2026/2/1大约 4 分钟
GUI 自动化
GUI 自动化通过模拟用户操作来自动化桌面应用程序和浏览器操作。
PyAutoGUI
基础操作
import pyautogui
import time
# 获取屏幕尺寸
screen_width, screen_height = pyautogui.size()
print(f"屏幕尺寸: {screen_width} x {screen_height}")
# 获取鼠标位置
current_x, current_y = pyautogui.position()
print(f"鼠标位置: ({current_x}, {current_y})")
# 移动鼠标
pyautogui.moveTo(100, 100, duration=0.5) # 绝对位置
pyautogui.moveRel(50, 0, duration=0.3) # 相对移动
# 点击
pyautogui.click(x=100, y=100) # 单击
pyautogui.doubleClick(x=200, y=200) # 双击
pyautogui.rightClick(x=300, y=300) # 右键
# 拖动
pyautogui.dragTo(400, 400, duration=1) # 绝对拖动
pyautogui.dragRel(100, 0, duration=0.5) # 相对拖动键盘操作
import pyautogui
# 输入文本
pyautogui.write('Hello, World!', interval=0.05)
# 按键
pyautogui.press('enter')
pyautogui.press('tab')
pyautogui.press('space')
# 组合键
pyautogui.hotkey('ctrl', 'c') # 复制
pyautogui.hotkey('ctrl', 'v') # 粘贴
pyautogui.hotkey('ctrl', 'a') # 全选
pyautogui.hotkey('ctrl', 's') # 保存
# 特殊按键
pyautogui.press('f1') # 功能键
pyautogui.press('esc') # ESC
pyautogui.press('delete') # Delete屏幕操作
import pyautogui
# 截图
screenshot = pyautogui.screenshot()
screenshot.save('screenshot.png')
# 区域截图
region_screenshot = pyautogui.screenshot(region=(0, 0, 300, 400))
# 查找图像
location = pyautogui.locateOnScreen('button.png')
if location:
center = pyautogui.center(location)
pyautogui.click(center)
# 查找所有匹配
locations = pyautogui.locateAllOnScreen('icon.png')
for location in locations:
center = pyautogui.center(location)
pyautogui.click(center)安全措施
import pyautogui
# 启用故障保护
pyautogui.FAILSAFE = True
# 设置操作暂停
pyautogui.PAUSE = 0.5
# 紧急停止：移动鼠标到屏幕左上角
# 防止检测（添加随机性）
import random
def human_like_click(x, y):
# 添加随机偏移
offset_x = random.randint(-5, 5)
offset_y = random.randint(-5, 5)
# 随机移动时间
duration = random.uniform(0.2, 0.5)
pyautogui.click(x + offset_x, y + offset_y, duration=duration)Selenium
基础使用
from selenium import webdriver
from selenium.webdriver.common.by import By
from selenium.webdriver.common.keys import Keys
from selenium.webdriver.support.ui import WebDriverWait
from selenium.webdriver.support import expected_conditions as EC
# 创建浏览器
driver = webdriver.Chrome()
# 访问页面
driver.get('https://www.example.com')
# 查找元素
element = driver.find_element(By.ID, 'my-id')
elements = driver.find_elements(By.CLASS_NAME, 'my-class')
# 查找方式
driver.find_element(By.ID, 'id')
driver.find_element(By.NAME, 'name')
driver.find_element(By.XPATH, '//input[@id="id"]')
driver.find_element(By.CSS_SELECTOR, '#id')
driver.find_element(By.LINK_TEXT, '链接文本')
driver.find_element(By.PARTIAL_LINK_TEXT, '部分链接')
# 元素操作
element.click()
element.send_keys('文本内容')
element.clear()
element.submit()
# 获取元素信息
text = element.text
value = element.get_attribute('value')
is_displayed = element.is_displayed()
is_enabled = element.is_enabled()
# 关闭浏览器
driver.quit()等待策略
from selenium.webdriver.support.ui import WebDriverWait
from selenium.webdriver.support import expected_conditions as EC
from selenium.webdriver.common.by import By
# 显式等待
wait = WebDriverWait(driver, 10)
# 等待元素出现
element = wait.until(
EC.presence_of_element_located((By.ID, 'my-id'))
)
# 等待元素可见
element = wait.until(
EC.visibility_of_element_located((By.ID, 'my-id'))
)
# 等待元素可点击
element = wait.until(
EC.element_to_be_clickable((By.ID, 'my-id'))
)
# 等待文本出现
wait.until(
EC.text_to_be_present_in_element((By.ID, 'my-id'), '预期文本')
)
# 自定义等待条件
def custom_condition(driver):
element = driver.find_element(By.ID, 'status')
return element.get_attribute('data-ready') == 'true'
wait.until(custom_condition)表单处理
from selenium.webdriver.support.ui import Select
from selenium.webdriver.common.by import By
# 文本输入
text_input = driver.find_element(By.NAME, 'username')
text_input.clear()
text_input.send_keys('myusername')
# 下拉选择
select_element = driver.find_element(By.NAME, 'country')
select = Select(select_element)
# 选择方式
select.select_by_visible_text('China')
select.select_by_value('cn')
select.select_by_index(1)
# 多选
select.deselect_by_visible_text('China')
select.deselect_all()
