Python PDF 转 Word:保留排版批量转换
import os
from pdf2docx import Converter
def pdf_to_word(folder):
for name in os.listdir(folder):
if not name.lower().endswith('.pdf'):
import os
from pdf2docx import Converter
def pdf_to_word(folder):
for name in os.listdir(folder):
if not name.lower().endswith('.pdf'):
import os
import win32com.client
def excel_to_pdf(folder):
excel = win32com.client.Dispatch('Excel.Application')
excel.Visible = False
for
from playwright.sync_api import sync_playwright
def html_to_pdf(url, output='page.pdf'):
with sync_playwright() as p:
browser = p.chromium.
from selenium import webdriverfrom selenium.webdriver.chrome.options import Options def screenshot(url, output='screenshot.png'): options = Options()
import osfrom pyzbar.pyzbar import decodefrom PIL import Image def decode_qrcodes(folder): for name in os.listdir(folder): if not name.lower().endswit
import qrcodefrom qrcode.image.styledpil import StyledPilImagefrom qrcode.image.styles.moduledrawers import RoundedModuleDrawer def gen_qrcode(data, output='
import os
from PIL import Image, ImageDraw, ImageFont
def add_watermark(folder, text='© 我的网站'):
out_dir = os.path.join(folder, 'waterm
import os
from PIL import Image
def compress_images(folder, quality=75, max_size=(1920, 1920)):
out_dir = os.path.join(folder, 'compressed')
os
import os
def batch_rename(folder, prefix='file_', start=1):
"""批量给文件夹里的文件加前缀+序号"""
files = sorted([f for f in os.listdir(folder
import os
import shutil
folder = './下载'
rules = {
'图片': ['.jpg', '.jpeg', '.png', '.gif', '.webp
import pandas as pd df = pd.read_excel('数据.xlsx') # 去重df.drop_duplicates(subset=['手机号'], keep='first').to_excel('去重后.xlsx&
import smtplibfrom email.message import EmailMessage def send_mail(to_list, subject, body, attachment=None): msg = EmailMessage() msg['From'] = &#
import time
from plyer import notification
def pomodoro(minutes=25):
print(f"专注 {minutes} 分钟,开始!")
time.sleep(minutes * 60)
notification.notify
import os
from PIL import Image
folder = './图片'
size = (800, 800) # 目标尺寸
for f in os.listdir(folder):
if f.lower().endswith(('.jpg',
from pypdf import PdfWriter, PdfReader
# 合并
writer = PdfWriter()
for pdf in ['a.pdf', 'b.pdf', 'c.pdf']:
writer.append(pdf)
write
import os
from docx import Document
folder = './文档'
for f in os.listdir(folder):
if f.endswith('.docx'):
doc = Document(os.path.j
import pandas as pd
df = pd.read_excel('总表.xlsx')
for name, group in df.groupby('部门'):
group.to_excel(f'拆分_{name}.xlsx', inde
import pandas as pd
import glob
files = glob.glob('./表格/*.xlsx')
df_list = [pd.read_excel(f) for f in files]
merged = pd.concat(df_list, ignore_index
import os
def batch_rename(folder, prefix='file_', start=1):
files = sorted([f for f in os.listdir(folder) if os.path.isfile(os.path.join(folder, f
import os def split_file(input_file, lines_per_file=2000, output_dir='output'): """ 将 input_file 按每 lines_per_file 行分割成多个文件,保存到 outpu