phoenix-oss/llama-stack-mirror
0
0
Fork 1
mirror of https://github.com/meta-llama/llama-stack.git synced 2025-12-12 12:06:04 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions

address comments

Browse source
This commit is contained in:
Swapna Lekkala 2025-10-14 13:30:42 -07:00
parent 74be622f57
commit 9a4d3d7576
1 changed files with 5 additions and 13 deletions
Download patch file Download diff file Expand all files Collapse all files

18
llama_stack/providers/inline/agents/meta_reference/responses/streaming.py
View file

@ -134,16 +134,6 @@ class StreamingResponseOrchestrator:
# Track if we've sent a refusal response
self.violation_detected = False
async def _apply_guardrails(self, text: str, context: str = "content") -> str | None:
"""Apply guardrails to text content. Returns violation message if blocked."""
if not self.guardrail_ids or not text:
return None
violation_message = await run_multiple_guardrails(self.safety_api, text, self.guardrail_ids)
if violation_message:
logger.info(f"{context.capitalize()} guardrail violation: {violation_message}")
return violation_message
async def _create_refusal_response(self, violation_message: str) -> OpenAIResponseObjectStream:
"""Create a refusal response to replace streaming content."""
refusal_content = OpenAIResponseContentPartRefusal(refusal=violation_message)
@ -205,8 +195,9 @@ class StreamingResponseOrchestrator:
# Input safety validation - check messages before processing
if self.guardrail_ids:
combined_text = interleaved_content_as_str([msg.content for msg in self.ctx.messages])
input_violation_message = await self._apply_guardrails(combined_text, "input")
input_violation_message = await run_multiple_guardrails(self.safety_api, combined_text, self.guardrail_ids)
if input_violation_message:
logger.info(f"Input guardrail violation: {input_violation_message}")
# Return refusal response immediately
yield await self._create_refusal_response(input_violation_message)
return
@ -715,10 +706,11 @@ class StreamingResponseOrchestrator:
) + tool_call.function.arguments
# Output Safety Validation for a chunk
if chat_response_content:
if self.guardrail_ids:
accumulated_text = "".join(chat_response_content)
violation_message = await self._apply_guardrails(accumulated_text, "output")
violation_message = await run_multiple_guardrails(self.safety_api, accumulated_text, self.guardrail_ids)
if violation_message:
logger.info(f"Output guardrail violation: {violation_message}")
yield await self._create_refusal_response(violation_message)
self.violation_detected = True
return