phoenix-oss/llama-stack-mirror
0
0
Fork 1
mirror of https://github.com/meta-llama/llama-stack.git synced 2025-10-11 05:38:38 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions

test

Browse source 
# What does this PR do?


## Test Plan
This commit is contained in:
Eric Huang 2025-10-09 20:53:19 -07:00
parent f50ce11a3b
commit 4a3d1e33f8
31 changed files with 727 additions and 892 deletions
Download patch file Download diff file Expand all files Collapse all files

53
llama_stack/providers/remote/inference/bedrock/bedrock.py
View file

@ -6,21 +6,20 @@
import json
from collections.abc import AsyncIterator
from typing import Any
from botocore.client import BaseClient
from llama_stack.apis.inference import (
ChatCompletionRequest,
Inference,
OpenaiChatCompletionRequest,
OpenAICompletionRequest,
OpenAIEmbeddingsResponse,
)
from llama_stack.apis.inference.inference import (
OpenAIChatCompletion,
OpenAIChatCompletionChunk,
OpenAICompletion,
OpenAIMessageParam,
OpenAIResponseFormatParam,
)
from llama_stack.providers.remote.inference.bedrock.config import BedrockConfig
from llama_stack.providers.utils.bedrock.client import create_bedrock_client
@ -135,56 +134,12 @@ class BedrockInferenceAdapter(
async def openai_completion(
self,
# Standard OpenAI completion parameters
model: str,
prompt: str | list[str] | list[int] | list[list[int]],
best_of: int | None = None,
echo: bool | None = None,
frequency_penalty: float | None = None,
logit_bias: dict[str, float] | None = None,
logprobs: bool | None = None,
max_tokens: int | None = None,
n: int | None = None,
presence_penalty: float | None = None,
seed: int | None = None,
stop: str | list[str] | None = None,
stream: bool | None = None,
stream_options: dict[str, Any] | None = None,
temperature: float | None = None,
top_p: float | None = None,
user: str | None = None,
# vLLM-specific parameters
guided_choice: list[str] | None = None,
prompt_logprobs: int | None = None,
# for fill-in-the-middle type completion
suffix: str | None = None,
params: OpenAICompletionRequest,
) -> OpenAICompletion:
raise NotImplementedError("OpenAI completion not supported by the Bedrock provider")
async def openai_chat_completion(
self,
model: str,
messages: list[OpenAIMessageParam],
frequency_penalty: float | None = None,
function_call: str | dict[str, Any] | None = None,
functions: list[dict[str, Any]] | None = None,
logit_bias: dict[str, float] | None = None,
logprobs: bool | None = None,
max_completion_tokens: int | None = None,
max_tokens: int | None = None,
n: int | None = None,
parallel_tool_calls: bool | None = None,
presence_penalty: float | None = None,
response_format: OpenAIResponseFormatParam | None = None,
seed: int | None = None,
stop: str | list[str] | None = None,
stream: bool | None = None,
stream_options: dict[str, Any] | None = None,
temperature: float | None = None,
tool_choice: str | dict[str, Any] | None = None,
tools: list[dict[str, Any]] | None = None,
top_logprobs: int | None = None,
top_p: float | None = None,
user: str | None = None,
params: OpenaiChatCompletionRequest,
) -> OpenAIChatCompletion | AsyncIterator[OpenAIChatCompletionChunk]:
raise NotImplementedError("OpenAI chat completion not supported by the Bedrock provider")

26
llama_stack/providers/remote/inference/databricks/databricks.py
View file

@ -5,11 +5,14 @@
# the root directory of this source tree.
from collections.abc import Iterable
from typing import Any
from typing import TYPE_CHECKING
from databricks.sdk import WorkspaceClient
from llama_stack.apis.inference import OpenAICompletion
if TYPE_CHECKING:
from llama_stack.apis.inference import OpenAICompletionRequest
from llama_stack.log import get_logger
from llama_stack.providers.utils.inference.openai_mixin import OpenAIMixin
@ -43,25 +46,6 @@ class DatabricksInferenceAdapter(OpenAIMixin):
async def openai_completion(
self,
model: str,
prompt: str | list[str] | list[int] | list[list[int]],
best_of: int | None = None,
echo: bool | None = None,
frequency_penalty: float | None = None,
logit_bias: dict[str, float] | None = None,
logprobs: bool | None = None,
max_tokens: int | None = None,
n: int | None = None,
presence_penalty: float | None = None,
seed: int | None = None,
stop: str | list[str] | None = None,
stream: bool | None = None,
stream_options: dict[str, Any] | None = None,
temperature: float | None = None,
top_p: float | None = None,
user: str | None = None,
guided_choice: list[str] | None = None,
prompt_logprobs: int | None = None,
suffix: str | None = None,
params: "OpenAICompletionRequest",
) -> OpenAICompletion:
raise NotImplementedError()

