from selenium import webdriver from selenium.webdriver.common.by import By from selenium.webdriver.common.keys import Keys from openpyxl import Workbook, load_workbook import pytesseract from PIL import Image, ImageFilter import time

优化后的代码

driver = webdriver.Chrome() driver.maximize_window()

加载Excel文件

wb = load_workbook('data.xlsx') ws = wb.active

定义常量

IMG_PATH = 'captcha.png' SLEEP_TIME = 3

获取验证码图片并识别

def get_captcha(): captcha = driver.find_element(By.XPATH, '//img[@id="captcha_img"]') captcha.screenshot(IMG_PATH) img = Image.open(IMG_PATH) text = pytesseract.image_to_string(img) return text.strip()

遍历Excel表格并填写表单

for row in ws.iter_rows(min_row=2): name = row[0].value email = row[1].value message = row[2].value

# 打开网页并填写表单
driver.get('https://www.example.com/contact')
driver.find_element(By.NAME, 'name').send_keys(name)
driver.find_element(By.NAME, 'email').send_keys(email)
driver.find_element(By.NAME, 'message').send_keys(message)

# 获取验证码并填写
captcha_text = get_captcha()
driver.find_element(By.NAME, 'captcha').send_keys(captcha_text)

# 提交表单并等待页面加载
driver.find_element(By.NAME, 'submit').click()
time.sleep(SLEEP_TIME)

关闭浏览器

driver.quit(