shangfangjian
0a27b18c27
Profeto — 给 LLM 提供数据,让 LLM 预测足球比分。 核心模块: - FastAPI 后端 + PostgreSQL (SQLAlchemy async) - 多 Agent LLM 预测 (5 专家 + 终裁) - 数据采集 (bzzoiro / understat / injuries) - React 前端 (Vite + Tailwind) 包含: - 数据源抽象 (DataSource 协议 + 注册表) - Alembic 数据库迁移 - Prompt 模板 (单/多 Agent) - 核心路径单元测试
115 lines
3.4 KiB
Python
115 lines
3.4 KiB
Python
"""多提供商 LLM 抽象(OpenAI-compatible 接口)。
|
|
|
|
支持: OpenAI / Deepseek / Ollama / 任何 OpenAI-compatible 网关。
|
|
"""
|
|
from __future__ import annotations
|
|
|
|
import json
|
|
import logging
|
|
import time
|
|
from dataclasses import dataclass, field
|
|
from typing import Any
|
|
|
|
from src.core.config import settings
|
|
from src.core.http_client import get_client
|
|
|
|
logger = logging.getLogger(__name__)
|
|
|
|
|
|
@dataclass
|
|
class LLMResponse:
|
|
content: str
|
|
parsed: dict | None = None
|
|
prompt_tokens: int | None = None
|
|
completion_tokens: int | None = None
|
|
latency_ms: int | None = None
|
|
raw: dict | None = None
|
|
error: str | None = None
|
|
|
|
|
|
@dataclass
|
|
class LLMProvider:
|
|
"""OpenAI-compatible async provider。"""
|
|
|
|
api_key: str = ""
|
|
base_url: str = "https://api.openai.com/v1"
|
|
model: str = "gpt-4o"
|
|
timeout: int = 60
|
|
extra_headers: dict = field(default_factory=dict)
|
|
|
|
async def chat(
|
|
self,
|
|
system: str,
|
|
user: str,
|
|
*,
|
|
json_mode: bool = True,
|
|
temperature: float = 0.3,
|
|
max_tokens: int = 1000,
|
|
) -> LLMResponse:
|
|
"""发请求,返回结构化响应。"""
|
|
headers = {
|
|
"Authorization": f"Bearer {self.api_key}",
|
|
"Content-Type": "application/json",
|
|
**self.extra_headers,
|
|
}
|
|
payload: dict[str, Any] = {
|
|
"model": self.model,
|
|
"messages": [
|
|
{"role": "system", "content": system},
|
|
{"role": "user", "content": user},
|
|
],
|
|
"temperature": temperature,
|
|
"max_tokens": max_tokens,
|
|
}
|
|
if json_mode:
|
|
payload["response_format"] = {"type": "json_object"}
|
|
|
|
start = time.perf_counter()
|
|
try:
|
|
client = get_client()
|
|
resp = await client.post(
|
|
f"{self.base_url}/chat/completions",
|
|
headers=headers,
|
|
json=payload,
|
|
timeout=self.timeout,
|
|
)
|
|
resp.raise_for_status()
|
|
data = resp.json()
|
|
latency = int((time.perf_counter() - start) * 1000)
|
|
usage = data.get("usage", {})
|
|
content = data["choices"][0]["message"]["content"]
|
|
parsed = None
|
|
if json_mode:
|
|
try:
|
|
parsed = json.loads(content)
|
|
except json.JSONDecodeError:
|
|
# 尝试从代码块提取
|
|
import re
|
|
m = re.search(r"```(?:json)?\s*([\s\S]*?)\s*```", content)
|
|
if m:
|
|
try:
|
|
parsed = json.loads(m.group(1))
|
|
except json.JSONDecodeError:
|
|
pass
|
|
return LLMResponse(
|
|
content=content,
|
|
parsed=parsed,
|
|
prompt_tokens=usage.get("prompt_tokens"),
|
|
completion_tokens=usage.get("completion_tokens"),
|
|
latency_ms=latency,
|
|
raw=data,
|
|
)
|
|
except Exception as e:
|
|
latency = int((time.perf_counter() - start) * 1000)
|
|
logger.error("LLM request failed: %s", e)
|
|
return LLMResponse(content="", error=str(e), latency_ms=latency)
|
|
|
|
|
|
def get_default_provider() -> LLMProvider:
|
|
return LLMProvider(
|
|
api_key=settings.LLM_API_KEY,
|
|
base_url=settings.LLM_BASE_URL,
|
|
model=settings.LLM_MODEL,
|
|
timeout=settings.LLM_TIMEOUT,
|
|
)
|