Selenium - 翻页（常用代码）

本文基于 macOS 10.15.7 | selenium - Version: 4.21.0 | Python 3.11.5

import time
import os 
import sys
from selenium  import webdriver
from selenium.webdriver.common.by import By

list_path = 'list.txt'

def next_page(driver):
    
    # 示例：获取页面列表连接数据
    a_nodes = driver.find_elements(by=By.XPATH, value='//div[@class="xxx"]//a')
    print(len(a_nodes))  
    for anode in a_nodes: 
        # print(anode.get_attribute('href'))
        
        with open(list_path, 'a') as fa:
            fa.write(anode.get_attribute('href').strip() + '\n')
     
    next_btn = driver.find_elements(by=By.XPATH, value='//a[@class="next"]')[-1] 
    print('-- next_btn : ', next_btn) 

    status = next_btn.get_attribute('aria-disabled') 
    print('-- next status : ', status) 
 
    if status == 'true':return 
    next_btn.click() 
    time.sleep(5) 
    next_page(driver)



def all_page():

    options = webdriver.ChromeOptions()
    # options.add_argument('headless') 
    options.add_argument('window-size=1920x1080')
    options.add_argument('start-maximized')
    options.add_argument('blink-settings=imagesEnabled=false')
    driver = webdriver.Chrome(options=options) 
    driver.set_page_load_timeout(30) 

    url = 'xxxx'
    driver.get(url)  
    time.sleep(5) 
    
    next_page(driver)

all_page()