26
llama_stack/providers/remote/inference/llama_openai_compat/llama.py
View file

@ -3,9 +3,12 @@
#
# This source code is licensed under the terms described in the LICENSE file in
# the root directory of this source tree.
from typing import Any
from typing import TYPE_CHECKING
from llama_stack.apis.inference.inference import OpenAICompletion, OpenAIEmbeddingsResponse
if TYPE_CHECKING:
from llama_stack.apis.inference import OpenAICompletionRequest
from llama_stack.log import get_logger
from llama_stack.providers.remote.inference.llama_openai_compat.config import LlamaCompatConfig
from llama_stack.providers.utils.inference.openai_mixin import OpenAIMixin
@ -34,26 +37,7 @@ class LlamaCompatInferenceAdapter(OpenAIMixin):
async def openai_completion(
self,
model: str,
prompt: str | list[str] | list[int] | list[list[int]],
best_of: int | None = None,
echo: bool | None = None,
frequency_penalty: float | None = None,
logit_bias: dict[str, float] | None = None,
logprobs: bool | None = None,
max_tokens: int | None = None,
n: int | None = None,
presence_penalty: float | None = None,
seed: int | None = None,
stop: str | list[str] | None = None,
stream: bool | None = None,
stream_options: dict[str, Any] | None = None,
temperature: float | None = None,
top_p: float | None = None,
user: str | None = None,
guided_choice: list[str] | None = None,
prompt_logprobs: int | None = None,
suffix: str | None = None,
params: "OpenAICompletionRequest",
) -> OpenAICompletion:
raise NotImplementedError()

