56 lines
1.6 KiB
Python
56 lines
1.6 KiB
Python
|
import subprocess
|
||
|
|
import os
|
||
|
|
import asyncio
|
||
|
|
from configs.basic_config import *
|
||
|
|
|
||
|
|
|
||
|
|
async def convert_doc_to_docx(file_path: str) -> bool:
|
||
|
|
"""使用 libreoffice 将 doc 文件转换为 docx 文件, 替换原来的文件"""
|
||
|
|
|
||
|
|
try:
|
||
|
|
# 检查文件是否存在
|
||
|
|
if not os.path.exists(file_path):
|
||
|
|
raise FileNotFoundError(f"文件不存在: {file_path}")
|
||
|
|
|
||
|
|
# 获取文件所在目录和文件名
|
||
|
|
file_dir = os.path.dirname(file_path)
|
||
|
|
file_name = os.path.basename(file_path)
|
||
|
|
|
||
|
|
# 构建 libreoffice 命令
|
||
|
|
cmd = [
|
||
|
|
"soffice",
|
||
|
|
"--headless",
|
||
|
|
"--convert-to",
|
||
|
|
"docx",
|
||
|
|
"--outdir",
|
||
|
|
file_dir,
|
||
|
|
file_path,
|
||
|
|
]
|
||
|
|
|
||
|
|
# 执行转换
|
||
|
|
process = await asyncio.create_subprocess_exec(
|
||
|
|
*cmd, stdout=subprocess.PIPE, stderr=subprocess.PIPE
|
||
|
|
)
|
||
|
|
stdout, stderr = await process.communicate()
|
||
|
|
|
||
|
|
if process.returncode != 0:
|
||
|
|
raise Exception(f"转换失败: {stderr.decode()}")
|
||
|
|
|
||
|
|
# 获取转换后的文件路径
|
||
|
|
docx_file = os.path.join(file_dir, os.path.splitext(file_name)[0] + ".docx")
|
||
|
|
|
||
|
|
# 检查转换后的文件是否存在
|
||
|
|
if not os.path.exists(docx_file):
|
||
|
|
raise FileNotFoundError(f"转换后的文件不存在: {docx_file}")
|
||
|
|
|
||
|
|
# 删除原文件并重命名新文件
|
||
|
|
os.remove(file_path)
|
||
|
|
os.rename(docx_file, file_path)
|
||
|
|
|
||
|
|
logger.info(f"成功将 {file_path} 转换为 docx 格式")
|
||
|
|
return True
|
||
|
|
|
||
|
|
except Exception as e:
|
||
|
|
logger.error(f"转换 doc 到 docx 失败: {str(e)}")
|
||
|
|
raise
|