85f674c9cb
- ChromaDB → Qdrant 전환 (맥미니 기존 인스턴스, tk_qc_issues 컬렉션) - Ollama 임베딩/텍스트 생성 URL 분리 (임베딩: 맥미니, 텍스트: GPU서버) - MLX fallback 제거, Ollama 단일 경로로 단순화 - ds923 docker-compose에서 ai-service 제거 - gateway/system3-web nginx: ai-service 프록시를 ai.hyungi.net 경유로 변경 - resolver + 변수 기반 proxy_pass로 런타임 DNS 해석 (컨테이너 시작 실패 방지) Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
83 lines
3.3 KiB
Python
83 lines
3.3 KiB
Python
import asyncio
|
|
import httpx
|
|
from config import settings
|
|
|
|
|
|
class OllamaClient:
|
|
def __init__(self):
|
|
self.text_url = settings.OLLAMA_BASE_URL # GPU서버 (텍스트 생성)
|
|
self.embed_url = settings.OLLAMA_EMBED_URL # 맥미니 (임베딩)
|
|
self.timeout = httpx.Timeout(float(settings.OLLAMA_TIMEOUT), connect=10.0)
|
|
self._client: httpx.AsyncClient | None = None
|
|
|
|
async def _get_client(self) -> httpx.AsyncClient:
|
|
if self._client is None or self._client.is_closed:
|
|
self._client = httpx.AsyncClient(timeout=self.timeout)
|
|
return self._client
|
|
|
|
async def close(self):
|
|
if self._client and not self._client.is_closed:
|
|
await self._client.aclose()
|
|
self._client = None
|
|
|
|
async def generate_embedding(self, text: str) -> list[float]:
|
|
client = await self._get_client()
|
|
response = await client.post(
|
|
f"{self.embed_url}/api/embeddings",
|
|
json={"model": settings.OLLAMA_EMBED_MODEL, "prompt": text},
|
|
)
|
|
response.raise_for_status()
|
|
return response.json()["embedding"]
|
|
|
|
async def batch_embeddings(self, texts: list[str], concurrency: int = 5) -> list[list[float]]:
|
|
semaphore = asyncio.Semaphore(concurrency)
|
|
|
|
async def _embed(text: str) -> list[float]:
|
|
async with semaphore:
|
|
return await self.generate_embedding(text)
|
|
|
|
return await asyncio.gather(*[_embed(t) for t in texts])
|
|
|
|
async def generate_text(self, prompt: str, system: str = None) -> str:
|
|
messages = []
|
|
if system:
|
|
messages.append({"role": "system", "content": system})
|
|
messages.append({"role": "user", "content": prompt})
|
|
client = await self._get_client()
|
|
response = await client.post(
|
|
f"{self.text_url}/api/chat",
|
|
json={
|
|
"model": settings.OLLAMA_TEXT_MODEL,
|
|
"messages": messages,
|
|
"stream": False,
|
|
"think": False,
|
|
"options": {"temperature": 0.3, "num_predict": 2048},
|
|
},
|
|
)
|
|
response.raise_for_status()
|
|
return response.json()["message"]["content"]
|
|
|
|
async def check_health(self) -> dict:
|
|
result = {}
|
|
short_timeout = httpx.Timeout(5.0, connect=3.0)
|
|
# GPU서버 Ollama (텍스트 생성)
|
|
try:
|
|
async with httpx.AsyncClient(timeout=short_timeout) as c:
|
|
response = await c.get(f"{self.text_url}/api/tags")
|
|
models = response.json().get("models", [])
|
|
result["ollama_text"] = {"status": "connected", "url": self.text_url, "models": [m["name"] for m in models]}
|
|
except Exception:
|
|
result["ollama_text"] = {"status": "disconnected", "url": self.text_url}
|
|
# 맥미니 Ollama (임베딩)
|
|
try:
|
|
async with httpx.AsyncClient(timeout=short_timeout) as c:
|
|
response = await c.get(f"{self.embed_url}/api/tags")
|
|
models = response.json().get("models", [])
|
|
result["ollama_embed"] = {"status": "connected", "url": self.embed_url, "models": [m["name"] for m in models]}
|
|
except Exception:
|
|
result["ollama_embed"] = {"status": "disconnected", "url": self.embed_url}
|
|
return result
|
|
|
|
|
|
ollama_client = OllamaClient()
|