feat: AI 서비스 MLX 듀얼 백엔드 및 모델 최적화

- MLX(맥미니 27B) 우선 → Ollama(조립컴 9B) fallback 구조
- pydantic-settings 기반 config 전환
- health check에 MLX 상태 추가
- 텍스트 모델 qwen3:8b → qwen3.5:9b-q8_0 변경

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

ai-service/routers/rag.py

@@ -1,4 +1,4 @@
-from fastapi import APIRouter
+from fastapi import APIRouter, HTTPException
 from pydantic import BaseModel
 from services.rag_service import (
     rag_suggest_solution,
@@ -30,7 +30,7 @@ async def suggest_solution(issue_id: int):
     try:
         return await rag_suggest_solution(issue_id)
     except Exception as e:
-        return {"available": False, "error": str(e)}
+        raise HTTPException(status_code=500, detail="AI 서비스 처리 중 오류가 발생했습니다")
 
 
 @router.post("/rag/ask")
@@ -38,7 +38,7 @@ async def ask_question(req: AskRequest):
     try:
         return await rag_ask(req.question, req.project_id)
     except Exception as e:
-        return {"available": False, "error": str(e)}
+        raise HTTPException(status_code=500, detail="AI 서비스 처리 중 오류가 발생했습니다")
 
 
 @router.post("/rag/pattern")
@@ -46,7 +46,7 @@ async def analyze_pattern(req: PatternRequest):
     try:
         return await rag_analyze_pattern(req.description, req.n_results)
     except Exception as e:
-        return {"available": False, "error": str(e)}
+        raise HTTPException(status_code=500, detail="AI 서비스 처리 중 오류가 발생했습니다")
 
 
 @router.post("/rag/classify")
@@ -54,4 +54,4 @@ async def classify_with_rag(req: ClassifyRequest):
     try:
         return await rag_classify_with_context(req.description, req.detail_notes)
     except Exception as e:
-        return {"available": False, "error": str(e)}
+        raise HTTPException(status_code=500, detail="AI 서비스 처리 중 오류가 발생했습니다")