2023.12.12 已修复错误，感谢小伙伴指出错误！！！

京东有反爬措施，爬取需要登陆，代码第一次使用需要手动登录依次获取cookie（自动保存至当前目录下，cookie相当于密码，注意保护好隐私！！！），可以根据自己网速设置相应的等待时间（这里默认10s）,之后再运行可直接自动登录。

ps:悄悄说一句，登录过于频繁会被限制哦，所以尽量一次登录成功。

效果如图：

import time
import re
import csv
import json
from selenium import webdriver
from selenium.webdriver.chrome.options import Options
from bs4 import BeautifulSoup
import os
# 可使用Chrome浏览器驱动程序并将其设置为无头模式
chrome_options = Options()
driver = webdriver.Chrome(options=chrome_options)
chrome_options.add_argument("--disable-gpu")
# 可使用Chrome浏览器驱动程序并将其设置为无头模式
#chrome_options.add_argument("--headless")
chrome_options.add_argument("--no-sandbox")
chrome_options.add_argument("--disable-dev-shm-usage")

driver.get("https://passport.jd.com/new/login.aspx?/")
cookie_file = 'cookie.txt'
# 判断是否有 cookie.txt 文件
if os.path.exists(cookie_file):
    # 读取cookie文件中的内容
    driver.get(f"https://www.jd.com/")
    time.sleep(2)
    with open(cookie_file, 'r') as file:
        # 读取文件中的 cookie
        cookies = json.load(file)
        # 加载cookie信息
        for cookie in cookies:
            driver.add_cookie(cookie)
    print('使用已保存的cookie登录')
else:
    driver.get("https://passport.jd.com/new/login.aspx?/")
    # 等待用户登录并获取cookie
    time.sleep(10)  # 第一次使用需要用户手动登录获取cookie,可根据网络状况修改
    dictcookies = driver.get_cookies()
    jsoncookies = json.dumps(dictcookies)
    with open('cookie.txt', 'w') as f:
        f.write(jsoncookies)
    print('cookies已保存')

# 发送请求并检索网页内容
key="人工智能"  #搜索关键词
url = (f"https://search.jd.com/Search?keyword={key}")
driver.get(url)

# 初始化变量
page_number = 1
max_pages = 10  # 设置要爬取的页面数，爬的页数太多会被反爬机制阻止，建议设置sleep
data = []

while page_number <= max_pages:
    print("正在爬取第", page_number, "页")

    # 检索页面完全加载后的html内容
    html_content = driver.page_source

    # 使用BeautifulSoup解析html内容
    soup = BeautifulSoup(html_content, "html.parser")

    # 查找所有包含产品信息的class为“gl-i-wrap”的div
    div_list = soup.find_all("div", class_="gl-i-wrap")

    # 从每个div中提取文本信息
    for div in div_list:
        name = div.find("div", class_="p-name").get_text().strip().replace("\n","")
        price = div.find("div", class_="p-price").get_text().strip()
        commit = div.find("div", class_="p-commit").get_text().strip()
        commit = commit.replace('条评价', '').replace('+', '')
        if '万' in commit:
            commit = float(commit.replace('万', '')) * 10000

        # 模拟点击书名，获取新页面中的信息
        link = div.find("div", class_="p-name").find("a").get("href")
        if "http" not in link:
            link = "https:" + link

        # 打开新标签页
        driver.execute_script(f'''window.open("{link}","_blank");''')
        # 切换到新标签页
        windows = driver.window_handles
        driver.switch_to.window(windows[-1])
        time.sleep(5)
        soup_new = BeautifulSoup(driver.page_source, "html.parser")

        time.sleep(6)
        publisher = soup_new.find("li", title=True, clstag="shangpin|keycount|product|chubanshe_3")["title"] \
            if soup_new.find("li", title=True, clstag="shangpin|keycount|product|chubanshe_3") is not None else '未找到'
        publish_date_tag = soup_new.find("li", string=re.compile(r"出版时间："))
        publish_date = publish_date_tag.get_text().replace('出版时间：', '').strip() if publish_date_tag else '未找到'
        # 点击“商品评价”按钮
        shop_button = driver.find_elements_by_xpath("//*[@id='detail']/div[1]/ul/li[5]")[0]
        shop_button.click()
        time.sleep(2)  # 爬取并输出评价信息（好评、中评、差评数目）
        good_comments = driver.find_elements_by_xpath("//*[@id='comment']/div[2]/div[2]/div[1]/ul/li[5]/a/em")
        for comment in good_comments:
            comment_text = comment.text.strip("()+")
            if "万" in comment_text:
                comment_text = str(int(float(comment_text.strip("万")) * 10000))
            good_comments_count = int(comment_text)

        medium_comments = driver.find_elements_by_xpath("//*[@id='comment']/div[2]/div[2]/div[1]/ul/li[6]/a/em")
        for comment in medium_comments:
            comment_text = comment.text.strip("()+")
            if "万" in comment_text:
                comment_text = str(int(float(comment_text.strip("万")) * 10000))
            medium_comments_count = int(comment_text)

        bad_comments = driver.find_elements_by_xpath("//*[@id='comment']/div[2]/div[2]/div[1]/ul/li[7]/a/em")
        for comment in bad_comments:
            comment_text = comment.text.strip("()+")
            if "万" in comment_text:
                comment_text = str(int(float(comment_text.strip("万")) * 10000))
            bad_comments_count = int(comment_text)
        driver.close()
        driver.switch_to.window(windows[0])

        # 将数据添加到列表中
        info = {
            "书名": name,
            "价格": price,
            "评论数": commit,
            "出版社": publisher,
            "出版年份": publish_date,
            "好评": good_comments_count,
            "中评": medium_comments_count,
            "差评": bad_comments_count
        }
        data.append(info)

        # 打印书籍信息
        print(info)

    # 点击下一页按钮（如果可用）
    next_page_button = driver.find_element_by_class_name("pn-next")
    if next_page_button:
        next_page_button.click()
        time.sleep(3)  # 延迟以完全加载下一页
    else:
        break

    page_number += 1

# 关闭浏览器驱动程序
driver.quit()
# 读取JSON文件中的cookie信息
with open('cookie.txt', 'r') as file:
    cookies = json.load(file)

# 使用cookie创建WebDriver对象
driver = webdriver.Chrome(options=chrome_options)
driver.get(url)

# 将cookie信息添加到WebDriver对象中
for cookie in cookies:
    driver.add_cookie(cookie)
# 将数据保存到CSV文件中
filename = "book_info.csv"
fields = ["书名", "价格", "评论数", "出版社", "出版年份", "好评", "中评", "差评"]

with open(filename, 'w', newline='', encoding='utf-8') as csvfile:
    writer = csv.DictWriter(csvfile, fieldnames=fields)
    writer.writeheader()
    writer.writerows(data)

print("数据已保存到", filename)

感谢朋友们阅读，下期再见！！！