText",
"description": "Text formatting configuration for the response"
},
"tool_choice": {
"$ref": "#/components/schemas/OpenAIResponsesToolChoice"
},
"tools": {
"type": "array",
"items": {
"$ref": "#/components/schemas/OpenAIResponsesTool"
}
},
"top_logprobs": {
"type": "integer"
},
"top_p": {
"type": "number",
"description": "(Optional) Nucleus sampling parameter used for generation"
},
"truncation": {
"type": "string",
"description": "(Optional) Truncation strategy applied to the response"
},
"user": {
"type": "string",
"description": "(Optional) User identifier associated with the request"
},
"truncation": {
"type": "string",
"description": "(Optional) Truncation strategy applied to the response"
}
},
"additionalProperties": false,
@ -8659,11 +8739,10 @@
"object",
"output",
"parallel_tool_calls",
"status",
"text"
"status"
],
"title": "OpenAIResponseObject",
"description": "Complete OpenAI response object containing generation results and metadata."
"description": "Complete OpenAI response object containing generation results and metadata.\nBased on OpenAI Responses API schema: https://github.com/openai/openai-python/blob/34014aedbb8946c03e97e5c8d72e03ad2259cd7c/src/openai/types/responses/response.py#L38"
},
"OpenAIResponseOutput": {
"oneOf": [
@ -8821,6 +8900,103 @@
"title": "OpenAIResponseOutputMessageMCPListTools",
"description": "MCP list tools output message containing available tools from an MCP server."
},
"OpenAIResponsePrompt": {
"type": "object",
"properties": {
"id": {
"type": "string",
"description": "The unique identifier of the prompt template to use."
},
"variables": {
"type": "object",
"additionalProperties": {
"oneOf": [
{
"type": "null"
},
{
"type": "boolean"
},
{
"type": "number"
},
{
"type": "string"
},
{
"type": "array"
},
{
"type": "object"
}
]
},
"description": "(Optional) Map of values to substitute in for variables in your prompt. The substitution values can either be strings, or other Response input types like images or files."
},
"version": {
"type": "string",
"description": "(Optional) Version of the prompt template."
}
},
"additionalProperties": false,
"required": [
"id"
],
"title": "OpenAIResponsePrompt",
"description": "Reference to a prompt template and its variables."
},
"OpenAIResponseReasoning": {
"type": "object",
"properties": {
"effort": {
"type": "string",
"enum": [
"low",
"medium",
"high",
"minimal"
],
"description": "(Optional) The effort level to use for reasoning."
},
"generate_summary": {
"type": "string",
"description": "Deprecated. Use the generate_summary_text field instead. (Optional) Whether to generate a summary of the reasoning process."
},
"summary": {
"type": "string"
}
},
"additionalProperties": false,
"title": "OpenAIResponseReasoning",
"description": "Configuration options for reasoning models."
},
"OpenAIResponsesTool": {
"type": "object",
"properties": {
"description": {
"type": "string"
},
"name": {
"type": "string"
},
"parameters": {
"type": "object",
"title": "object",
"description": "The base class of the class hierarchy.\nWhen called, it accepts no arguments and returns a new featureless\ninstance that has no instance attributes and cannot be given any."
},
"type": {
"type": "string",
"const": "function"
}
},
"additionalProperties": false,
"title": "OpenAIResponsesTool"
},
"OpenAIResponsesToolChoice": {
"type": "object",
"title": "OpenAIResponsesToolChoice",
"description": "Type alias.\nType aliases are created through the type statement::\n\n type Alias = int\n\nIn this example, Alias and int will be treated equivalently by static\ntype checkers.\n\nAt runtime, Alias is an instance of TypeAliasType. The __name__\nattribute holds the name of the type alias. The value of the type alias\nis stored in the __value__ attribute. It is evaluated lazily, so the\nvalue is computed only if the attribute is accessed.\n\nType aliases can also be generic::\n\n type ListOrSet[T] = list[T] | set[T]\n\nIn this case, the type parameters of the alias are stored in the\n__type_params__ attribute.\n\nSee PEP 695 for more information."
