apunkt/plano
1
0
Fork 0
mirror of https://github.com/katanemo/plano.git synced 2026-06-17 15:25:17 +02:00
Code Issues Projects Releases Packages Wiki Activity

clean code + remove cmts

Browse source
This commit is contained in:
co tran 2025-04-01 01:19:29 +00:00
parent 81c7bfff84
commit b7916ce192
2 changed files with 0 additions and 26 deletions
Download patch file Download diff file Expand all files Collapse all files

20
model_server/src/core/function_calling.py
View file

@ -369,15 +369,6 @@ class ArchFunctionHandler(ArchBaseHandler):
model_response += chunk.choices[0].delta.content
logger.info(f"[Agent Orchestrator]: response received: {model_response}")
else:
# *********************************************************************************************
# TODO:
# Update the following logic for hallucination check
# 1. If the model response starts wtth `tool_calls`, continue halluciantion check:
# - If hallucination detected, start prompt prefilling
# - Otherwise, continue until the end
# 2. Otherwise, stop it
# *********************************************************************************************
# initialize the hallucination handler, which is an iterator
self.hallucination_state = HallucinationState(
response_iterator=response, function=req.tools
@ -416,17 +407,6 @@ class ArchFunctionHandler(ArchBaseHandler):
model_response = self.default_prefix + "".join(
self.hallucination_state.tokens
)
# else:
# # start parameter gathering if the model is not generating tool calls
# prefill_response = self._engage_parameter_gathering(messages)
# model_response = prefill_response.choices[0].message.content
# # *********************************************************************************************\
# # TODO: Remove the following for loop after updating hallucination check
# # *********************************************************************************************
# for chunk in response:
# if len(chunk.choices) > 0 and chunk.choices[0].delta.content:
# model_response += chunk.choices[0].delta.content
logger.info(f"[arch-fc]: raw model response: {model_response}")
# Extract tool calls from model response

6
model_server/src/main.py
View file

@ -97,9 +97,6 @@ async def function_calling(req: ChatMessage, res: Response):
elif final_response.choices[0].message.tool_calls:
final_response.metadata["function_latency"] = str(round(latency * 1000, 3))
# *********************************************************************************************
# TODO: Put the following code back when hallucination check is ready
# *********************************************************************************************
if not use_agent_orchestrator:
final_response.metadata["hallucination"] = str(
model_handler.hallucination_state.hallucination
@ -111,9 +108,6 @@ async def function_calling(req: ChatMessage, res: Response):
if not use_agent_orchestrator:
final_response.metadata["intent_latency"] = str(round(latency * 1000, 3))
# *********************************************************************************************
# TODO: Put the following code back when hallucination check is ready
# *********************************************************************************************
final_response.metadata["hallucination"] = str(
model_handler.hallucination_state.hallucination
)