35 lines
1.2 KiB
Python
35 lines
1.2 KiB
Python
from __future__ import annotations
|
|
|
|
import json
|
|
|
|
import httpx
|
|
|
|
from app.core.config import Settings
|
|
from app.domain.schemas import PromptOutput
|
|
|
|
|
|
class LLMClient:
|
|
def __init__(self, settings: Settings):
|
|
self.settings = settings
|
|
|
|
def extract_json(self, system_prompt: str, user_text: str) -> PromptOutput | None:
|
|
if not self.settings.llm_enabled or not self.settings.llm_base_url or not self.settings.llm_api_key:
|
|
return None
|
|
|
|
payload = {
|
|
"model": self.settings.llm_model,
|
|
"messages": [
|
|
{"role": "system", "content": system_prompt},
|
|
{"role": "user", "content": user_text},
|
|
],
|
|
"temperature": 0.1,
|
|
"response_format": {"type": "json_object"},
|
|
}
|
|
headers = {"Authorization": f"Bearer {self.settings.llm_api_key}"}
|
|
with httpx.Client(timeout=30.0) as client:
|
|
response = client.post(f"{self.settings.llm_base_url.rstrip('/')}/chat/completions", json=payload, headers=headers)
|
|
response.raise_for_status()
|
|
data = response.json()
|
|
raw_text = data["choices"][0]["message"]["content"]
|
|
return PromptOutput(content=json.loads(raw_text), raw_text=raw_text)
|