},
"OpenAIResponseContentPart": {
"oneOf": [
{
@ -12744,6 +12920,39 @@
"type": "string",
"description": "Unique identifier for this response"
},
"incomplete_details": {
"$ref": "#/components/schemas/OpenAIResponseIncompleteDetails",
"description": "(Optional) Incomplete details if the response is incomplete"
},
"instructions": {
"oneOf": [
{
"type": "string"
},
{
"type": "array",
"items": {
"type": "string"
}
}
],
"description": "(Optional) A system (or developer) message inserted into the model's context."
},
"max_output_tokens": {
"type": "integer",
"description": "(Optional) An upper bound for the number of tokens that can be generated for a response, including visible output tokens and reasoning tokens."
},
"max_tool_calls": {
"type": "integer",
"description": "(Optional) The maximum number of total calls to built-in tools that can be processed in a response."
},
"metadata": {
"type": "object",
"additionalProperties": {
"type": "string"
},
"description": "(Optional) Set of 16 key-value pairs that can be attached to an object. This can be useful for storing additional information about the object in a structured format, and querying for objects via API or the dashboard."
},
"model": {
"type": "string",
"description": "Model identifier used for generation"
@ -12770,6 +12979,26 @@
"type": "string",
"description": "(Optional) ID of the previous response in a conversation"
},
"prompt": {
"$ref": "#/components/schemas/OpenAIResponsePrompt",
"description": "(Optional) Reference to a prompt template and its variables."
},
"prompt_cache_key": {
"type": "string",
"description": "(Optional)Used to cache responses for similar requests to optimize your cache hit rates. Replaces the user field."
},
"reasoning": {
"$ref": "#/components/schemas/OpenAIResponseReasoning",
"description": "(Optional) Configuration options for reasoning models."
},
"safety_identifier": {
"type": "string",
"description": "(Optional) A stable identifier used to help detect users of your application that may be violating OpenAI's usage policies."
},
"service_tier": {
"type": "string",
"description": "(Optional) Specifies the processing type used for serving the request."
},
"status": {
"type": "string",
"description": "Current status of the response generation"
@ -12782,18 +13011,30 @@
"$ref": "#/components/schemas/OpenAIResponseText",
"description": "Text formatting configuration for the response"
},
"tool_choice": {
"$ref": "#/components/schemas/OpenAIResponsesToolChoice"
},
"tools": {
"type": "array",
"items": {
"$ref": "#/components/schemas/OpenAIResponsesTool"
}
},
"top_logprobs": {
"type": "integer"
},
"top_p": {
"type": "number",
"description": "(Optional) Nucleus sampling parameter used for generation"
},
"truncation": {
"type": "string",
"description": "(Optional) Truncation strategy applied to the response"
},
"user": {
"type": "string",
"description": "(Optional) User identifier associated with the request"
},
"truncation": {
"type": "string",
"description": "(Optional) Truncation strategy applied to the response"
},
"input": {
"type": "array",
"items": {
@ -12811,7 +13052,6 @@
"output",
"parallel_tool_calls",
"status",
"text",
"input"
],
"title": "OpenAIResponseObjectWithInput",

260
docs/_static/llama-stack-spec.yaml vendored
View file

@ -6249,6 +6249,17 @@ components:
title: OpenAIResponseError
description: >-
Error details for failed OpenAI response requests.
OpenAIResponseIncompleteDetails:
type: object
properties:
reason:
type: string
description: Reason for the response being incomplete
additionalProperties: false
required:
- reason
title: OpenAIResponseIncompleteDetails
description: Incomplete details for OpenAI responses.
