phoenix-oss/llama-stack-mirror
0
0
Fork 1
mirror of https://github.com/meta-llama/llama-stack.git synced 2025-10-12 13:57:57 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions

featu: support passing "extra body" throught to providers

Browse source 
# What does this PR do?
Allows passing through extra_body parameters to inference providers.


closes #2720

## Test Plan
CI and added new test
This commit is contained in:
Eric Huang 2025-10-10 16:10:35 -07:00
parent 80d58ab519
commit 579aa96b2c
42 changed files with 3147 additions and 202 deletions
Download patch file Download diff file Expand all files Collapse all files

8
llama_stack/providers/inline/batches/reference/batches.py
View file

@ -22,8 +22,8 @@ from llama_stack.apis.files import Files, OpenAIFilePurpose
from llama_stack.apis.inference import (
Inference,
OpenAIAssistantMessageParam,
OpenAIChatCompletionRequest,
OpenAICompletionRequest,
OpenAIChatCompletionRequestWithExtraBody,
OpenAICompletionRequestWithExtraBody,
OpenAIDeveloperMessageParam,
OpenAIMessageParam,
OpenAISystemMessageParam,
@ -608,7 +608,7 @@ class ReferenceBatchesImpl(Batches):
# TODO(SECURITY): review body for security issues
if request.url == "/v1/chat/completions":
request.body["messages"] = [convert_to_openai_message_param(msg) for msg in request.body["messages"]]
chat_params = OpenAIChatCompletionRequest(**request.body)
chat_params = OpenAIChatCompletionRequestWithExtraBody(**request.body)
chat_response = await self.inference_api.openai_chat_completion(chat_params)
# this is for mypy, we don't allow streaming so we'll get the right type
@ -623,7 +623,7 @@ class ReferenceBatchesImpl(Batches):
},
}
elif request.url == "/v1/completions":
completion_params = OpenAICompletionRequest(**request.body)
completion_params = OpenAICompletionRequestWithExtraBody(**request.body)
completion_response = await self.inference_api.openai_completion(completion_params)
# this is for mypy, we don't allow streaming so we'll get the right type