mirror of
https://github.com/meta-llama/llama-stack.git
synced 2025-06-28 19:04:19 +00:00
making the API python based with a converter script
This commit is contained in:
Ashwin Bharambe
parent
1a2b17af7f
commit
722d20c6de
5 changed files with 1565 additions and 0 deletions
2
requirements.txt
Normal file
2
requirements.txt
Normal file
|
|
@ -0,0 +1,2 @@
|
|||
json-strong-typing
|
||||
python-openapi
|
||||
226
source/defn.py
Normal file
226
source/defn.py
Normal file
|
|
@ -0,0 +1,226 @@
|
|||
from dataclasses import dataclass, field
|
||||
from enum import Enum
|
||||
from typing import Any, Dict, List, Optional, Protocol, Union
|
||||
|
||||
import yaml
|
||||
|
||||
from pyopenapi import Info, Options, Server, Specification, webmethod
|
||||
from strong_typing.schema import json_schema_type
|
||||
|
||||
|
||||
@json_schema_type(
|
||||
schema={"type": "string", "format": "uri", "pattern": "^(https?://|file://|data:)"}
|
||||
)
|
||||
@dataclass
|
||||
class URL:
|
||||
url: str
|
||||
|
||||
def __str__(self) -> str:
|
||||
return self.url
|
||||
|
||||
|
||||
@json_schema_type
|
||||
@dataclass
|
||||
class Attachment:
|
||||
"""
|
||||
Attachments are used to refer to external resources, such as images, videos, audio, etc.
|
||||
|
||||
"""
|
||||
|
||||
url: URL
|
||||
mime_type: str
|
||||
|
||||
|
||||
Content = Union[
|
||||
str,
|
||||
Attachment,
|
||||
List[Union[str, Attachment]],
|
||||
]
|
||||
|
||||
|
||||
class Role(Enum):
|
||||
system = "system"
|
||||
user = "user"
|
||||
assistant = "assistant"
|
||||
tool = "tool"
|
||||
|
||||
|
||||
class StopReason(Enum):
|
||||
"""
|
||||
Stop reasons are used to indicate why the model stopped generating text.
|
||||
"""
|
||||
|
||||
not_stopped = "not_stopped"
|
||||
finished_ok = "finished_ok"
|
||||
max_tokens = "max_tokens"
|
||||
|
||||
|
||||
@dataclass
|
||||
class ToolCall:
|
||||
"""
|
||||
A tool call is a request to a tool.
|
||||
"""
|
||||
|
||||
tool_name: str
|
||||
arguments: Dict[str, Any]
|
||||
|
||||
|
||||
@dataclass
|
||||
class ToolResponse:
|
||||
tool_name: str
|
||||
response: str
|
||||
|
||||
|
||||
@dataclass
|
||||
class ToolDefinition:
|
||||
tool_name: str
|
||||
parameters: Dict[str, Any]
|
||||
|
||||
|
||||
@json_schema_type
|
||||
@dataclass
|
||||
class Message:
|
||||
role: Role
|
||||
|
||||
# input to the model or output from the model
|
||||
content: Content
|
||||
|
||||
# zero-shot tool definitions as input to the model
|
||||
tool_definitions: List[ToolDefinition] = field(default_factory=list)
|
||||
|
||||
# output from the model
|
||||
tool_calls: List[ToolCall] = field(default_factory=list)
|
||||
|
||||
# input to the model
|
||||
tool_responses: List[ToolResponse] = field(default_factory=list)
|
||||
|
||||
|
||||
@json_schema_type
|
||||
@dataclass
|
||||
class CompletionResponse:
|
||||
"""Normal completion response."""
|
||||
content: Content
|
||||
stop_reason: StopReason
|
||||
logprobs: Optional[Dict[str, Any]] = None
|
||||
|
||||
|
||||
@json_schema_type
|
||||
@dataclass
|
||||
class StreamedCompletionResponse:
|
||||
"""streamed completion response."""
|
||||
text_delta: str
|
||||
stop_reason: StopReason
|
||||
logprobs: Optional[Dict[str, Any]] = None
|
||||
|
||||
|
||||
@json_schema_type
|
||||
@dataclass
|
||||
class ChatCompletionResponse:
|
||||
"""Normal chat completion response."""
|
||||
|
||||
content: Content
|
||||
stop_reason: StopReason
|
||||
tool_calls: List[ToolCall] = field(default_factory=list)
|
||||
logprobs: Optional[Dict[str, Any]] = None
|
||||
|
||||
|
||||
@json_schema_type
|
||||
@dataclass
|
||||
class StreamedChatCompletionResponse:
|
||||
"""Streamed chat completion response."""
|
||||
|
||||
text_delta: str
|
||||
stop_reason: StopReason
|
||||
tool_call: Optional[ToolCall] = None
|
||||
|
||||
|
||||
@dataclass
|
||||
class SamplingParams:
|
||||
temperature: float = 0.0
|
||||
strategy: str = "greedy"
|
||||
top_p: float = 0.95
|
||||
top_k: int = 0
|
||||
|
||||
|
||||
class PretrainedModel(Enum):
|
||||
llama3_8b = "llama3_8b"
|
||||
llama3_70b = "llama3_70b"
|
||||
|
||||
|
||||
class InstructModel(Enum):
|
||||
llama3_8b_chat = "llama3_8b_chat"
|
||||
llama3_70b_chat = "llama3_70b_chat"
|
||||
|
||||
|
||||
@json_schema_type
|
||||
@dataclass
|
||||
class CompletionRequest:
|
||||
content: Content
|
||||
model: PretrainedModel = PretrainedModel.llama3_8b
|
||||
sampling_params: SamplingParams = SamplingParams()
|
||||
max_tokens: int = 0
|
||||
stream: bool = False
|
||||
logprobs: bool = False
|
||||
|
||||
|
||||
@json_schema_type
|
||||
@dataclass
|
||||
class ChatCompletionRequest:
|
||||
message: Message
|
||||
message_history: List[Message] = None
|
||||
model: InstructModel = InstructModel.llama3_8b_chat
|
||||
sampling_params: SamplingParams = SamplingParams()
|
||||
max_tokens: int = 0
|
||||
stream: bool = False
|
||||
logprobs: bool = False
|
||||
|
||||
|
||||
class Inference(Protocol):
|
||||
|
||||
def post_completion(
|
||||
self,
|
||||
request: CompletionRequest,
|
||||
) -> Union[CompletionResponse, StreamedCompletionResponse]: ...