OpenAIResponseObject:
type: object
properties:
@ -6263,6 +6274,36 @@ components:
id:
type: string
description: Unique identifier for this response
incomplete_details:
$ref: '#/components/schemas/OpenAIResponseIncompleteDetails'
description: >-
(Optional) Incomplete details if the response is incomplete
instructions:
oneOf:
- type: string
- type: array
items:
type: string
description: >-
(Optional) A system (or developer) message inserted into the model's context.
max_output_tokens:
type: integer
description: >-
(Optional) An upper bound for the number of tokens that can be generated
for a response, including visible output tokens and reasoning tokens.
max_tool_calls:
type: integer
description: >-
(Optional) The maximum number of total calls to built-in tools that can
be processed in a response.
metadata:
type: object
additionalProperties:
type: string
description: >-
(Optional) Set of 16 key-value pairs that can be attached to an object.
This can be useful for storing additional information about the object
in a structured format, and querying for objects via API or the dashboard.
model:
type: string
description: Model identifier used for generation
@ -6287,6 +6328,28 @@ components:
type: string
description: >-
(Optional) ID of the previous response in a conversation
prompt:
$ref: '#/components/schemas/OpenAIResponsePrompt'
description: >-
(Optional) Reference to a prompt template and its variables.
prompt_cache_key:
type: string
description: >-
(Optional)Used to cache responses for similar requests to optimize your
cache hit rates. Replaces the user field.
reasoning:
$ref: '#/components/schemas/OpenAIResponseReasoning'
description: >-
(Optional) Configuration options for reasoning models.
safety_identifier:
type: string
description: >-
(Optional) A stable identifier used to help detect users of your application
that may be violating OpenAI's usage policies.
service_tier:
type: string
description: >-
(Optional) Specifies the processing type used for serving the request.
status:
type: string
description: >-
@ -6299,18 +6362,26 @@ components:
$ref: '#/components/schemas/OpenAIResponseText'
description: >-
Text formatting configuration for the response
tool_choice:
$ref: '#/components/schemas/OpenAIResponsesToolChoice'
tools:
type: array
items:
$ref: '#/components/schemas/OpenAIResponsesTool'
top_logprobs:
type: integer
top_p:
type: number
description: >-
(Optional) Nucleus sampling parameter used for generation
truncation:
type: string
description: >-
(Optional) Truncation strategy applied to the response
user:
type: string
description: >-
(Optional) User identifier associated with the request
truncation:
type: string
description: >-
(Optional) Truncation strategy applied to the response
additionalProperties: false
required:
- created_at
@ -6320,10 +6391,11 @@ components:
- output
- parallel_tool_calls
- status
- text
title: OpenAIResponseObject
description: >-
Complete OpenAI response object containing generation results and metadata.
Based on OpenAI Responses API schema: https://github.com/openai/openai-python/blob/34014aedbb8946c03e97e5c8d72e03ad2259cd7c/src/openai/types/responses/response.py#L38
OpenAIResponseOutput:
oneOf:
- $ref: '#/components/schemas/OpenAIResponseMessage'
@ -6441,6 +6513,115 @@ components:
title: OpenAIResponseOutputMessageMCPListTools
description: >-
MCP list tools output message containing available tools from an MCP server.
OpenAIResponsePrompt:
type: object
properties:
id:
type: string
description: >-
The unique identifier of the prompt template to use.
variables:
type: object
additionalProperties:
oneOf:
- type: 'null'
- type: boolean
- type: number
- type: string
- type: array
- type: object
description: >-
(Optional) Map of values to substitute in for variables in your prompt.
The substitution values can either be strings, or other Response input
types like images or files.
version:
type: string
description: >-
(Optional) Version of the prompt template.
additionalProperties: false
required:
- id
title: OpenAIResponsePrompt
description: >-
Reference to a prompt template and its variables.
OpenAIResponseReasoning:
type: object
properties:
effort:
type: string
enum:
- low
- medium
- high
- minimal
description: >-
(Optional) The effort level to use for reasoning.
generate_summary:
type: string
description: >-
Deprecated. Use the generate_summary_text field instead. (Optional) Whether
to generate a summary of the reasoning process.
summary:
type: string
additionalProperties: false
title: OpenAIResponseReasoning
description: >-
Configuration options for reasoning models.