114
llama_stack/providers/remote/inference/passthrough/passthrough.py
View file

@ -13,15 +13,14 @@ from llama_stack.apis.inference import (
Inference,
OpenAIChatCompletion,
OpenAIChatCompletionChunk,
OpenaiChatCompletionRequest,
OpenAICompletion,
OpenAICompletionRequest,
OpenAIEmbeddingsResponse,
OpenAIMessageParam,
OpenAIResponseFormatParam,
)
from llama_stack.apis.models import Model
from llama_stack.core.library_client import convert_pydantic_to_json_value
from llama_stack.providers.utils.inference.model_registry import ModelRegistryHelper
from llama_stack.providers.utils.inference.openai_compat import prepare_openai_completion_params
from .config import PassthroughImplConfig
@ -80,110 +79,33 @@ class PassthroughInferenceAdapter(Inference):
async def openai_completion(
self,
model: str,
prompt: str | list[str] | list[int] | list[list[int]],
best_of: int | None = None,
echo: bool | None = None,
frequency_penalty: float | None = None,
logit_bias: dict[str, float] | None = None,
logprobs: bool | None = None,
max_tokens: int | None = None,
n: int | None = None,
presence_penalty: float | None = None,
seed: int | None = None,
stop: str | list[str] | None = None,
stream: bool | None = None,
stream_options: dict[str, Any] | None = None,
temperature: float | None = None,
top_p: float | None = None,
user: str | None = None,
guided_choice: list[str] | None = None,
prompt_logprobs: int | None = None,
suffix: str | None = None,
params: OpenAICompletionRequest,
) -> OpenAICompletion:
client = self._get_client()
model_obj = await self.model_store.get_model(model)
model_obj = await self.model_store.get_model(params.model)
params = await prepare_openai_completion_params(
model=model_obj.provider_resource_id,
prompt=prompt,
best_of=best_of,
echo=echo,
frequency_penalty=frequency_penalty,
logit_bias=logit_bias,
logprobs=logprobs,
max_tokens=max_tokens,
n=n,
presence_penalty=presence_penalty,
seed=seed,
stop=stop,
stream=stream,
stream_options=stream_options,
temperature=temperature,
top_p=top_p,
user=user,
guided_choice=guided_choice,
prompt_logprobs=prompt_logprobs,
)
# Copy params to avoid mutating the original
params = params.model_copy()
params.model = model_obj.provider_resource_id
return await client.inference.openai_completion(**params)
request_params = params.model_dump(exclude_none=True)
return await client.inference.openai_completion(**request_params)
async def openai_chat_completion(
self,
model: str,
messages: list[OpenAIMessageParam],
frequency_penalty: float | None = None,
function_call: str | dict[str, Any] | None = None,
functions: list[dict[str, Any]] | None = None,
logit_bias: dict[str, float] | None = None,
logprobs: bool | None = None,
max_completion_tokens: int | None = None,
max_tokens: int | None = None,
n: int | None = None,
parallel_tool_calls: bool | None = None,
presence_penalty: float | None = None,
response_format: OpenAIResponseFormatParam | None = None,
seed: int | None = None,
stop: str | list[str] | None = None,
stream: bool | None = None,
stream_options: dict[str, Any] | None = None,
temperature: float | None = None,
tool_choice: str | dict[str, Any] | None = None,
tools: list[dict[str, Any]] | None = None,
top_logprobs: int | None = None,
top_p: float | None = None,
user: str | None = None,
params: OpenaiChatCompletionRequest,
) -> OpenAIChatCompletion | AsyncIterator[OpenAIChatCompletionChunk]:
client = self._get_client()
model_obj = await self.model_store.get_model(model)
model_obj = await self.model_store.get_model(params.model)
params = await prepare_openai_completion_params(
model=model_obj.provider_resource_id,
messages=messages,
frequency_penalty=frequency_penalty,
function_call=function_call,
functions=functions,
logit_bias=logit_bias,
logprobs=logprobs,
max_completion_tokens=max_completion_tokens,
max_tokens=max_tokens,
n=n,
parallel_tool_calls=parallel_tool_calls,
presence_penalty=presence_penalty,
response_format=response_format,
seed=seed,
stop=stop,
stream=stream,
stream_options=stream_options,
temperature=temperature,
tool_choice=tool_choice,
tools=tools,
top_logprobs=top_logprobs,
top_p=top_p,
user=user,
)
# Copy params to avoid mutating the original
params = params.model_copy()
params.model = model_obj.provider_resource_id
return await client.inference.openai_chat_completion(**params)
request_params = params.model_dump(exclude_none=True)
return await client.inference.openai_chat_completion(**request_params)
def cast_value_to_json_dict(self, request_params: dict[str, Any]) -> dict[str, Any]:
json_params = {}

66
llama_stack/providers/remote/inference/runpod/runpod.py
View file

@ -4,11 +4,12 @@
# This source code is licensed under the terms described in the LICENSE file in
# the root directory of this source tree.
from typing import Any
from collections.abc import AsyncIterator
from llama_stack.apis.inference import (
OpenAIMessageParam,
OpenAIResponseFormatParam,
OpenAIChatCompletion,
OpenAIChatCompletionChunk,
OpenaiChatCompletionRequest,
)
from llama_stack.providers.utils.inference.openai_mixin import OpenAIMixin
@ -34,56 +35,13 @@ class RunpodInferenceAdapter(OpenAIMixin):
async def openai_chat_completion(
self,
model: str,
messages: list[OpenAIMessageParam],
frequency_penalty: float | None = None,
function_call: str | dict[str, Any] | None = None,
functions: list[dict[str, Any]] | None = None,
logit_bias: dict[str, float] | None = None,
logprobs: bool | None = None,
max_completion_tokens: int | None = None,
max_tokens: int | None = None,
n: int | None = None,
parallel_tool_calls: bool | None = None,
presence_penalty: float | None = None,
response_format: OpenAIResponseFormatParam | None = None,
seed: int | None = None,
stop: str | list[str] | None = None,
stream: bool | None = None,
stream_options: dict[str, Any] | None = None,
temperature: float | None = None,
tool_choice: str | dict[str, Any] | None = None,
tools: list[dict[str, Any]] | None = None,
top_logprobs: int | None = None,
top_p: float | None = None,
user: str | None = None,
):
params: OpenaiChatCompletionRequest,
) -> OpenAIChatCompletion | AsyncIterator[OpenAIChatCompletionChunk]:
"""Override to add RunPod-specific stream_options requirement."""
if stream and not stream_options:
stream_options = {"include_usage": True}
# Copy params to avoid mutating the original
params = params.model_copy()
return await super().openai_chat_completion(
model=model,
messages=messages,
frequency_penalty=frequency_penalty,
function_call=function_call,
functions=functions,
logit_bias=logit_bias,
logprobs=logprobs,
max_completion_tokens=max_completion_tokens,
max_tokens=max_tokens,
n=n,
parallel_tool_calls=parallel_tool_calls,
presence_penalty=presence_penalty,
response_format=response_format,
seed=seed,
stop=stop,
stream=stream,
stream_options=stream_options,
temperature=temperature,
tool_choice=tool_choice,
tools=tools,
top_logprobs=top_logprobs,
top_p=top_p,
user=user,
)
if params.stream and not params.stream_options:
params.stream_options = {"include_usage": True}
return await super().openai_chat_completion(params)