|
||||
|
||||
def post_chat_completion(
|
||||
self,
|
||||
request: ChatCompletionRequest,
|
||||
) -> Union[ChatCompletionResponse, StreamedChatCompletionResponse]: ...
|
||||
|
||||
|
||||
|
||||
@json_schema_type
|
||||
@dataclass
|
||||
class AgenticSystemExecuteRequest:
|
||||
message: Message
|
||||
message_history: List[Message] = None
|
||||
model: InstructModel = InstructModel.llama3_8b_chat
|
||||
sampling_params: SamplingParams = SamplingParams()
|
||||
|
||||
class AgenticSystem(Protocol):
|
||||
|
||||
@webmethod(route="/agentic/system/execute")
|
||||
def create_agentic_system_execute(self,) -> str: ...
|
||||
|
||||
|
||||
class Endpoint(Inference, AgenticSystem): ...
|
||||
|
||||
|
||||
if __name__ == "__main__":
|
||||
print("Converting the spec to YAML (openapi.yaml) and HTML (openapi.html)")
|
||||
spec = Specification(
|
||||
Endpoint,
|
||||
Options(
|
||||
server=Server(url="http://llama.meta.com"),
|
||||
info=Info(
|
||||
title="Llama Stack specification",
|
||||
version="0.1",
|
||||
description="This is the llama stack",
|
||||
),
|
||||
),
|
||||
)
|
||||
with open("openapi.yaml", "w", encoding="utf-8") as fp:
|
||||
yaml.dump(spec.get_json(), fp, allow_unicode=True)
|
||||
|
||||
with open("openapi.html", "w") as fp:
|
||||
spec.write_html(fp, pretty_print=True)
|
||||
842
source/openapi.html
Normal file
842
source/openapi.html
Normal file
|
|
@ -0,0 +1,842 @@
|
|||
<!DOCTYPE html>
|
||||
<html>
|
||||
|
||||
<head>
|
||||
<meta charset="utf-8" />
|
||||
<meta name="viewport" content="width=device-width, initial-scale=1">
|
||||
<title>OpenAPI specification</title>
|
||||
<link href="https://fonts.googleapis.com/css?family=Montserrat:300,400,700|Roboto:300,400,700" rel="stylesheet">
|
||||
<style>
|
||||
body {
|
||||
margin: 0;
|
||||
padding: 0;
|
||||
}
|
||||
|
||||
</style>
|
||||
<script defer="defer" src="https://cdn.redoc.ly/redoc/latest/bundles/redoc.standalone.js"></script>
|
||||
<script defer="defer">
|
||||
document.addEventListener("DOMContentLoaded", function () {
|
||||
spec = {
|
||||
"openapi": "3.1.0",
|
||||
"info": {
|
||||
"title": "Llama Stack specification",
|
||||
"version": "0.1",
|
||||
"description": "This is the llama stack"
|
||||
},
|
||||
"servers": [
|
||||
{
|
||||
"url": "http://llama.meta.com"
|
||||
}
|
||||
],
|
||||
"paths": {
|
||||
"/agentic/system/execute": {
|
||||
"get": {
|
||||
"responses": {
|
||||
"200": {
|
||||
"description": "OK",
|
||||
"content": {
|
||||
"application/json": {
|
||||
"schema": {
|
||||
"type": "string"
|
||||
}
|
||||
}
|
||||
}
|
||||
}
|
||||
},
|
||||
"tags": [
|
||||
"AgenticSystem"
|
||||
],
|
||||
"parameters": []
|
||||
}
|
||||
},
|
||||
"/chat_completion": {
|
||||
"post": {
|
||||
"responses": {
|
||||
"200": {
|
||||
"description": "Normal chat completion response. **OR** Streamed chat completion response.",
|
||||
"content": {
|
||||
"application/json": {
|
||||
"schema": {
|
||||
"oneOf": [
|
||||
{
|
||||
"$ref": "#/components/schemas/ChatCompletionResponse"
|
||||
},
|
||||
{
|
||||
"$ref": "#/components/schemas/StreamedChatCompletionResponse"
|
||||
}
|
||||
]
|
||||
}
|
||||
}
|
||||
}
|
||||
}
|
||||
},
|
||||
"tags": [
|
||||
"Inference"
|
||||
],
|
||||
"parameters": [],
|
||||
"requestBody": {
|
||||
"content": {
|
||||
"application/json": {
|
||||
"schema": {
|
||||
"$ref": "#/components/schemas/ChatCompletionRequest"
|
||||
}
|
||||
}
|
||||
},
|
||||
"required": true
|
||||
}
|
||||
}
|
||||
},
|
||||
"/completion": {
|
||||
"post": {
|
||||
"responses": {
|
||||
"200": {
|
||||
"description": "Normal completion response. **OR** streamed completion response.",
|
||||
"content": {
|
||||
"application/json": {
|
||||
"schema": {
|
||||
"oneOf": [
|
||||
{
|
||||
"$ref": "#/components/schemas/CompletionResponse"
|
||||
},
|
||||
{
|
||||
"$ref": "#/components/schemas/StreamedCompletionResponse"
|
||||
}
|
||||
]
|
||||
}
|
||||
}
|
||||
}
|
||||
}
|
||||
},
|
||||
"tags": [
|
||||
"Inference"
|
||||
],
|
||||
"parameters": [],
|
||||
"requestBody": {
|
||||
"content": {
|
||||
"application/json": {
|
||||
"schema": {
|
||||
"$ref": "#/components/schemas/CompletionRequest"
|
||||
}
|
||||
}
|
||||
},
|
||||
"required": true
|
||||
}
|
||||
}
|
||||
}
|
||||
},
|
||||
"jsonSchemaDialect": "https://json-schema.org/draft/2020-12/schema",
|
||||
"components": {
|
||||
"schemas": {
|
||||
"Attachment": {
|
||||
"type": "object",
|
||||
"properties": {
|
||||
"url": {
|
||||
"$ref": "#/components/schemas/URL"
|
||||
},
|
||||
"mime_type": {
|
||||
"type": "string"
|
||||
}
|
||||
},
|
||||
"additionalProperties": false,
|
||||
"required": [
|
||||
"url",
|
||||
"mime_type"
|
||||
],
|
||||
"title": "Attachments are used to refer to external resources, such as images, videos, audio, etc."
