Python爬虫之基于 selenium 实现文献信息获取_python使用selenium爬取ieee和arxiv文献摘要

每个条目包含题目、作者、来源等信息

通过对当前页面分析，发现每条文献条目的 xpath 是有规律的

#题名
/html/body/div[3]/div[2]/div[2]/div[2]/form/div/table/tbody/tr[1]/td[2]

#作者
/html/body/div[3]/div[2]/div[2]/div[2]/form/div/table/tbody/tr[1]/td[3]

tr[1] 表示本页第一条条目，而 td[2] 中的2-6 分别代表作者、来源、发表时间和数据库

我们在当前页面是无法获取到文献的摘要、关键字等信息，需要进一步点击进入相关文献条目

进入到相关文献页面之后，根据 class name来获取摘要、关键字、是否为CSSCI 这些元素

完成以上知网页面的分析后，我们就可以根据需求开始写代码了！

代码实现

导入所需包

import time 
from selenium import webdriver
from selenium.webdriver.support.ui import WebDriverWait
from selenium.webdriver.support import expected_conditions as EC
from selenium.webdriver.common.by import By
from selenium.webdriver.common.desired_capabilities import DesiredCapabilities

创建浏览器对象

这里我用的是 Edge 浏览器

# get直接返回，不再等待界面加载完成
desired_capabilities = DesiredCapabilities.EDGE
desired_capabilities["pageLoadStrategy"] = "none"

# 设置 Edge 驱动器的环境
options = webdriver.EdgeOptions()
# 设置 Edge 不加载图片，提高速度
options.add_experimental_option("prefs", {"profile.managed\_default\_content\_settings.images": 2})

# 创建一个 Edge 驱动器
driver = webdriver.Edge(options=options)

传入 url 参数然后模拟对浏览器进行人为操作

适当的加入 time.sleep() 方法，等待页面加载完成

不然页面还没完全加载就执行下一步操作的话会报错

# 打开页面
driver.get("https://kns.cnki.net/kns8/AdvSearch")
time.sleep(2)
    
# 传入关键字
WebDriverWait(driver, 100).until(
        EC.presence_of_element_located((By.XPATH, '''//\*[@id="gradetxt"]/dd[3]/div[2]/input'''))).send_keys(theme)
time.sleep(2)
    
# 点击搜索
WebDriverWait(driver, 100).until(
        EC.presence_of_element_located((By.XPATH, "/html/body/div[2]/div/div[2]/div/div[1]/div[1]/div[2]/div[2]/input"))).click()
time.sleep(3)

 # 点击切换中文文献
WebDriverWait(driver, 100).until(
        EC.presence_of_element_located((By.XPATH, "/html/body/div[3]/div[1]/div/div/div/a[1]"))).click()
time.sleep(3)

获取总文献数和页数

res_unm = WebDriverWait(driver, 100).until(EC.presence_of_element_located(
        (By.XPATH, "/html/body/div[3]/div[2]/div[2]/div[2]/form/div/div[1]/div[1]/span[1]/em"))).text

# 去除千分位的逗号
res_unm = int(res_unm.replace(",", '')) 
page_unm = int(res_unm / 20) + 1
print(f"共找到 {res\_unm} 条结果, {page\_unm} 页。")

