thomasm291678
feat: 纳入 DAPT AI 引擎 vetnetv2.0 + 测试报告修复(语言约束/must_not_miss兜底/降温度/sdpa) + KB 走 git-lfs
9a71f08 | #!/usr/bin/env python3 | |
| import os, sys, time, json, logging | |
| from contextlib import asynccontextmanager | |
| from fastapi import FastAPI, HTTPException | |
| from fastapi.middleware.cors import CORSMiddleware | |
| from pydantic import BaseModel | |
| sys.path.insert(0, os.path.dirname(os.path.abspath(__file__))) | |
| from model import VetModel | |
| logging.basicConfig(level=logging.INFO, format='%(asctime)s %(levelname)s %(message)s') | |
| log = logging.getLogger('vetnet') | |
| MODEL = None | |
| class RunRequest(BaseModel): | |
| text: str | |
| species: str = '狗' | |
| class ChatRequest(BaseModel): | |
| question: str | |
| species: str = '狗' | |
| async def lifespan(app: FastAPI): | |
| global MODEL | |
| log.info('loading model...') | |
| t0 = time.time() | |
| api_key = os.getenv('DEEPSEEK_API_KEY', '') | |
| local_path = os.getenv('VET_LOCAL_MODEL', './vet-qwen3b-merged') | |
| if api_key and os.path.isdir(local_path): | |
| MODEL = VetModel(api_key=api_key, local_model_path=local_path) | |
| log.info(f'mode=pipeline loaded in {int((time.time()-t0)*1000)}ms') | |
| elif os.path.isdir(local_path): | |
| MODEL = VetModel(local_model_path=local_path) | |
| log.info(f'mode=local loaded in {int((time.time()-t0)*1000)}ms') | |
| elif api_key: | |
| MODEL = VetModel(api_key=api_key) | |
| log.info(f'mode=api loaded in {int((time.time()-t0)*1000)}ms') | |
| else: | |
| log.error('DEEPSEEK_API_KEY or VET_LOCAL_MODEL required') | |
| raise RuntimeError('DEEPSEEK_API_KEY or VET_LOCAL_MODEL required') | |
| yield | |
| if hasattr(MODEL, 'llm') and hasattr(MODEL.llm, 'unload'): | |
| MODEL.llm.unload() | |
| app = FastAPI(title='VetNet v1.2', lifespan=lifespan) | |
| app.add_middleware(CORSMiddleware, allow_origins=['*'], allow_methods=['*'], allow_headers=['*']) | |
| def health(): | |
| return {'status': 'ok', 'mode': MODEL.mode, 'ready': MODEL.is_ready} | |
| def run(req: RunRequest): | |
| if not MODEL: raise HTTPException(503, 'model not loaded') | |
| t0 = time.time() | |
| try: | |
| r = MODEL.run(req.text, req.species) | |
| r['server_ms'] = int((time.time() - t0) * 1000) | |
| return r | |
| except Exception as e: | |
| log.exception('run failed') | |
| raise HTTPException(500, str(e)) | |
| def chat(req: ChatRequest): | |
| if not MODEL: raise HTTPException(503, 'model not loaded') | |
| try: | |
| return {'answer': MODEL.chat(req.question, req.species)} | |
| except Exception as e: | |
| log.exception('chat failed') | |
| raise HTTPException(500, str(e)) | |
| if __name__ == '__main__': | |
| import uvicorn | |
| port = int(os.getenv('PORT', '8000')) | |
| uvicorn.run(app, host='0.0.0.0', port=port) | |