zbparse/flask_app/货物标/提取采购需求main.py

import concurrent.futures
import json
import time

from flask_app.general.doubao import pdf2txt
from flask_app.货物标.技术参数要求提取 import get_technical_requirements
from flask_app.general.通义千问long import upload_file
from flask_app.货物标.商务服务其他要求提取 import get_business_requirements


#获取采购清单
def fetch_procurement_reqs(procurement_path, invalid_path):
    # procurement_docpath = pdf2docx(procurement_path)  # 采购需求docx
    # 定义默认的 procurement_reqs 字典
    DEFAULT_PROCUREMENT_REQS = {
        "采购需求": "",
        "技术要求": "",
        "商务要求": "",
        "服务要求": "",
        "其他要求": ""
    }

    # 如果 procurement_docpath 是空字符串，直接返回包含空字符串的字典
    if not procurement_path:
        return DEFAULT_PROCUREMENT_REQS.copy()

    try:
        processed_filepath = pdf2txt(procurement_path)  # 纯文本提取
        # 使用 ThreadPoolExecutor 并行处理 get_technical_requirements 和 get_business_requirements
        with concurrent.futures.ThreadPoolExecutor() as executor:
            # 提交任务给线程池
            future_technical = executor.submit(get_technical_requirements, procurement_path, invalid_path,processed_filepath)
            time.sleep(0.5)  # 保持原有的延时
            future_business = executor.submit(get_business_requirements, procurement_path,processed_filepath)

            # 获取并行任务的结果
            technical_requirements = future_technical.result()
            business_requirements = future_business.result()

        # 构建最终的采购需求字典
        procurement_reqs = {
            "采购需求": technical_requirements.get("采购需求", {})
        }

        # 合并 business_requirements 到 procurement_reqs 中
        # 这样无论 business_requirements 包含什么键（如 "技术要求"、"服务要求" 或 "技术、服务要求"），都将被保留
        procurement_reqs.update(business_requirements)

        # 如果需要确保所有默认键存在，可以取消下面的注释
        # for key, default_value in DEFAULT_PROCUREMENT_REQS.items():
        #     procurement_reqs.setdefault(key, default_value)

        return procurement_reqs

    except Exception as e:
        print(f"Error in fetch_procurement_reqs: {e}")
        # 在出错时返回默认的包含空字符串的字典
        return DEFAULT_PROCUREMENT_REQS.copy()


#TODO:技术要求可以在技术参数之后执行，把完整的技术参数输入，问大模型，除了上述内容还有哪些，这样的话把技术标和其他的区分开。
if __name__ == "__main__":
    start_time=time.time()
    output_folder = "C:\\Users\\Administrator\\Desktop\\货物标\\货物标output"
    # file_path="C:\\Users\\Administrator\\Desktop\\货物标\\output1\\2-招标文件（2020年广水市中小学教师办公电脑系统及多媒体“班班通”设备采购安装项目）_procurement.pdf"
    procurement_path = r"C:\Users\Administrator\Desktop\fsdownload\5901b181-b55f-4107-9f30-c85d607b1fa0\ztbfile_procurement.pdf"
    procurement_docpath=r"C:\Users\Administrator\Desktop\fsdownload\5901b181-b55f-4107-9f30-c85d607b1fa0"
    invalid_path="C:\\Users\\Administrator\\Desktop\\fsdownload\\db79e9e0-830e-442c-8cb6-1d036215f8ff\\ztbfile.pdf"
    res=fetch_procurement_reqs(procurement_path,invalid_path)
    print(json.dumps(res, ensure_ascii=False, indent=4))
    end_time=time.time()
    print("耗时:"+str(end_time-start_time))