dsProject/dsRag/Start.py

import urllib.parse

from Util.SearchUtil import *
# 初始化日志
logger = logging.getLogger(__name__)
logger.setLevel(logging.INFO)

# 配置日志处理器
log_file = os.path.join(os.path.dirname(__file__), 'Logs', 'app.log')
os.makedirs(os.path.dirname(log_file), exist_ok=True)

# 文件处理器
file_handler = RotatingFileHandler(
    log_file, maxBytes=1024 * 1024, backupCount=5, encoding='utf-8')
file_handler.setFormatter(logging.Formatter(
    '%(asctime)s - %(name)s - %(levelname)s - %(message)s'))

# 控制台处理器
console_handler = logging.StreamHandler()
console_handler.setFormatter(logging.Formatter(
    '%(asctime)s - %(name)s - %(levelname)s - %(message)s'))

logger.addHandler(file_handler)
logger.addHandler(console_handler)



async def lifespan(app: FastAPI):
    # 初始化阿里云大模型工具
    app.state.aliyun_util = ALiYunUtil()

    # 抑制HTTPS相关警告
    warnings.filterwarnings('ignore', message='Connecting to .* using TLS with verify_certs=False is insecure')
    warnings.filterwarnings('ignore', message='Unverified HTTPS request is being made to host')
    yield


app = FastAPI(lifespan=lifespan)

# 挂载静态文件目录
app.mount("/static", StaticFiles(directory="Static"), name="static")


@app.post("/api/save-word")
async def save_to_word(request: Request):
    output_file = None
    try:
        # Parse request data
        try:
            data = await request.json()
            markdown_content = data.get('markdown_content', '')
            if not markdown_content:
                raise ValueError("Empty MarkDown content")
        except Exception as e:
            logger.error(f"Request parsing failed: {str(e)}")
            raise HTTPException(status_code=400, detail=f"Invalid request: {str(e)}")

        # 创建临时Markdown文件
        temp_md = os.path.join(tempfile.gettempdir(), uuid.uuid4().hex + ".md")
        with open(temp_md, "w", encoding="utf-8") as f:
            f.write(markdown_content)

        # 使用pandoc转换
        output_file = os.path.join(tempfile.gettempdir(), "【理想大模型】问答.docx")
        subprocess.run(['pandoc', temp_md, '-o', output_file, '--resource-path=static'], check=True)

        # 读取生成的Word文件
        with open(output_file, "rb") as f:
            stream = BytesIO(f.read())

        # 返回响应
        encoded_filename = urllib.parse.quote("【理想大模型】问答.docx")
        return StreamingResponse(
            stream,
            media_type="application/vnd.openxmlformats-officedocument.wordprocessingml.document",
            headers={"Content-Disposition": f"attachment; filename*=UTF-8''{encoded_filename}"})

    except HTTPException:
        raise
    except Exception as e:
        logger.error(f"Unexpected error: {str(e)}")
        raise HTTPException(status_code=500, detail="Internal server error")
    finally:
        # 清理临时文件
        try:
            if temp_md and os.path.exists(temp_md):
                os.remove(temp_md)
            if output_file and os.path.exists(output_file):
                os.remove(output_file)
        except Exception as e:
            logger.warning(f"Failed to clean up temp files: {str(e)}")



@app.post("/api/rag")
async def rag(request: Request):
    data = await request.json()
    query = data.get('query', '')
    query_tags = data.get('tags', [])

    # 调用es进行混合搜索
    search_results = queryByEs(query, query_tags)

    # 调用大模型
    markdown_content = callLLM(request, query, search_results)

    # 如果有正确的结果
    if markdown_content:
        return {"data": markdown_content, "format": "markdown"}

    return {"data": "没有在知识库中找到相关的信息，无法回答此问题。"}


if __name__ == "__main__":
    uvicorn.run(app, host="0.0.0.0", port=8000)