|
||||
},
|
||||
"ChatCompletionRequest": {
|
||||
"type": "object",
|
||||
"properties": {
|
||||
"message": {
|
||||
"$ref": "#/components/schemas/Message"
|
||||
},
|
||||
"message_history": {
|
||||
"type": "array",
|
||||
"items": {
|
||||
"$ref": "#/components/schemas/Message"
|
||||
}
|
||||
},
|
||||
"model": {
|
||||
"type": "string",
|
||||
"enum": [
|
||||
"llama3_8b_chat",
|
||||
"llama3_70b_chat"
|
||||
],
|
||||
"default": "llama3_8b_chat"
|
||||
},
|
||||
"sampling_params": {
|
||||
"type": "object",
|
||||
"properties": {
|
||||
"temperature": {
|
||||
"type": "number",
|
||||
"default": 0.0
|
||||
},
|
||||
"strategy": {
|
||||
"type": "string",
|
||||
"default": "greedy"
|
||||
},
|
||||
"top_p": {
|
||||
"type": "number",
|
||||
"default": 0.95
|
||||
},
|
||||
"top_k": {
|
||||
"type": "integer",
|
||||
"default": 0
|
||||
}
|
||||
},
|
||||
"additionalProperties": false,
|
||||
"required": [
|
||||
"temperature",
|
||||
"strategy",
|
||||
"top_p",
|
||||
"top_k"
|
||||
]
|
||||
},
|
||||
"max_tokens": {
|
||||
"type": "integer",
|
||||
"default": 0
|
||||
},
|
||||
"stream": {
|
||||
"type": "boolean",
|
||||
"default": false
|
||||
},
|
||||
"logprobs": {
|
||||
"type": "boolean",
|
||||
"default": false
|
||||
}
|
||||
},
|
||||
"additionalProperties": false,
|
||||
"required": [
|
||||
"message",
|
||||
"message_history",
|
||||
"model",
|
||||
"sampling_params",
|
||||
"max_tokens",
|
||||
"stream",
|
||||
"logprobs"
|
||||
]
|
||||
},
|
||||
"Message": {
|
||||
"type": "object",
|
||||
"properties": {
|
||||
"role": {
|
||||
"type": "string",
|
||||
"enum": [
|
||||
"system",
|
||||
"user",
|
||||
"assistant",
|
||||
"tool"
|
||||
]
|
||||
},
|
||||
"content": {
|
||||
"oneOf": [
|
||||
{
|
||||
"type": "string"
|
||||
},
|
||||
{
|
||||
"$ref": "#/components/schemas/Attachment"
|
||||
},
|
||||
{
|
||||
"type": "array",
|
||||
"items": {
|
||||
"oneOf": [
|
||||
{
|
||||
"type": "string"
|
||||
},
|
||||
{
|
||||
"$ref": "#/components/schemas/Attachment"
|
||||
}
|
||||
]
|
||||
}
|
||||
}
|
||||
]
|
||||
},
|
||||
"tool_definitions": {
|
||||
"type": "array",
|
||||
"items": {
|
||||
"type": "object",
|
||||
"properties": {
|
||||
"tool_name": {
|
||||
"type": "string"
|
||||
},
|
||||
"parameters": {
|
||||
"type": "object",
|
||||
"additionalProperties": {
|
||||
"oneOf": [
|
||||
{
|
||||
"type": "null"
|
||||
},
|
||||
{
|
||||
"type": "boolean"
|
||||
},
|
||||
{
|
||||
"type": "number"
|
||||
},
|
||||
{
|
||||
"type": "string"
|
||||
},
|
||||
{
|
||||
"type": "array"
|
||||
},
|
||||
{
|
||||
"type": "object"
|
||||
}
|
||||
]
|
||||
}
|
||||
}
|
||||
},
|
||||
"additionalProperties": false,
|
||||
"required": [
|
||||
"tool_name",
|
||||
"parameters"
|
||||
]
|
||||
}
|
||||
},
|
||||
"tool_calls": {
|
||||
"type": "array",
|
||||
"items": {
|
||||
"type": "object",
|
||||
"properties": {
|
||||
"tool_name": {
|
||||
"type": "string"
|
||||
},
|
||||
"arguments": {
|
||||
"type": "object",
|
||||
"additionalProperties": {
|
||||
"oneOf": [
|
||||
{
|
||||
"type": "null"
|
||||
},
|
||||
{
|
||||
"type": "boolean"
|
||||
},
|
||||
{
|
||||
"type": "number"
|
||||
},
|
||||
{
|
||||
"type": "string"
|
||||
},
|
||||
{
|
||||
"type": "array"
|
||||
},
|
||||
{
|
||||
"type": "object"
|
||||
}
|
||||
]
|
||||
}
|
||||
}
|
||||
},
|
||||
"additionalProperties": false,
|
||||
"required": [
|
||||
"tool_name",
|
||||
"arguments"
|
||||
],
|
||||
"title": "A tool call is a request to a tool."
