## 前言 總算入行上班幾個月了，不得不說業務是真的不消停啊。。 本人工作上經常遇到一種場景：為甲方做自動化介面處理工具，登入需要簡訊驗證碼，， 嘛算是摸索出了一套selenium程式碼模板，主要解決如下痛點 + 會話超時/斷開時，又要找甲方問簡訊等驗證碼登入 + 除錯途中增減修改功能，算是除錯中熱更新 分享一下 ## 模板程式碼 ### app.py ```python #!/usr/bin/python # -*- coding: utf-8 -*- import os import importlib from selenium import webdriver from selenium.webdriver import ActionChains from selenium.webdriver.common.by import By from selenium.webdriver.support.ui import WebDriverWait from selenium.webdriver.support import expected_conditions as EC import backend basepath = os.path.abspath(os.path.dirname(__file__)) driver_path = os.path.join(basepath, 'chromedriver.exe') logger = backend.logger def init_browser(driver_path=None): options = webdriver.ChromeOptions() options.add_argument('--no-sandbox') options.add_argument('--disable-gpu') prefs = { 'profile.default_content_setting_values': { 'notifications': 2 }} options.add_experimental_option('prefs', prefs) options.add_experimental_option('excludeSwitches', ['enable-automation']) options.add_experimental_option("useAutomationExtension", False) browser = webdriver.Chrome(options=options, executable_path=driver_path) browser.maximize_window() browser.execute_cdp_cmd("Page.addScriptToEvaluateOnNewDocument", { "source": """ Object.defineProperty(navigator, 'webdriver', { get: () => undefined }) """ }) return browser def jump_security(wait, mouse): wait.until(EC.presence_of_element_located((By.ID, 'details-button'))).click() ele = wait.until(EC.presence_of_element_located((By.ID, 'proceed-link'))) mouse.move_to_element(ele).click().perform() def init_login(driver, wait, mouse): username_inp = wait.until(EC.presence_of_element_located((By.ID, "username"))) username_inp.send_keys("user") password_inp = driver.find_element_by_id("password") password_inp.send_keys("password") class App(object): def __new__(cls, *args, **kwargs): if not hasattr(cls, '_instance'): cls.error_num = 0 cls.driver = init_browser(driver_path) cls.wait = WebDriverWait(cls.driver, 20) cls.mouse = ActionChains(cls.driver) cls.driver.get('https://www.target.com/login') # jump_security(cls.wait, cls.mouse) init_login(cls.driver, cls.wait, cls.mouse) cls._instance = object.__new__(cls) return cls._instance # 模式1：client無限迴圈 def run_unlimited(): while True: try: obj = App() input('等待登入並進入目標頁面後，回此處按回車 >>> ') back = backend.Backend(obj) results = back.main() except Exception as e: pass finally: mode = input('供backend修改的阻塞暫停') importlib.reload(backend) # 模式2：構建本地api服務 from flask import Flask app = Flask(__name__) @app.route("/", methods=["GET"]) def main(): importlib.reload(backend) back = backend.Backend(App()) results = back.main() if __name__ == '__main__': os.system('taskkill /im chromedriver.exe /F') # win專用，清殘留程序 os.system('taskkill /im chrome.exe /F') run_unlimited() # app.run() ``` 前端有兩部分，一是單例的selenium，二是此自動化處理工具的形式：client迴圈形式 / api服務形式 1. 單例的 \__new__ 裡init一些屬性，處理登入那部分也可以放後臺 2. 兩種形式其實就是看形式是要主動觸發還是被動觸發，至於具體做什麼就放後臺 --- ### backend.py ```python #!/usr/bin/python # -*- coding: utf-8 -*- import json import os import re from concurrent.futures import ThreadPoolExecutor, as_completed import requests import simplejson from loguru import logger from retry import retry from tqdm import tqdm, trange import urllib3 urllib3.disable_warnings(urllib3.exceptions.InsecureRequestWarning) basepath = os.path.abspath('./') logger.add(f'{basepath}/logs/{os.path.basename(__file__)[:-3]}.log', format="{level} | {time:YYYY-MM-DD HH:mm:ss} | {function}:{line} - {message}", level="INFO", retention='5 days') class Backend(object): def __init__(self, obj): self.sess = requests.session() self.driver = obj.driver self.sess.headers = {'Accept': 'application/json, text/javascript, */*; q=0.01', 'Accept-Encoding': 'gzip, deflate', 'Accept-Language': 'zh-Hans-CN, zh-Hans; q=0.5', 'Cache-Control': 'no-cache', 'Connection': 'Keep-Alive', 'Content-Length': '561', 'Content-Type': 'application/x-www-form-urlencoded; charset=UTF-8', 'Cookie': 'SESSION=abcdefg', 'Host': 'www.target.com', 'Referer': 'https://www.target.com/path', 'User-Agent': 'Mozilla/5.0 (Windows NT 10.0; WOW64; Trident/7.0; rv:11.0) like Gecko', 'X-Requested-With': 'XMLHttpRequest' } def get_cookie(self): self.driver.find_element_by_xpath('//input[@class="e.g:trigger btn"]').click() cookies = {_["name"]: _["value"] for _ in self.driver.get_cookies()} return cookies def get_headers(self): cookies = self.get_cookie() token = self.driver.execute_script('return window.sessionStorage.getItem("token")') self.sess.headers.update({ 'Authorization': token, 'Cookie': f'SESSION={cookies["SESSION"]}; acw_tc={cookies["acw_tc"]}' }) @retry((json.decoder.JSONDecodeError, simplejson.errors.JSONDecodeError, KeyError, ValueError), tries=3, delay=1) def do_api(self): url = 'https://www.target.com/api/path' payload = { 'params': '31b1xu0', } self.get_headers() resp = self.sess.post(url, json=payload, verify=False, timeout=10) if resp.status_code == 200: self.pre_api_task(resp.json()) # do what you need todo else: raise ValueError(f'do_api failed:: {resp.text}') def do_selenium_command(self): self.driver.execute_script("$('p[class=imgShow]').click()") self.driver.execute_script("document.getElementsByClassName('supportRadioOptional1 checked')[0].click();") pagenum = int(re.search(r'共 (\d+) 頁', self.driver.page_source).group(1)) for _ in trange(pagenum, ncols=40): self.pre_page_task() # do what you need todo self.driver.execute_script(f"PaginationpageTable.gotoPage('next', '{_+2}', '50');") def main(self): self.do_selenium_command() self.do_api() if __name__ == '__main__': requests.get('127.0.0.1:5000') ``` 基於前面說的簡訊驗證碼，讓甲方登入後selenium一頓操作就把api的headers補完了，可以愉快地請求介面了 需要js取引數的話可以這樣寫`token = self.driver.execute_script('return window.sessionStorage.getItem("token")')` 目前遇到的一些注意點： 1. 渲染的頁面帶frame，需要switch_to再xpath等處理，可把`driver.page_source`寫進檔案判斷是否該目標頁順帶測定位 2. 有時`driver.find_element_by_*`無法定位，試試用js；有些JS/Jquery功能在老版IE上用不了，回用mouse處理（套娃呢喂）；連續使用js時要注意響應等待時間 + basepath處用`'./'`取巧了一下(與pyinstaller打包有關)，可以基於此變數做一些本地檔案處理 --- ## Last 畢竟最終是為甲方做的，程式要以甲方裝置為準 即使它是win7,用pywin32定位控制代碼出現相容問題即使業務網站只相容IE核心，js部分功能無法用頭髮掉光了啊 畢竟是個人摸索出的，可能有更優解，如大佬路過還請不要吝嗇交(p)流(y)一