OpenAIResponsesTool:
type: object
properties:
description:
type: string
name:
type: string
parameters:
type: object
title: object
description: >-
The base class of the class hierarchy.
When called, it accepts no arguments and returns a new featureless
instance that has no instance attributes and cannot be given any.
type:
type: string
const: function
additionalProperties: false
title: OpenAIResponsesTool
OpenAIResponsesToolChoice:
type: object
title: OpenAIResponsesToolChoice
description: >-
Type alias.
Type aliases are created through the type statement::
type Alias = int
In this example, Alias and int will be treated equivalently by static
type checkers.
At runtime, Alias is an instance of TypeAliasType. The __name__
attribute holds the name of the type alias. The value of the type alias
is stored in the __value__ attribute. It is evaluated lazily, so the
value is computed only if the attribute is accessed.
Type aliases can also be generic::
type ListOrSet[T] = list[T] | set[T]
In this case, the type parameters of the alias are stored in the
__type_params__ attribute.
See PEP 695 for more information.
OpenAIResponseContentPart:
oneOf:
- $ref: '#/components/schemas/OpenAIResponseContentPartOutputText'
@ -9464,6 +9645,36 @@ components:
id:
type: string
description: Unique identifier for this response
incomplete_details:
$ref: '#/components/schemas/OpenAIResponseIncompleteDetails'
description: >-
(Optional) Incomplete details if the response is incomplete
instructions:
oneOf:
- type: string
- type: array
items:
type: string
description: >-
(Optional) A system (or developer) message inserted into the model's context.
max_output_tokens:
type: integer
description: >-
(Optional) An upper bound for the number of tokens that can be generated
for a response, including visible output tokens and reasoning tokens.
max_tool_calls:
type: integer
description: >-
(Optional) The maximum number of total calls to built-in tools that can
be processed in a response.
metadata:
type: object
additionalProperties:
type: string
description: >-
(Optional) Set of 16 key-value pairs that can be attached to an object.
This can be useful for storing additional information about the object
in a structured format, and querying for objects via API or the dashboard.
model:
type: string
description: Model identifier used for generation
@ -9488,6 +9699,28 @@ components:
type: string
description: >-
(Optional) ID of the previous response in a conversation
prompt:
$ref: '#/components/schemas/OpenAIResponsePrompt'
description: >-
(Optional) Reference to a prompt template and its variables.
prompt_cache_key:
type: string
description: >-
(Optional)Used to cache responses for similar requests to optimize your
cache hit rates. Replaces the user field.
reasoning:
$ref: '#/components/schemas/OpenAIResponseReasoning'
description: >-
(Optional) Configuration options for reasoning models.
safety_identifier:
type: string
description: >-
(Optional) A stable identifier used to help detect users of your application
that may be violating OpenAI's usage policies.
service_tier:
type: string
description: >-
(Optional) Specifies the processing type used for serving the request.
status:
type: string
description: >-
@ -9500,18 +9733,26 @@ components:
$ref: '#/components/schemas/OpenAIResponseText'
description: >-
Text formatting configuration for the response
tool_choice:
$ref: '#/components/schemas/OpenAIResponsesToolChoice'
tools:
type: array
items:
$ref: '#/components/schemas/OpenAIResponsesTool'
top_logprobs:
type: integer
top_p:
type: number
description: >-
(Optional) Nucleus sampling parameter used for generation
truncation:
type: string
description: >-
(Optional) Truncation strategy applied to the response
user:
type: string
description: >-
(Optional) User identifier associated with the request
truncation:
type: string
description: >-
(Optional) Truncation strategy applied to the response
input:
type: array
items:
@ -9527,7 +9768,6 @@ components:
- output
- parallel_tool_calls
- status
- text
- input
title: OpenAIResponseObjectWithInput
description: >-

1
docs/source/distributions/k8s-benchmark/benchmark.py
View file

@ -1,4 +1,3 @@
#!/usr/bin/env python3
# Copyright (c) Meta Platforms, Inc. and affiliates.
# All rights reserved.
#