爬取instagram留言 - Selenium

这边我是打API爬的，所以先写了序列化：

class IgCommentsSerializer(serializers.Serializer):    post = serializers.CharField(max_length=1000)    poster = serializers.CharField(max_length=200)

一开始先写selenium的webserver基本设定，要抓取全部留言的话需先登入完ig，再跳转到要的post页面：

class IgComments(APIView):    def __init__(self):        self.path = 'chromedriver的路径'        self.sbaccount = '帐号'        self.sbpd = '密码'    def post(self, request):        options = Options()        options.add_argument("--headless") # 执行时浏览器只在背景执行        driver = webdriver.Chrome(self.path, options=options)        driver.implicitly_wait(3)        driver.get('https://www.instagram.com/')         time.sleep(2)        account = driver.find_elements_by_name('username')[0]        pd = driver.find_elements_by_name('password')[0]        # 登入        account.send_keys(self.sbaccount)        pd.send_keys(self.sbpd)        driver.find_element_by_xpath('//*[@id="loginForm"]/div/div[3]/button').click()          time.sleep(3)        driver.get('https://www.instagram.com/p/CYXqAMuBX0e/') # 直接跳转到post        more_xpath = '//*[@id="react-root"]/section/main/div/div[1]/article/div/div[2]/div/div[2]/div[1]/ul/li/div/button/div'        time.sleep(2)

ig一次只会载入12则留言(我没记错的话XD)，这边需要借助selenium的力量，自动化点选更多留言按钮。而为了要抓取所有留言，这边透过while迴圈直到没有更多留言按钮可点选，之后就可以一次性地抓取所有留言文字了：

...接下上部分程式码...        while True:            try:                time.sleep(2)                driver.find_element_by_xpath(more_xpath).click()                print('下一页')            except:                print('最后一页')                break        crawl_comments = []        comments = driver.find_element_by_class_name("XQXOT").find_elements_by_class_name("Mr508")        n = 1        for c in comments:            poster = c.find_element_by_css_selector('h3._6lAjh span').text            post_xpath = f'//*[@id="react-root"]/section/main/div/div[1]/article/div/div[2]/div/div[2]/div[1]/ul/ul[{n}]/div/li/div/div/div[2]/span'.format(n=n)            time.sleep(2)            post = c.find_element_by_xpath(post_xpath).text            crawl_comments.append({'poster':poster, 'post':post})            n+=1        ser = IgCommentsSerializer(crawl_comments, many=True)        return Response(ser.data)