|
||||
}
|
||||
},
|
||||
"tool_responses": {
|
||||
"type": "array",
|
||||
"items": {
|
||||
"type": "object",
|
||||
"properties": {
|
||||
"tool_name": {
|
||||
"type": "string"
|
||||
},
|
||||
"response": {
|
||||
"type": "string"
|
||||
}
|
||||
},
|
||||
"additionalProperties": false,
|
||||
"required": [
|
||||
"tool_name",
|
||||
"response"
|
||||
]
|
||||
}
|
||||
}
|
||||
},
|
||||
"additionalProperties": false,
|
||||
"required": [
|
||||
"role",
|
||||
"content",
|
||||
"tool_definitions",
|
||||
"tool_calls",
|
||||
"tool_responses"
|
||||
]
|
||||
},
|
||||
"URL": {
|
||||
"type": "string",
|
||||
"format": "uri",
|
||||
"pattern": "^(https?://|file://|data:)"
|
||||
},
|
||||
"ChatCompletionResponse": {
|
||||
"type": "object",
|
||||
"properties": {
|
||||
"content": {
|
||||
"oneOf": [
|
||||
{
|
||||
"type": "string"
|
||||
},
|
||||
{
|
||||
"$ref": "#/components/schemas/Attachment"
|
||||
},
|
||||
{
|
||||
"type": "array",
|
||||
"items": {
|
||||
"oneOf": [
|
||||
{
|
||||
"type": "string"
|
||||
},
|
||||
{
|
||||
"$ref": "#/components/schemas/Attachment"
|
||||
}
|
||||
]
|
||||
}
|
||||
}
|
||||
]
|
||||
},
|
||||
"stop_reason": {
|
||||
"type": "string",
|
||||
"enum": [
|
||||
"not_stopped",
|
||||
"finished_ok",
|
||||
"max_tokens"
|
||||
],
|
||||
"title": "Stop reasons are used to indicate why the model stopped generating text."
|
||||
},
|
||||
"tool_calls": {
|
||||
"type": "array",
|
||||
"items": {
|
||||
"type": "object",
|
||||
"properties": {
|
||||
"tool_name": {
|
||||
"type": "string"
|
||||
},
|
||||
"arguments": {
|
||||
"type": "object",
|
||||
"additionalProperties": {
|
||||
"oneOf": [
|
||||
{
|
||||
"type": "null"
|
||||
},
|
||||
{
|
||||
"type": "boolean"
|
||||
},
|
||||
{
|
||||
"type": "number"
|
||||
},
|
||||
{
|
||||
"type": "string"
|
||||
},
|
||||
{
|
||||
"type": "array"
|
||||
},
|
||||
{
|
||||
"type": "object"
|
||||
}
|
||||
]
|
||||
}
|
||||
}
|
||||
},
|
||||
"additionalProperties": false,
|
||||
"required": [
|
||||
"tool_name",
|
||||
"arguments"
|
||||
],
|
||||
"title": "A tool call is a request to a tool."
|
||||
}
|
||||
},
|
||||
"logprobs": {
|
||||
"type": "object",
|
||||
"additionalProperties": {
|
||||
"oneOf": [
|
||||
{
|
||||
"type": "null"
|
||||
},
|
||||
{
|
||||
"type": "boolean"
|
||||
},
|
||||
{
|
||||
"type": "number"
|
||||
},
|
||||
{
|
||||
"type": "string"
|
||||
},
|
||||
{
|
||||
"type": "array"
|
||||
},
|
||||
{
|
||||
"type": "object"
|
||||
}
|
||||
]
|
||||
}
|
||||
}
|
||||
},
|
||||
"additionalProperties": false,
|
||||
"required": [
|
||||
"content",
|
||||
"stop_reason",
|
||||
"tool_calls"
|
||||
],
|
||||
"title": "Normal chat completion response."
|
||||
},
|
||||
"StreamedChatCompletionResponse": {
|
||||
"type": "object",
|
||||
"properties": {
|
||||
"text_delta": {
|
||||
"type": "string"
|
||||
},
|
||||
"stop_reason": {
|
||||
"type": "string",
|
||||
"enum": [
|
||||
"not_stopped",
|
||||
"finished_ok",
|
||||
"max_tokens"
|
||||
],
|
||||
"title": "Stop reasons are used to indicate why the model stopped generating text."
|
||||
},
|
||||
"tool_call": {
|
||||
"type": "object",
|
||||
"properties": {
|
||||
"tool_name": {
|
||||
"type": "string"
|
||||
},
|
||||
"arguments": {
|
||||
"type": "object",
|
||||
"additionalProperties": {
|
||||
"oneOf": [
|
||||
{
|
||||
"type": "null"
|
||||
},
|
||||
{
|
||||
"type": "boolean"
|
||||
},
|
||||
{
|
||||
"type": "number"
|
||||
},
|
||||
{
|
||||
"type": "string"
|
||||
},
|
||||
{
|
||||
"type": "array"
|
||||
},
|
||||
{
|
||||
"type": "object"
|
||||
}
|
||||
]
|
||||
}
|
||||
}
|
||||
},
|
||||
"additionalProperties": false,
|
||||
"required": [
|
||||
"tool_name",
|
||||
"arguments"
|
||||
],
|
||||
"title": "A tool call is a request to a tool."
|
||||
}
|
||||
},
|
||||
"additionalProperties": false,
|
||||
"required": [
|
||||
"text_delta",
|
||||
"stop_reason"
|
||||
],
|
||||
"title": "Streamed chat completion response."
