36 lines
994 B
Python
36 lines
994 B
Python
|
"""Read-only smoke check for the configured LLM chat and embedding endpoints."""
|
||
|
|
from __future__ import annotations
|
||
|
|
|
||
|
|
import asyncio
|
||
|
|
import sys
|
||
|
|
from pathlib import Path
|
||
|
|
|
||
|
|
sys.path.insert(0, str(Path(__file__).resolve().parents[1]))
|
||
|
|
|
||
|
|
from tool.llm import llm
|
||
|
|
|
||
|
|
|
||
|
|
async def check() -> None:
|
||
|
|
try:
|
||
|
|
answer = await llm.chat(
|
||
|
|
[
|
||
|
|
{"role": "system", "content": "你是一个合规的基金投顾助手。"},
|
||
|
|
{"role": "user", "content": "只回复：连接正常"},
|
||
|
|
],
|
||
|
|
temperature=0,
|
||
|
|
max_tokens=16,
|
||
|
|
)
|
||
|
|
print("chat: ok", answer[:32])
|
||
|
|
except Exception as exc:
|
||
|
|
print(f"chat: {type(exc).__name__}: {exc}")
|
||
|
|
|
||
|
|
try:
|
||
|
|
vectors = await llm.embed(["投顾记忆联调"])
|
||
|
|
print("embedding: ok", len(vectors), len(vectors[0]) if vectors else 0)
|
||
|
|
except Exception as exc:
|
||
|
|
print(f"embedding: {type(exc).__name__}: {exc}")
|
||
|
|
|
||
|
|
|
||
|
|
if __name__ == "__main__":
|
||
|
|
asyncio.run(check())
|