使用python-docx和win32com库实现批量转换doc文件为docx文件

pip install python-docx
pip install pywin32 

C:\Users\wgx58>pip install python-docx
Collecting python-docx
  Using cached python_docx-1.2.0-py3-none-any.whl.metadata (2.0 kB)
Requirement already satisfied: lxml>=3.1.0 in c:\python\lib\site-packages (from python-docx) (6.0.2)
Requirement already satisfied: typing_extensions>=4.9.0 in c:\python\lib\site-packages (from python-docx) (4.15.0)
Using cached python_docx-1.2.0-py3-none-any.whl (252 kB)
Installing collected packages: python-docx
Successfully installed python-docx-1.2.0

C:\Users\wgx58>pip install pywin32
Collecting pywin32
  Downloading pywin32-311-cp313-cp313-win_amd64.whl.metadata (10 kB)
Downloading pywin32-311-cp313-cp313-win_amd64.whl (9.5 MB)
   ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 9.5/9.5 MB 12.7 kB/s  0:09:26
Installing collected packages: pywin32
Successfully installed pywin32-311

import os
import glob
from win32com import client as wc
from docx import Document
import pythoncom

def doc_to_docx(input_path, output_path):
    """
    将单个doc文件转换为docx文件
    """
    try:
        # 初始化COM组件（单线程）
        pythoncom.CoInitialize()
        
        # 创建Word应用程序对象
        word = wc.Dispatch("Word.Application")
        word.Visible = False  # 不显示Word界面
        
        # 打开doc文件
        doc = word.Documents.Open(input_path)
        
        # 保存为docx格式
        doc.SaveAs(output_path, 16)  # 16表示docx格式
        
        # 关闭文档和Word应用程序
        doc.Close()
        word.Quit()
        
        print(f"成功转换: {input_path} -> {output_path}")
        return True
        
    except Exception as e:
        print(f"转换失败 {input_path}: {str(e)}")
        return False
    finally:
        # 释放COM组件
        pythoncom.CoUninitialize()

def batch_doc_to_docx(input_folder, output_folder=None):
    """
    批量转换文件夹中的所有doc文件为docx格式
    
    Args:
        input_folder: 输入文件夹路径
        output_folder: 输出文件夹路径，如果为None则使用输入文件夹
    """
    if output_folder is None:
        output_folder = input_folder
    
    # 确保输出文件夹存在
    if not os.path.exists(output_folder):
        os.makedirs(output_folder)
    
    # 查找所有doc文件
    doc_files = glob.glob(os.path.join(input_folder, "*.doc"))
    
    if not doc_files:
        print(f"在文件夹 {input_folder} 中未找到任何doc文件")
        return
    
    print(f"找到 {len(doc_files)} 个doc文件，开始转换...")
    
    success_count = 0
    failed_files = []
    
    for doc_file in doc_files:
        # 生成输出文件路径
        filename = os.path.basename(doc_file)
        docx_filename = os.path.splitext(filename)[0] + ".docx"
        output_path = os.path.join(output_folder, docx_filename)
        
        # 转换文件
        if doc_to_docx(doc_file, output_path):
            success_count += 1
        else:
            failed_files.append(doc_file)
    
    # 输出转换结果
    print(f"\n转换完成！")
    print(f"成功: {success_count}/{len(doc_files)}")
    
    if failed_files:
        print(f"失败的文件:")
        for file in failed_files:
            print(f"  - {file}")

def main():
    """
    主函数，提供用户交互界面
    """
    print("=== DOC 转 DOCX 批量转换工具 ===")
    
    # 获取输入文件夹
    input_folder = input("请输入包含doc文件的文件夹路径: ").strip()
    
    if not os.path.exists(input_folder):
        print("错误：指定的文件夹不存在！")
        return
    
    # 询问输出文件夹
    use_same_folder = input("是否将docx文件保存在同一文件夹？(y/n): ").strip().lower()
    
    if use_same_folder == 'y':
        output_folder = input_folder
    else:
        output_folder = input("请输入输出文件夹路径: ").strip()
    
    # 开始转换
    batch_doc_to_docx(input_folder, output_folder)

if __name__ == "__main__":
    main()