|
||||
},
|
||||
"CompletionRequest": {
|
||||
"type": "object",
|
||||
"properties": {
|
||||
"content": {
|
||||
"oneOf": [
|
||||
{
|
||||
"type": "string"
|
||||
},
|
||||
{
|
||||
"$ref": "#/components/schemas/Attachment"
|
||||
},
|
||||
{
|
||||
"type": "array",
|
||||
"items": {
|
||||
"oneOf": [
|
||||
{
|
||||
"type": "string"
|
||||
},
|
||||
{
|
||||
"$ref": "#/components/schemas/Attachment"
|
||||
}
|
||||
]
|
||||
}
|
||||
}
|
||||
]
|
||||
},
|
||||
"model": {
|
||||
"type": "string",
|
||||
"enum": [
|
||||
"llama3_8b",
|
||||
"llama3_70b"
|
||||
],
|
||||
"default": "llama3_8b"
|
||||
},
|
||||
"sampling_params": {
|
||||
"type": "object",
|
||||
"properties": {
|
||||
"temperature": {
|
||||
"type": "number",
|
||||
"default": 0.0
|
||||
},
|
||||
"strategy": {
|
||||
"type": "string",
|
||||
"default": "greedy"
|
||||
},
|
||||
"top_p": {
|
||||
"type": "number",
|
||||
"default": 0.95
|
||||
},
|
||||
"top_k": {
|
||||
"type": "integer",
|
||||
"default": 0
|
||||
}
|
||||
},
|
||||
"additionalProperties": false,
|
||||
"required": [
|
||||
"temperature",
|
||||
"strategy",
|
||||
"top_p",
|
||||
"top_k"
|
||||
]
|
||||
},
|
||||
"max_tokens": {
|
||||
"type": "integer",
|
||||
"default": 0
|
||||
},
|
||||
"stream": {
|
||||
"type": "boolean",
|
||||
"default": false
|
||||
},
|
||||
"logprobs": {
|
||||
"type": "boolean",
|
||||
"default": false
|
||||
}
|
||||
},
|
||||
"additionalProperties": false,
|
||||
"required": [
|
||||
"content",
|
||||
"model",
|
||||
"sampling_params",
|
||||
"max_tokens",
|
||||
"stream",
|
||||
"logprobs"
|
||||
]
|
||||
},
|
||||
"CompletionResponse": {
|
||||
"type": "object",
|
||||
"properties": {
|
||||
"content": {
|
||||
"oneOf": [
|
||||
{
|
||||
"type": "string"
|
||||
},
|
||||
{
|
||||
"$ref": "#/components/schemas/Attachment"
|
||||
},
|
||||
{
|
||||
"type": "array",
|
||||
"items": {
|
||||
"oneOf": [
|
||||
{
|
||||
"type": "string"
|
||||
},
|
||||
{
|
||||
"$ref": "#/components/schemas/Attachment"
|
||||
}
|
||||
]
|
||||
}
|
||||
}
|
||||
]
|
||||
},
|
||||
"stop_reason": {
|
||||
"type": "string",
|
||||
"enum": [
|
||||
"not_stopped",
|
||||
"finished_ok",
|
||||
"max_tokens"
|
||||
],
|
||||
"title": "Stop reasons are used to indicate why the model stopped generating text."
|
||||
},
|
||||
"logprobs": {
|
||||
"type": "object",
|
||||
"additionalProperties": {
|
||||
"oneOf": [
|
||||
{
|
||||
"type": "null"
|
||||
},
|
||||
{
|
||||
"type": "boolean"
|
||||
},
|
||||
{
|
||||
"type": "number"
|
||||
},
|
||||
{
|
||||
"type": "string"
|
||||
},
|
||||
{
|
||||
"type": "array"
|
||||
},
|
||||
{
|
||||
"type": "object"
|
||||
}
|
||||
]
|
||||
}
|
||||
}
|
||||
},
|
||||
"additionalProperties": false,
|
||||
"required": [
|
||||
"content",
|
||||
"stop_reason"
|
||||
],
|
||||
"title": "Normal completion response."
|
||||
},
|
||||
"StreamedCompletionResponse": {
|
||||
"type": "object",
|
||||
"properties": {
|
||||
"text_delta": {
|
||||
"type": "string"
|
||||
},
|
||||
"stop_reason": {
|
||||
"type": "string",
|
||||
"enum": [
|
||||
"not_stopped",
|
||||
"finished_ok",
|
||||
"max_tokens"
|
||||
],
|
||||
"title": "Stop reasons are used to indicate why the model stopped generating text."
|
||||
},
|
||||
"logprobs": {
|
||||
"type": "object",
|
||||
"additionalProperties": {
|
||||
"oneOf": [
|
||||
{
|
||||
"type": "null"
|
||||
},
|
||||
{
|
||||
"type": "boolean"
|
||||
},
|
||||
{
|
||||
"type": "number"
|
||||
},
|
||||
{
|
||||
"type": "string"
|
||||
},
|
||||
{
|
||||
"type": "array"
|
||||
},
|
||||
{
|
||||
"type": "object"
|
||||
}
|
||||
]
|
||||
}
|
||||
}
|
||||
},
|
||||
"additionalProperties": false,
|
||||
"required": [
|
||||
"text_delta",
|
||||
"stop_reason"
|
||||
],
|
||||
"title": "streamed completion response."