65
llama_stack/providers/remote/inference/vllm/vllm.py
View file

@ -4,7 +4,6 @@
# This source code is licensed under the terms described in the LICENSE file in
# the root directory of this source tree.
from collections.abc import AsyncIterator
from typing import Any
from urllib.parse import urljoin
import httpx
@ -15,8 +14,7 @@ from pydantic import ConfigDict
from llama_stack.apis.inference import (
OpenAIChatCompletion,
OpenAIMessageParam,
OpenAIResponseFormatParam,
OpenaiChatCompletionRequest,
ToolChoice,
)
from llama_stack.log import get_logger
@ -79,61 +77,20 @@ class VLLMInferenceAdapter(OpenAIMixin):
async def openai_chat_completion(
self,
model: str,
messages: list[OpenAIMessageParam],
frequency_penalty: float | None = None,
function_call: str | dict[str, Any] | None = None,
functions: list[dict[str, Any]] | None = None,
logit_bias: dict[str, float] | None = None,
logprobs: bool | None = None,
max_completion_tokens: int | None = None,
max_tokens: int | None = None,
n: int | None = None,
parallel_tool_calls: bool | None = None,
presence_penalty: float | None = None,
response_format: OpenAIResponseFormatParam | None = None,
seed: int | None = None,
stop: str | list[str] | None = None,
stream: bool | None = None,
stream_options: dict[str, Any] | None = None,
temperature: float | None = None,
tool_choice: str | dict[str, Any] | None = None,
tools: list[dict[str, Any]] | None = None,
top_logprobs: int | None = None,
top_p: float | None = None,
user: str | None = None,
params: "OpenaiChatCompletionRequest",
) -> OpenAIChatCompletion | AsyncIterator[OpenAIChatCompletionChunk]:
max_tokens = max_tokens or self.config.max_tokens
# Copy params to avoid mutating the original
params = params.model_copy()
# Apply vLLM-specific defaults
if params.max_tokens is None and self.config.max_tokens:
params.max_tokens = self.config.max_tokens
# This is to be consistent with OpenAI API and support vLLM <= v0.6.3
# References:
# * https://platform.openai.com/docs/api-reference/chat/create#chat-create-tool_choice
# * https://github.com/vllm-project/vllm/pull/10000
if not tools and tool_choice is not None:
tool_choice = ToolChoice.none.value
if not params.tools and params.tool_choice is not None:
params.tool_choice = ToolChoice.none.value
return await super().openai_chat_completion(
model=model,
messages=messages,
frequency_penalty=frequency_penalty,
function_call=function_call,
functions=functions,
logit_bias=logit_bias,
logprobs=logprobs,
max_completion_tokens=max_completion_tokens,
max_tokens=max_tokens,
n=n,
parallel_tool_calls=parallel_tool_calls,
presence_penalty=presence_penalty,
response_format=response_format,
seed=seed,
stop=stop,
stream=stream,
stream_options=stream_options,
temperature=temperature,
tool_choice=tool_choice,
tools=tools,
top_logprobs=top_logprobs,
top_p=top_p,
user=user,
)
return await super().openai_chat_completion(params)