python批量转换word文档为pdf文件的示例代码

pip install python-docx
pip install comtypes

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 comtypes
collecting comtypes
  downloading comtypes-1.4.12-py3-none-any.whl.metadata (7.3 kb)
downloading comtypes-1.4.12-py3-none-any.whl (253 kb)
installing collected packages: comtypes
successfully installed comtypes-1.4.12

import os
import comtypes.client
from pathlib import path
import tkinter as tk
from tkinter import filedialog, messagebox
import threading


class wordtopdfconverter:
    def __init__(self):
        self.root = tk.tk()
        self.root.title("word转pdf批量转换器")
        self.root.geometry("500x400")

        self.setup_ui()

    def setup_ui(self):
        """设置用户界面"""
        # 标题
        title_label = tk.label(self.root, text="word转pdf批量转换器",
                               font=("arial", 16, "bold"))
        title_label.pack(pady=20)

        # 选择文件夹按钮
        self.select_btn = tk.button(self.root, text="选择word文档文件夹",
                                    command=self.select_folder,
                                    font=("arial", 12), bg="#4caf50", fg="white")
        self.select_btn.pack(pady=10)

        # 显示选择的文件夹
        self.folder_label = tk.label(self.root, text="未选择文件夹",
                                     font=("arial", 10), wraplength=400)
        self.folder_label.pack(pady=5)

        # 输出文件夹选择
        self.output_btn = tk.button(self.root, text="选择pdf输出文件夹",
                                    command=self.select_output_folder,
                                    font=("arial", 12), bg="#2196f3", fg="white")
        self.output_btn.pack(pady=10)

        self.output_label = tk.label(self.root, text="未选择输出文件夹",
                                     font=("arial", 10), wraplength=400)
        self.output_label.pack(pady=5)

        # 转换按钮
        self.convert_btn = tk.button(self.root, text="开始转换",
                                     command=self.start_conversion,
                                     font=("arial", 14), bg="#ff9800", fg="white",
                                     state="disabled")
        self.convert_btn.pack(pady=20)

        # 进度显示
        self.progress_label = tk.label(self.root, text="", font=("arial", 10))
        self.progress_label.pack(pady=5)

        # 日志文本框
        self.log_text = tk.text(self.root, height=10, width=60)
        self.log_text.pack(pady=10, padx=20, fill=tk.both, expand=true)

        # 滚动条
        scrollbar = tk.scrollbar(self.log_text)
        scrollbar.pack(side=tk.right, fill=tk.y)
        self.log_text.config(yscrollcommand=scrollbar.set)
        scrollbar.config(command=self.log_text.yview)

        self.input_folder = ""
        self.output_folder = ""

    def log_message(self, message):
        """添加日志消息"""
        self.log_text.insert(tk.end, f"{message}\n")
        self.log_text.see(tk.end)
        self.root.update()

    def select_folder(self):
        """选择包含word文档的文件夹"""
        folder = filedialog.askdirectory(title="选择包含word文档的文件夹")
        if folder:
            self.input_folder = folder
            self.folder_label.config(text=f"已选择: {folder}")
            self.check_ready()

    def select_output_folder(self):
        """选择pdf输出文件夹"""
        folder = filedialog.askdirectory(title="选择pdf输出文件夹")
        if folder:
            self.output_folder = folder
            self.output_label.config(text=f"已选择: {folder}")
            self.check_ready()

    def check_ready(self):
        """检查是否准备好转换"""
        if self.input_folder and self.output_folder:
            self.convert_btn.config(state="normal")

    def start_conversion(self):
        """开始转换过程"""
        self.convert_btn.config(state="disabled")
        thread = threading.thread(target=self.convert_word_to_pdf)
        thread.daemon = true
        thread.start()

    def convert_word_to_pdf(self):
        """执行word到pdf的转换"""
        try:
            self.log_message("开始扫描word文档...")

            # 支持的word文档扩展名
            word_extensions = ['.doc', '.docx', '.docm']
            word_files = []

            # 递归查找所有word文档
            for root, dirs, files in os.walk(self.input_folder):
                for file in files:
                    if any(file.lower().endswith(ext) for ext in word_extensions):
                        word_files.append(os.path.join(root, file))

            if not word_files:
                self.log_message("未找到任何word文档！")
                self.convert_btn.config(state="normal")
                return

            self.log_message(f"找到 {len(word_files)} 个word文档")
            self.log_message("开始转换...")

            success_count = 0
            error_count = 0

            # 初始化word应用程序
            word_app = comtypes.client.createobject('word.application')
            word_app.visible = false

            for i, word_file in enumerate(word_files, 1):
                try:
                    # 更新进度
                    self.progress_label.config(text=f"处理中: {i}/{len(word_files)}")

                    # 构建输出路径
                    relative_path = os.path.relpath(word_file, self.input_folder)
                    pdf_filename = os.path.splitext(relative_path)[0] + '.pdf'
                    pdf_path = os.path.join(self.output_folder, pdf_filename)

                    # 确保输出目录存在
                    os.makedirs(os.path.dirname(pdf_path), exist_ok=true)

                    # 转换文档
                    doc = word_app.documents.open(word_file)
                    doc.saveas(pdf_path, fileformat=17)  # 17代表pdf格式
                    doc.close()

                    self.log_message(f"✓ 成功转换: {os.path.basename(word_file)}")
                    success_count += 1

                except exception as e:
                    self.log_message(f"✗ 转换失败: {os.path.basename(word_file)} - {str(e)}")
                    error_count += 1

            # 关闭word应用程序
            word_app.quit()

            # 显示结果
            self.progress_label.config(text="")
            self.log_message(f"\n转换完成！")
            self.log_message(f"成功: {success_count} 个文件")
            self.log_message(f"失败: {error_count} 个文件")

            messagebox.showinfo("转换完成",
                                f"转换完成！\n成功: {success_count} 个文件\n失败: {error_count} 个文件")

        except exception as e:
            self.log_message(f"转换过程中发生错误: {str(e)}")
            messagebox.showerror("错误", f"转换过程中发生错误: {str(e)}")
        finally:
            self.convert_btn.config(state="normal")


def main():
    """主函数"""
    converter = wordtopdfconverter()
    converter.root.mainloop()


if __name__ == "__main__":
    main()