selenium  pip install selenium

from selenium import webdriver
import time  # 時間模塊, 可以用于程序的延遲
import random  # 隨機數(shù)模塊
from constants import TAO_USERNAME1, TAO_PASSWORD1
import csv  # 數(shù)據(jù)保存的模塊



def search_product(keyword):
    """搜索商品數(shù)據(jù), 登錄用戶"""
    driver.find_element_by_xpath('//*[@id="q"]').send_keys(keyword)
    time.sleep(random.randint(1, 3))  # 盡量避免人機檢測  隨機延遲

    driver.f
def parse_data():
    """解析商品數(shù)據(jù)"""
    divs = driver.find_elements_by_xpath('//div[@class="grid g-clearfx"]/div/div')  #  所有的div標簽

    for div in divs:
        try:
            info = div.find_element_by_xpath('.//div[@class="row row-2 title"]/a').text
            price = div.find_element_by_xpath('.//strong').text + '元'
            deal = div.find_element_by_xpath('.//div[@class="deal-cnt"]').text
            name = div.find_element_by_xpath('.//div[@class="shop"]/a/span[2]').text
            location = div.find_element_by_xpath('.//div[@class="location"]').te'.//div[@class="pic"]/a').get_attribute('href')

            print(info, price, deal, name, location, detail_url)

            # 保存
            with open('某寶.csv', mode='a', encoding='utf-8', newline='') as f:
                csv_write = csv.writer(f)
                csv_write.writerow([info, price, deal, name, location, detail_url])
        except:
            continue


word = input('請輸入你要搜索商品的關(guān)鍵字:')
# 創(chuàng)建一個瀏覽器
driver = webdriver.Chrome()

# selenium操作的瀏覽器被識別了, 無法登錄
# 修改瀏覽器的部分屬性, 繞過檢測
driver.execute_cdp_cmd("Page.addScriptToEvaluateOnNewDocument",
            {"source": """Object.defineProperty(navigator, 'webdriver', {get: () => false})"""})


# 執(zhí)行自動化瀏覽器的操作
driver.get('https://www.taobao.com/')
driver.implicitly_wait(10)  # 設(shè)置瀏覽器的等待,加載數(shù)據(jù)
driver.maximize_window()  # 最大化瀏覽器


# 調(diào)用商品搜索的函數(shù)
search_product(word)

for page in range(100): # 012
    print(f'\n==================正在抓取第{page + 1}頁數(shù)據(jù)====================')
    url = f'https://s.taobao.com/search?q=%E5%B7%B4%E9%BB%8E%E4%B8%96%E5%AE%B6&s={page * 44}'
    # 解析商品數(shù)據(jù)
    parse_data()
    time.sleep(random.randint(1, 3))  # 盡量避免人機檢測  隨機延遲