|
||||
}
|
||||
},
|
||||
"responses": {}
|
||||
},
|
||||
"security": [
|
||||
{
|
||||
"Default": []
|
||||
}
|
||||
],
|
||||
"tags": [
|
||||
{
|
||||
"name": "AgenticSystem"
|
||||
},
|
||||
{
|
||||
"name": "Inference"
|
||||
},
|
||||
{
|
||||
"name": "Attachment",
|
||||
"description": "Attachments are used to refer to external resources, such as images, videos, audio, etc.\n\n<SchemaDefinition schemaRef=\"#/components/schemas/Attachment\" />"
|
||||
},
|
||||
{
|
||||
"name": "ChatCompletionRequest",
|
||||
"description": "<SchemaDefinition schemaRef=\"#/components/schemas/ChatCompletionRequest\" />"
|
||||
},
|
||||
{
|
||||
"name": "Message",
|
||||
"description": "<SchemaDefinition schemaRef=\"#/components/schemas/Message\" />"
|
||||
},
|
||||
{
|
||||
"name": "URL",
|
||||
"description": "<SchemaDefinition schemaRef=\"#/components/schemas/URL\" />"
|
||||
},
|
||||
{
|
||||
"name": "ChatCompletionResponse",
|
||||
"description": "Normal chat completion response.\n\n<SchemaDefinition schemaRef=\"#/components/schemas/ChatCompletionResponse\" />"
|
||||
},
|
||||
{
|
||||
"name": "StreamedChatCompletionResponse",
|
||||
"description": "Streamed chat completion response.\n\n<SchemaDefinition schemaRef=\"#/components/schemas/StreamedChatCompletionResponse\" />"
|
||||
},
|
||||
{
|
||||
"name": "CompletionRequest",
|
||||
"description": "<SchemaDefinition schemaRef=\"#/components/schemas/CompletionRequest\" />"
|
||||
},
|
||||
{
|
||||
"name": "CompletionResponse",
|
||||
"description": "Normal completion response.\n\n<SchemaDefinition schemaRef=\"#/components/schemas/CompletionResponse\" />"
|
||||
},
|
||||
{
|
||||
"name": "StreamedCompletionResponse",
|
||||
"description": "streamed completion response.\n\n<SchemaDefinition schemaRef=\"#/components/schemas/StreamedCompletionResponse\" />"
|
||||
}
|
||||
],
|
||||
"x-tagGroups": [
|
||||
{
|
||||
"name": "Operations",
|
||||
"tags": [
|
||||
"AgenticSystem",
|
||||
"Inference"
|
||||
]
|
||||
},
|
||||
{
|
||||
"name": "Types",
|
||||
"tags": [
|
||||
"Attachment",
|
||||
"ChatCompletionRequest",
|
||||
"ChatCompletionResponse",
|
||||
"CompletionRequest",
|
||||
"CompletionResponse",
|
||||
"Message",
|
||||
"StreamedChatCompletionResponse",
|
||||
"StreamedCompletionResponse",
|
||||
"URL"
|
||||
]
|
||||
}
|
||||
]
|
||||
};
|
||||
options = {
|
||||
downloadFileName: "openapi.json",
|
||||
expandResponses: "200",
|
||||
expandSingleSchemaField: true,
|
||||
jsonSampleExpandLevel: "all",
|
||||
schemaExpansionLevel: "all",
|
||||
};
|
||||
element = document.getElementById("openapi-container");
|
||||
Redoc.init(spec, options, element);
|
||||
|
||||
if (spec.info && spec.info.title) {
|
||||
document.title = spec.info.title;
|
||||
}
|
||||
});
|
||||
</script>
|
||||
</head>
|
||||
|
||||
<body>
|
||||
<div id="openapi-container"></div>
|
||||
</body>
|
||||
|
||||
</html>
|
||||
492
source/openapi.yaml
Normal file
492
source/openapi.yaml
Normal file
|
|
@ -0,0 +1,492 @@
|
|||
components:
|
||||
responses: {}
|
||||
schemas:
|
||||
Attachment:
|
||||
additionalProperties: false
|
||||
properties:
|
||||
mime_type:
|
||||
type: string
|
||||
url:
|
||||
$ref: '#/components/schemas/URL'
|
||||
required:
|
||||
- url
|
||||
- mime_type
|
||||
title: Attachments are used to refer to external resources, such as images,
|
||||
videos, audio, etc.
|
||||
type: object
|
||||
ChatCompletionRequest:
|
||||
additionalProperties: false
|
||||
properties:
|
||||
logprobs:
|
||||
default: false
|
||||
type: boolean
|
||||
max_tokens:
|
||||
default: 0
|
||||
type: integer
|
||||
message:
|
||||
$ref: '#/components/schemas/Message'
|
||||
message_history:
|
||||
items:
|
||||
$ref: '#/components/schemas/Message'
|
||||
type: array
|
||||
model:
|
||||
default: llama3_8b_chat
|
||||
enum:
|
||||
- llama3_8b_chat
|
||||
- llama3_70b_chat
|
||||
type: string
|
||||
sampling_params:
|
||||
additionalProperties: false
|
||||
properties:
|
||||
strategy:
|
||||
default: greedy
|
||||
type: string
|
||||
temperature:
|
||||
default: 0.0
|
||||
type: number
|
||||
top_k:
|
||||
default: 0
|
||||
type: integer
|
||||
top_p:
|
||||
default: 0.95
|
||||
type: number
|
||||
required:
|
||||
- temperature
|
||||
- strategy
|
||||
- top_p
|
||||
- top_k
|
||||
type: object
|
||||
stream:
|
||||
default: false
|
||||
type: boolean
|
||||
required:
|
||||
- message
|
||||
- message_history
|
||||
- model
|
||||
- sampling_params
|
||||
- max_tokens
|
||||
- stream
|
||||
- logprobs
|
||||
type: object
|
||||
ChatCompletionResponse:
|
||||
additionalProperties: false
|
||||
properties:
|
||||
content:
|
||||
oneOf:
|
||||
- type: string
|
||||
- $ref: '#/components/schemas/Attachment'
|
||||
- items:
|
||||
oneOf:
|
||||
- type: string
|
||||
- $ref: '#/components/schemas/Attachment'
|
||||
type: array
|
||||
logprobs:
|
||||
additionalProperties:
|
||||
oneOf:
|
||||
- type: 'null'
|
||||
- type: boolean
|
||||
- type: number
|
||||
- type: string
|
||||
- type: array
|
||||
- type: object
|
||||
type: object
|
||||
stop_reason:
|
||||
enum:
|
||||
- not_stopped
|
||||
- finished_ok
|
||||
- max_tokens
|
||||
title: Stop reasons are used to indicate why the model stopped generating
|
||||
text.
|
||||
type: string
|
||||
tool_calls:
|
||||
items:
|
||||
additionalProperties: false
|
||||
properties:
|
||||
arguments:
|
||||
additionalProperties:
|
||||
oneOf:
|
||||
- type: 'null'
|
||||
- type: boolean
|
||||
- type: number
|
||||
- type: string
|
||||
- type: array
|
||||
- type: object
|
||||
type: object
|
||||
tool_name:
|
||||
type: string
|
||||
required:
|
||||
- tool_name
|
||||
- arguments
|
||||
title: A tool call is a request to a tool.