# 单选按钮和复选框
radio = driver.find_element(By.ID, 'male')
radio.click()
checkbox = driver.find_element(By.ID, 'agree')
checkbox.click()
# 检查状态
is_selected = checkbox.is_selected()窗口和框架
# 切换窗口
driver.switch_to.window(driver.window_handles[1])
# 切换框架
driver.switch_to.frame('frame_name') # 通过名称
driver.switch_to.frame(0) # 通过索引
driver.switch_to.frame(driver.find_element(By.ID, 'frame_id'))
# 返回主内容
driver.switch_to.default_content()
# 窗口操作
driver.maximize_window()
driver.set_window_size(1024, 768)
driver.set_window_position(100, 100)
# 获取窗口信息
current_window = driver.current_window_handle
all_windows = driver.window_handlesCookie 和 JavaScript
# Cookie 操作
cookies = driver.get_cookies()
driver.add_cookie({'name': 'key', 'value': 'value'})
cookie = driver.get_cookie('key')
driver.delete_cookie('key')
driver.delete_all_cookies()
# 执行 JavaScript
result = driver.execute_script('return document.title;')
driver.execute_script('window.scrollTo(0, document.body.scrollHeight);')
# 高级 JavaScript
js_code = '''
var elements = document.querySelectorAll('.item');
return Array.from(elements).map(el => el.textContent);
'''
items = driver.execute_script(js_code)Playwright
基础使用
from playwright.sync_api import sync_playwright
with sync_playwright() as p:
# 启动浏览器
browser = p.chromium.launch(headless=False)
page = browser.new_page()
# 访问页面
page.goto('https://example.com')
# 查找元素
element = page.query_selector('#my-id')
elements = page.query_selector_all('.my-class')
# 操作元素
element.click()
element.fill('文本内容')
element.select_option('value')
# 等待
page.wait_for_selector('.loaded')
page.wait_for_url('**/success')
# 关闭
browser.close()高级功能
from playwright.sync_api import sync_playwright
with sync_playwright() as p:
browser = p.chromium.launch()
context = browser.new_context(
viewport={'width': 1920, 'height': 1080},
user_agent='Custom User Agent'
)
page = context.new_page()
# 拦截请求
def route_handler(route):
if route.request.resource_type == 'image':
route.abort()
else:
route.continue_()
page.route('**/*', route_handler)
# 监听响应
page.on('response', lambda response: print(f"状态: {response.status}"))
# 执行脚本
page.evaluate('() => document.title')
# 截图和 PDF
page.screenshot(path='screenshot.png')
page.pdf(path='page.pdf')
browser.close()自动化最佳实践
GUI 自动化建议
- 元素定位：优先使用稳定的定位方式
- 等待策略：使用显式等待而非固定延迟
- 异常处理：优雅处理元素未找到的情况
- 日志记录：记录操作步骤便于调试
- 模块化：封装可复用的操作函数
元素定位优先级
# 最稳定
element = driver.find_element(By.ID, 'unique-id')
# 次选
element = driver.find_element(By.NAME, 'unique-name')
# 可用但易变
element = driver.find_element(By.CLASS_NAME, 'class')
element = driver.find_element(By.CSS_SELECTOR, '#id')
element = driver.find_element(By.XPATH, '//div[@class="class"]')
# 最不稳定
element = driver.find_element(By.LINK_TEXT, '点击这里')页面对象模式
class LoginPage:
def __init__(self, driver):
self.driver = driver
self.username_input = (By.ID, 'username')
self.password_input = (By.ID, 'password')
self.login_button = (By.ID, 'login-btn')
def enter_username(self, username):
element = self.driver.find_element(*self.username_input)
element.clear()
element.send_keys(username)
def enter_password(self, password):
element = self.driver.find_element(*self.password_input)
element.send_keys(password)
def click_login(self):
element = self.driver.find_element(*self.login_button)
element.click()
def login(self, username, password):
self.enter_username(username)
self.enter_password(password)
self.click_login()反自动化检测
# 隐藏 webdriver 特征
driver.execute_cdp_cmd('Page.addScriptToEvaluateOnNewDocument', {
'source': '''
Object.defineProperty(navigator, 'webdriver', {
get: () => undefined
})
'''
})
# 设置 User-Agent
options.add_argument('--user-agent=Mozilla/5.0...')
# 使用代理
options.add_argument('--proxy-server=http://proxy:8080')性能优化
# 禁用图片加载
chrome_options = webdriver.ChromeOptions()
prefs = {'profile.managed_default_content_settings.images': 2}
chrome_options.add_experimental_option('prefs', prefs)
# 无头模式
chrome_options.add_argument('--headless')
# 禁用 GPU
chrome_options.add_argument('--disable-gpu')