对结果页进行解析

def crawl(driver, papers_need, theme):
    # 赋值序号, 控制爬取的文章数量
    count = 1

    # 当爬取数量小于需求时，循环网页页码
    while count <= papers_need:
        # 等待加载完全，休眠3S
        time.sleep(3)

        title_list = WebDriverWait(driver, 10).until(EC.presence_of_all_elements_located((By.CLASS_NAME, "fz14")))
        # 循环网页一页中的条目
        for i in range(len(title_list)):
            try:
                if count % 20 != 0:
                    term = count % 20  # 本页的第几个条目
                else:
                    term = 20
                title_xpath = f"/html/body/div[3]/div[2]/div[2]/div[2]/form/div/table/tbody/tr[{term}]/td[2]"
                author_xpath = f"/html/body/div[3]/div[2]/div[2]/div[2]/form/div/table/tbody/tr[{term}]/td[3]"
                source_xpath = f"/html/body/div[3]/div[2]/div[2]/div[2]/form/div/table/tbody/tr[{term}]/td[4]"
                date_xpath = f"/html/body/div[3]/div[2]/div[2]/div[2]/form/div/table/tbody/tr[{term}]/td[5]"
                database_xpath = f"/html/body/div[3]/div[2]/div[2]/div[2]/form/div/table/tbody/tr[{term}]/td[6]"
                title = WebDriverWait(driver, 10).until(EC.presence_of_element_located((By.XPATH, title_xpath))).text
                authors = WebDriverWait(driver, 10).until(EC.presence_of_element_located((By.XPATH, author_xpath))).text
                source = WebDriverWait(driver, 10).until(EC.presence_of_element_located((By.XPATH, source_xpath))).text
                date = WebDriverWait(driver, 10).until(EC.presence_of_element_located((By.XPATH, date_xpath))).text
                database = WebDriverWait(driver, 10).until(
                    EC.presence_of_element_located((By.XPATH, database_xpath))).text
                # 点击条目
                title_list[i].click()
                # 获取driver的句柄
                n = driver.window_handles
                # driver切换至最新生产的页面
                driver.switch_to.window(n[-1])
                time.sleep(3)
                # 开始获取页面信息
                title = WebDriverWait(driver, 10).until(EC.presence_of_element_located(
                    (By.XPATH, "/html/body/div[2]/div[1]/div[3]/div/div/div[3]/div/h1"))).text
                authors = WebDriverWait(driver, 10).until(EC.presence_of_element_located(
                    (By.XPATH, "/html/body/div[2]/div[1]/div[3]/div/div/div[3]/div/h3[1]"))).text
                institute = WebDriverWait(driver, 10).until(EC.presence_of_element_located(
                    (By.XPATH, "/html/body/div[2]/div[1]/div[3]/div/div/div[3]/div/h3[2]"))).text
                abstract = WebDriverWait(driver, 10).until(
                    EC.presence_of_element_located((By.CLASS_NAME, "abstract-text"))).text
                try:
                    keywords = WebDriverWait(driver, 10).until(
                        EC.presence_of_element_located((By.CLASS_NAME, "keywords"))).text[:-1]
                    cssci = WebDriverWait(driver, 10).until(
                        EC.presence_of_element_located((By.XPATH, "/html/body/div[2]/div[1]/div[3]/div/div/div[1]/div[1]/a[2]"))).text
                except:
                    keywords = '无'
                    cssci = 'NULL'
                url = driver.current_url

                # 写入文件
                res = f"{count}\t{title}\t{authors}\t{cssci}\t{institute}\t{date}\t{source}\t{database}\t{keywords}\t{abstract}\t{url}".replace(
                        "\n", "") + "\n"
                print(res)
                with open(f'{theme}.tsv', 'a', encoding='gbk') as f:
                    f.write(res)
            except:
                print(f" 第{count} 条爬取失败\n")
                # 跳过本条，接着下一个
                continue
            finally:
                # 如果有多个窗口，关闭第二个窗口， 切换回主页
                n2 = driver.window_handles
                if len(n2) > 1:
                    driver.close()
                    driver.switch_to.window(n2[0])
                # 计数,判断篇数是否超出限制
                count += 1
        if count == papers_need:
            break
        else:
            # 切换到下一页
            WebDriverWait(driver, 10).until(EC.presence_of_element_located((By.XPATH, "//a[@id='PageNext']"))).click()

结果展示：
结果是一个以制表符分隔的表格文件（用 excel 打开），其中包含了论文的基本信息，包括：题目、作者、是否 CSSCI、来源、摘要等

完整代码如下：

import time
from selenium import webdriver
from selenium.webdriver.support.ui import WebDriverWait
from selenium.webdriver.support import expected_conditions as EC
from selenium.webdriver.common.by import By
from selenium.webdriver.common.desired_capabilities import DesiredCapabilities
from urllib.parse import urljoin


def open\_page(driver, theme):
    # 打开页面
    driver.get("https://kns.cnki.net/kns8/AdvSearch")
    time.sleep(2)

    # 传入关键字
    WebDriverWait(driver, 100).until(
        EC.presence_of_element_located((By.XPATH, '''//\*[@id="gradetxt"]/dd[3]/div[2]/input'''))).send_keys(theme)
    time.sleep(2)

    # 点击搜索
    WebDriverWait(driver, 100).until(
        EC.presence_of_element_located((By.XPATH, "/html/body/div[2]/div/div[2]/div/div[1]/div[1]/div[2]/div[2]/input"))).click()
    time.sleep(3)

    # 点击切换中文文献
    WebDriverWait(driver, 100).until(
        EC.presence_of_element_located((By.XPATH, "/html/body/div[3]/div[1]/div/div/div/a[1]"))).click()
    time.sleep(3)

    # 跳转到第三页
    # WebDriverWait(driver, 100).until(
    # EC.presence\_of\_element\_located((By.XPATH, "/html/body/div[3]/div[2]/div[2]/div[2]/form/div/div[2]/a[2]"))).click()
    # time.sleep(3)

    # 获取总文献数和页数
    res_unm = WebDriverWait(driver, 100).until(EC.presence_of_element_located(
        (By.XPATH, "/html/body/div[3]/div[2]/div[2]/div[2]/form/div/div[1]/div[1]/span[1]/em"))).text