|
||||
type: object
|
||||
type: array
|
||||
required:
|
||||
- content
|
||||
- stop_reason
|
||||
- tool_calls
|
||||
title: Normal chat completion response.
|
||||
type: object
|
||||
CompletionRequest:
|
||||
additionalProperties: false
|
||||
properties:
|
||||
content:
|
||||
oneOf:
|
||||
- type: string
|
||||
- $ref: '#/components/schemas/Attachment'
|
||||
- items:
|
||||
oneOf:
|
||||
- type: string
|
||||
- $ref: '#/components/schemas/Attachment'
|
||||
type: array
|
||||
logprobs:
|
||||
default: false
|
||||
type: boolean
|
||||
max_tokens:
|
||||
default: 0
|
||||
type: integer
|
||||
model:
|
||||
default: llama3_8b
|
||||
enum:
|
||||
- llama3_8b
|
||||
- llama3_70b
|
||||
type: string
|
||||
sampling_params:
|
||||
additionalProperties: false
|
||||
properties:
|
||||
strategy:
|
||||
default: greedy
|
||||
type: string
|
||||
temperature:
|
||||
default: 0.0
|
||||
type: number
|
||||
top_k:
|
||||
default: 0
|
||||
type: integer
|
||||
top_p:
|
||||
default: 0.95
|
||||
type: number
|
||||
required:
|
||||
- temperature
|
||||
- strategy
|
||||
- top_p
|
||||
- top_k
|
||||
type: object
|
||||
stream:
|
||||
default: false
|
||||
type: boolean
|
||||
required:
|
||||
- content
|
||||
- model
|
||||
- sampling_params
|
||||
- max_tokens
|
||||
- stream
|
||||
- logprobs
|
||||
type: object
|
||||
CompletionResponse:
|
||||
additionalProperties: false
|
||||
properties:
|
||||
content:
|
||||
oneOf:
|
||||
- type: string
|
||||
- $ref: '#/components/schemas/Attachment'
|
||||
- items:
|
||||
oneOf:
|
||||
- type: string
|
||||
- $ref: '#/components/schemas/Attachment'
|
||||
type: array
|
||||
logprobs:
|
||||
additionalProperties:
|
||||
oneOf:
|
||||
- type: 'null'
|
||||
- type: boolean
|
||||
- type: number
|
||||
- type: string
|
||||
- type: array
|
||||
- type: object
|
||||
type: object
|
||||
stop_reason:
|
||||
enum:
|
||||
- not_stopped
|
||||
- finished_ok
|
||||
- max_tokens
|
||||
title: Stop reasons are used to indicate why the model stopped generating
|
||||
text.
|
||||
type: string
|
||||
required:
|
||||
- content
|
||||
- stop_reason
|
||||
title: Normal completion response.
|
||||
type: object
|
||||
Message:
|
||||
additionalProperties: false
|
||||
properties:
|
||||
content:
|
||||
oneOf:
|
||||
- type: string
|
||||
- $ref: '#/components/schemas/Attachment'
|
||||
- items:
|
||||
oneOf:
|
||||
- type: string
|
||||
- $ref: '#/components/schemas/Attachment'
|
||||
type: array
|
||||
role:
|
||||
enum:
|
||||
- system
|
||||
- user
|
||||
- assistant
|
||||
- tool
|
||||
type: string
|
||||
tool_calls:
|
||||
items:
|
||||
additionalProperties: false
|
||||
properties:
|
||||
arguments:
|
||||
additionalProperties:
|
||||
oneOf:
|
||||
- type: 'null'
|
||||
- type: boolean
|
||||
- type: number
|
||||
- type: string
|
||||
- type: array
|
||||
- type: object
|
||||
type: object
|
||||
tool_name:
|
||||
type: string
|
||||
required:
|
||||
- tool_name
|
||||
- arguments
|
||||
title: A tool call is a request to a tool.
|
||||
type: object
|
||||
type: array
|
||||
tool_definitions:
|
||||
items:
|
||||
additionalProperties: false
|
||||
properties:
|
||||
parameters:
|
||||
additionalProperties:
|
||||
oneOf:
|
||||
- type: 'null'
|
||||
- type: boolean
|
||||
- type: number
|
||||
- type: string
|
||||
- type: array
|
||||
- type: object
|
||||
type: object
|
||||
tool_name:
|
||||
type: string
|
||||
required:
|
||||
- tool_name
|
||||
- parameters
|
||||
type: object
|
||||
type: array
|
||||
tool_responses:
|
||||
items:
|
||||
additionalProperties: false
|
||||
properties:
|
||||
response:
|
||||
type: string
|
||||
tool_name:
|
||||
type: string
|
||||
required:
|
||||
- tool_name
|
||||
- response
|
||||
type: object
|
||||
type: array
|
||||
required:
|
||||
- role
|
||||
- content
|
||||
- tool_definitions
|
||||
- tool_calls
|
||||
- tool_responses
|
||||
type: object
|
||||
StreamedChatCompletionResponse:
|
||||
additionalProperties: false
|
||||
properties:
|
||||
stop_reason:
|
||||
enum:
|
||||
- not_stopped
|
||||
- finished_ok
|
||||
- max_tokens
|
||||
title: Stop reasons are used to indicate why the model stopped generating
|
||||
text.
|
||||
type: string
|
||||
text_delta:
|
||||
type: string
|
||||
tool_call:
|
||||
additionalProperties: false
|
||||
properties:
|
||||
arguments:
|
||||
additionalProperties:
|
||||
oneOf:
|
||||
- type: 'null'
|
||||
- type: boolean
|
||||
- type: number
|
||||
- type: string
|
||||
- type: array
|
||||
- type: object
|
||||
type: object
|
||||
tool_name:
|
||||
type: string
|
||||
required:
|
||||
- tool_name
|
||||
- arguments
|
||||
title: A tool call is a request to a tool.
