Блог

Я делаю page scraper и хочу останавливать загрузку страницы каждые 2 секунды. Поэтому я попытался нажать кнопку ESCAPE с помощью actionchains, но это никогда не работает. Ошибки не возникает, но ничего не произошло. Вот мой код ниже. (Я разорвал часть только для отправки ESCAPE-ключа) Заранее благодарю вас!

• Я загрузил свой полный код! извините за мой грязный код (…) Пожалуйста, будьте великодушны в этом!

 from selenium import webdriver
from selenium.webdriver.chrome.options import Options
from openpyxl import load_workbook
from selenium.webdriver.support.ui import WebDriverWait
from selenium.webdriver.common.by import By
from selenium.webdriver.support import expected_conditions as EC
from selenium.webdriver.common.keys import Keys
from selenium.webdriver.common.action_chains import ActionChains
from selenium.common.exceptions import NoSuchElementException
import urllib.request
import time
import pandas as pd

wb = load_workbook(filename = 'C:/Users/yoosimyung/Desktop/Project/selenium/Taobao/Info_Taobao.xlsx')
sheet = wb['Furniture_Table']

chrome_options = Options()
chrome_options.add_experimental_option("debuggerAddress", "127.0.0.1:9222")
chrome_driver = "C:/Users/yoosimyung/Desktop/Project/selenium/chromedriver.exe"
driver = webdriver.Chrome(chrome_driver, options=chrome_options)

Max = driver.find_element_by_css_selector("input.J_Input").get_attribute("max")
MaxN = int(Max)
print(Max)

p = 2
for i in range(1, MaxN 1):
   
    driver.execute_script("window.scrollTo(0, document.body.scrollHeight)")
    time.sleep(1)

    scheight = .01
    while scheight < 9:
        driver.execute_script("window.scrollTo(0, document.body.scrollHeight/%s);" % scheight)
        scheight  = .013

    th_images = driver.find_elements_by_css_selector(".item.J_MouserOnverReq:not(.item-ad) .J_ItemPic.img")
    count_th = 1
    for image in th_images:
        th_url = driver.find_elements_by_css_selector(".item.J_MouserOnverReq:not(.item-ad) .J_ItemPic.img")[count_th-1].get_attribute("src")
        urllib.request.urlretrieve(th_url, str(i)   "_"   str(count_th)   "_TH.jpg")
        count_th = count_th   1

    count_detail = 1
    for detail in th_images:

        try:
            driver.find_elements_by_css_selector(".item.J_MouserOnverReq:not(.item-ad) .row.row-2.title .J_ClickStat")[count_detail-1].click()
        except:
            driver.find_elements_by_css_selector(".item.J_MouserOnverReq:not(.item-ad) .J_ItemPic.img")[count_detail-1].click()
        driver.switch_to.window(driver.window_handles[-1])
       
        Currenturl = driver.current_url
        url = str(Currenturl)

        time.sleep(1)
        webdriver.Actionchains(driver).send_keys(Keys.ESCAPE).perform()

        if 'item.taobao.com' in url:
            try:
                driver.find_element_by_id("bd")
                sheet['I'   str(p)].value = url
                sheet['J'   str(p)].value = "=HYPERLINK(I"   str(p)   ")"

                scheight = .02
                while scheight < 8.5:
                    driver.execute_script("window.scrollTo(0, document.body.scrollHeight/%s);" % scheight)
                    scheight  = .022
                    time.sleep(0.005)

                contentPage = driver.find_element_by_id("J_DivItemDesc")
                detail_url = contentPage.find_elements_by_tag_name("img")
                for k in range(len(detail_url)):
                    try:
                        urllib.request.urlretrieve(detail_url[k].get_attribute("src"), str(i)   "_"   str(count_detail)   "_detail_"   str(k)   ".jpg")
                    except:
                        pass

                PDtext = driver.find_element_by_css_selector(".tb-main-title").text
                PDprice = driver.find_element_by_css_selector(".tb-rmb-num").text
                sheet['A'   str(p)].value = str(i)   "_"   str(count_detail)
                sheet['B'   str(p)].value = PDtext
                print(PDtext)
                if "-" in PDprice:
                    sheet['H'   str(p)].value = PDprice.split("-")[1]
                    print(PDprice.split("-")[1])
                else:
                    sheet['H'   str(p)].value = PDprice
                    print(PDprice)
                print(Currenturl)

            except NoSuchElementException:
                sheet['A'   str(p)].value = str(i)   "_"   str(count_detail)
                sheet['B'   str(p)].value = "PASS"

        else:
            try:
                driver.find_element_by_id("content")
                sheet['I'   str(p)].value = url
                sheet['J'   str(p)].value = "=HYPERLINK(I"   str(p)   ")"
   
                scheight = .02
                while scheight < 8.5:
                    driver.execute_script("window.scrollTo(0, document.body.scrollHeight/%s);" % scheight)
                    scheight  = .022
                    time.sleep(0.005)
               
                contentPage = driver.find_element_by_id("description")
                detail_url = contentPage.find_elements_by_class_name("img-ks-lazyload:not(.desc_anchor)")
                for k in range(len(detail_url)):
                    try:
                        urllib.request.urlretrieve(detail_url[k].get_attribute("src"), str(i)   "_"   str(count_detail)   "_detail_"   str(k)   ".jpg")
                    except:
                        pass

                PDtext = driver.find_element_by_css_selector(".tb-detail-hd > h1").text
                PDprice = driver.find_element_by_css_selector(".tm-price").text
                sheet['A'   str(p)].value = str(i)   "_"   str(count_th)
                sheet['B'   str(p)].value = PDtext
                print (PDtext)
                if "-" in PDprice:
                    sheet['H'   str(p)].value = PDprice.split("-")[1]
                    print(PDprice.split("-")[1])
                else:
                    sheet['H'   str(p)].value = PDprice
                    print(PDprice)
                print(Currenturl)

            except NoSuchElementException:
                sheet['A'   str(p)].value = str(i)   "_"   str(count_detail)
                sheet['B'   str(p)].value = "PASS"

        driver.close()
        driver.switch_to.window(driver.window_handles[0])
        count_detail = count_detail   1
        p = p   1
        wb.save(filename = 'C:/Users/yoosimyung/Desktop/Project/selenium/Taobao/Info_Taobao.xlsx')

    i = i   1
    driver.find_element_by_css_selector(".btn.J_Submit").click()
 

driver.find_element_by_xpath("//body").send_keys(Keys.ESCAPE)

Отправьте ключи в элемент body, если вы хотите остановить загрузку страницы и все