wehub-resource-sync
e768098d0e
tools_continuous_delivery / Private PyPI non-main branch release (push) Has been skipped
tools_continuous_delivery / Private PyPI main branch release (push) Failing after 2m42s
Publish Promptflow Doc / Build (push) Has been cancelled
Publish Promptflow Doc / Deploy (push) Has been cancelled
Flake8 Lint / flake8 (push) Has been cancelled
Spell check CI / Spell_Check (push) Has been cancelled
57 lines
1.8 KiB
Python
57 lines
1.8 KiB
Python
from promptflow.core import tool
|
|
from typing import Union
|
|
from promptflow.connections import AzureOpenAIConnection, OpenAIConnection
|
|
from openai import AzureOpenAI as AzureOpenAIClient
|
|
from openai import OpenAI as OpenAIClient
|
|
from promptflow.tools.common import parse_chat
|
|
|
|
|
|
def parse_questions(completion: str) -> list:
|
|
questions = []
|
|
|
|
for item in completion.choices:
|
|
response = getattr(item.message, "content", "")
|
|
print(response)
|
|
questions.append(response)
|
|
return questions
|
|
|
|
|
|
@tool
|
|
def call_llm_chat(
|
|
connection: Union[AzureOpenAIConnection, OpenAIConnection],
|
|
prompt: str,
|
|
question_count: int,
|
|
deployment_name_or_model: str,
|
|
stop: list = [],
|
|
) -> str:
|
|
|
|
messages = parse_chat(prompt)
|
|
params = {
|
|
"model": deployment_name_or_model,
|
|
"messages": messages,
|
|
"temperature": 1.0,
|
|
"top_p": 1.0,
|
|
"stream": False,
|
|
"stop": stop if stop else None,
|
|
"presence_penalty": 0.8,
|
|
"frequency_penalty": 0.8,
|
|
"max_tokens": None,
|
|
"n": question_count
|
|
}
|
|
if isinstance(connection, AzureOpenAIConnection):
|
|
client = AzureOpenAIClient(api_key=connection.api_key,
|
|
api_version=connection.api_version,
|
|
azure_endpoint=connection.api_base)
|
|
elif isinstance(connection, OpenAIConnection):
|
|
client = OpenAIClient(api_key=connection.api_key,
|
|
organization=connection.organization,
|
|
base_url=connection.base_url)
|
|
else:
|
|
raise ValueError("Unsupported connection type")
|
|
|
|
completion = client.chat.completions.create(**params)
|
|
print(completion)
|
|
questions = parse_questions(completion)
|
|
|
|
return "\n".join(questions)
|