|
||||
type: object
|
||||
required:
|
||||
- text_delta
|
||||
- stop_reason
|
||||
title: Streamed chat completion response.
|
||||
type: object
|
||||
StreamedCompletionResponse:
|
||||
additionalProperties: false
|
||||
properties:
|
||||
logprobs:
|
||||
additionalProperties:
|
||||
oneOf:
|
||||
- type: 'null'
|
||||
- type: boolean
|
||||
- type: number
|
||||
- type: string
|
||||
- type: array
|
||||
- type: object
|
||||
type: object
|
||||
stop_reason:
|
||||
enum:
|
||||
- not_stopped
|
||||
- finished_ok
|
||||
- max_tokens
|
||||
title: Stop reasons are used to indicate why the model stopped generating
|
||||
text.
|
||||
type: string
|
||||
text_delta:
|
||||
type: string
|
||||
required:
|
||||
- text_delta
|
||||
- stop_reason
|
||||
title: streamed completion response.
|
||||
type: object
|
||||
URL:
|
||||
format: uri
|
||||
pattern: ^(https?://|file://|data:)
|
||||
type: string
|
||||
info:
|
||||
description: This is the llama stack
|
||||
title: Llama Stack specification
|
||||
version: '0.1'
|
||||
jsonSchemaDialect: https://json-schema.org/draft/2020-12/schema
|
||||
openapi: 3.1.0
|
||||
paths:
|
||||
/agentic/system/execute:
|
||||
get:
|
||||
parameters: []
|
||||
responses:
|
||||
'200':
|
||||
content:
|
||||
application/json:
|
||||
schema:
|
||||
type: string
|
||||
description: OK
|
||||
tags:
|
||||
- AgenticSystem
|
||||
/chat_completion:
|
||||
post:
|
||||
parameters: []
|
||||
requestBody:
|
||||
content:
|
||||
application/json:
|
||||
schema:
|
||||
$ref: '#/components/schemas/ChatCompletionRequest'
|
||||
required: true
|
||||
responses:
|
||||
'200':
|
||||
content:
|
||||
application/json:
|
||||
schema:
|
||||
oneOf:
|
||||
- $ref: '#/components/schemas/ChatCompletionResponse'
|
||||
- $ref: '#/components/schemas/StreamedChatCompletionResponse'
|
||||
description: Normal chat completion response. **OR** Streamed chat completion
|
||||
response.
|
||||
tags:
|
||||
- Inference
|
||||
/completion:
|
||||
post:
|
||||
parameters: []
|
||||
requestBody:
|
||||
content:
|
||||
application/json:
|
||||
schema:
|
||||
$ref: '#/components/schemas/CompletionRequest'
|
||||
required: true
|
||||
responses:
|
||||
'200':
|
||||
content:
|
||||
application/json:
|
||||
schema:
|
||||
oneOf:
|
||||
- $ref: '#/components/schemas/CompletionResponse'
|
||||
- $ref: '#/components/schemas/StreamedCompletionResponse'
|
||||
description: Normal completion response. **OR** streamed completion response.
|
||||
tags:
|
||||
- Inference
|
||||
security:
|
||||
- Default: []
|
||||
servers:
|
||||
- url: http://llama.meta.com
|
||||
tags:
|
||||
- name: AgenticSystem
|
||||
- name: Inference
|
||||
- description: 'Attachments are used to refer to external resources, such as images,
|
||||
videos, audio, etc.
|
||||
|
||||
|
||||
<SchemaDefinition schemaRef="#/components/schemas/Attachment" />'
|
||||
name: Attachment
|
||||
- description: <SchemaDefinition schemaRef="#/components/schemas/ChatCompletionRequest"
|
||||
/>
|
||||
name: ChatCompletionRequest
|
||||
- description: <SchemaDefinition schemaRef="#/components/schemas/Message" />
|
||||
name: Message
|
||||
- description: <SchemaDefinition schemaRef="#/components/schemas/URL" />
|
||||
name: URL
|
||||
- description: 'Normal chat completion response.
|
||||
|
||||
|
||||
<SchemaDefinition schemaRef="#/components/schemas/ChatCompletionResponse" />'
|
||||
name: ChatCompletionResponse
|
||||
- description: 'Streamed chat completion response.
|
||||
|
||||
|
||||
<SchemaDefinition schemaRef="#/components/schemas/StreamedChatCompletionResponse"
|
||||
/>'
|
||||
name: StreamedChatCompletionResponse
|
||||
- description: <SchemaDefinition schemaRef="#/components/schemas/CompletionRequest"
|
||||
/>
|
||||
name: CompletionRequest
|
||||
- description: 'Normal completion response.
|
||||
|
||||
|
||||
<SchemaDefinition schemaRef="#/components/schemas/CompletionResponse" />'
|
||||
name: CompletionResponse
|
||||
- description: 'streamed completion response.
|
||||
|
||||
|
||||
<SchemaDefinition schemaRef="#/components/schemas/StreamedCompletionResponse"
|
||||
/>'
|
||||
name: StreamedCompletionResponse
|
||||
x-tagGroups:
|
||||
- name: Operations
|
||||
tags:
|
||||
- AgenticSystem
|
||||
- Inference
|
||||
- name: Types
|
||||
tags:
|
||||
- Attachment
|
||||
- ChatCompletionRequest
|
||||
- ChatCompletionResponse
|
||||
- CompletionRequest
|
||||
- CompletionResponse
|
||||
- Message
|
||||
- StreamedChatCompletionResponse
|
||||
- StreamedCompletionResponse
|
||||
- URL
|
||||
3
source/run.sh
Normal file
3
source/run.sh
Normal file
|
|
@ -0,0 +1,3 @@
|
|||
#!/bin/bash
|
||||
|
||||
PYTHONPATH=. python3 defn.py
|
||||
Loading…
Add table
Add a link
Reference in a new issue