    # 去除千分位里的逗号
    res_unm = int(res_unm.replace(",", ''))
    page_unm = int(res_unm / 20) + 1
    print(f"共找到 {res\_unm} 条结果, {page\_unm} 页。")
    return res_unm


def crawl(driver, papers_need, theme):
    # 赋值序号, 控制爬取的文章数量
    count = 1

    # 当爬取数量小于需求时，循环网页页码
    while count <= papers_need:
        # 等待加载完全，休眠3S
        time.sleep(3)

        title_list = WebDriverWait(driver, 10).until(EC.presence_of_all_elements_located((By.CLASS_NAME, "fz14")))
        # 循环网页一页中的条目
        for i in range(len(title_list)):
            try:
                if count % 20 != 0:
                    term = count % 20  # 本页的第几个条目
                else:
                    term = 20
                title_xpath = f"/html/body/div[3]/div[2]/div[2]/div[2]/form/div/table/tbody/tr[{term}]/td[2]"
                author_xpath = f"/html/body/div[3]/div[2]/div[2]/div[2]/form/div/table/tbody/tr[{term}]/td[3]"
                source_xpath = f"/html/body/div[3]/div[2]/div[2]/div[2]/form/div/table/tbody/tr[{term}]/td[4]"
                date_xpath = f"/html/body/div[3]/div[2]/div[2]/div[2]/form/div/table/tbody/tr[{term}]/td[5]"
                database_xpath = f"/html/body/div[3]/div[2]/div[2]/div[2]/form/div/table/tbody/tr[{term}]/td[6]"
                title = WebDriverWait(driver, 10).until(EC.presence_of_element_located((By.XPATH, title_xpath))).text
                authors = WebDriverWait(driver, 10).until(EC.presence_of_element_located((By.XPATH, author_xpath))).text
                source = WebDriverWait(driver, 10).until(EC.presence_of_element_located((By.XPATH, source_xpath))).text
                date = WebDriverWait(driver, 10).until(EC.presence_of_element_located((By.XPATH, date_xpath))).text
                database = WebDriverWait(driver, 10).until(
                    EC.presence_of_element_located((By.XPATH, database_xpath))).text
                # 点击条目
                title_list[i].click()
                # 获取driver的句柄
                n = driver.window_handles
                # driver切换至最新生产的页面
                driver.switch_to.window(n[-1])
                time.sleep(3)
                # 开始获取页面信息
                title = WebDriverWait(driver, 10).until(EC.presence_of_element_located(
                    (By.XPATH, "/html/body/div[2]/div[1]/div[3]/div/div/div[3]/div/h1"))).text
                authors = WebDriverWait(driver, 10).until(EC.presence_of_element_located(
                    (By.XPATH, "/html/body/div[2]/div[1]/div[3]/div/div/div[3]/div/h3[1]"))).text
                institute = WebDriverWait(driver, 10).until(EC.presence_of_element_located(
                    (By.XPATH, "/html/body/div[2]/div[1]/div[3]/div/div/div[3]/div/h3[2]"))).text
                abstract = WebDriverWait(driver, 10).until(
                    EC.presence_of_element_located((By.CLASS_NAME, "abstract-text"))).text
                try:
                    keywords = WebDriverWait(driver, 10).until(
                        EC.presence_of_element_located((By.CLASS_NAME, "keywords"))).text[:-1]
                    cssci = WebDriverWait(driver, 10).until(
                        EC.presence_of_element_located((By.XPATH, "/html/body/div[2]/div[1]/div[3]/div/div/div[1]/div[1]/a[2]"))).text
                except:
                    keywords = '无'
                    cssci = 'NULL'
                url = driver.current_url
                # 获取下载链接
                # link = WebDriverWait( driver, 10 ).until( EC.presence\_of\_all\_elements\_located((By.CLASS\_NAME ,"btn-dlcaj") ) )[0].get\_attribute('href')
                # link = urljoin(driver.current\_url, link)

                # 写入文件
                res = f"{count}\t{title}\t{authors}\t{cssci}\t{institute}\t{date}\t{source}\t{database}\t{keywords}\t{abstract}\t{url}".replace(
                        "\n", "") + "\n"
                print(res)
                # with open(f'CNKI\_{theme}.tsv', 'a', encoding='gbk') as f:
                # f.write(res)
                count += 1
                if count > papers_need:
                    break
            except:
                print(f" 第{count} 条爬取失败\n")
                # 跳过本条，接着下一个
                continue
            finally:
                # 如果有多个窗口，关闭第二个窗口， 切换回主页
                n2 = driver.window_handles
                if len(n2) > 1:
                    driver.close()
                    driver.switch_to.window(n2[0])
                # 计数,判断需求是否足够
                # count += 1
        if count > papers_need:
            break
        else:
            # 切换到下一页
            WebDriverWait(driver, 10).until(EC.presence_of_element_located((By.XPATH, "//a[@id='PageNext']"))).click()


def webserver(theme):
    # get直接返回，不再等待界面加载完成
    desired_capabilities = DesiredCapabilities.EDGE
    desired_capabilities["pageLoadStrategy"] = "none"

    # 设置微软驱动器的环境
    options = webdriver.EdgeOptions()
    # 设置chrome不加载图片，提高速度
    options.add_experimental_option("prefs", {"profile.managed\_default\_content\_settings.images": 2})

    # 创建一个微软驱动器
    driver = webdriver.Edge(options=options)

    # 设置所需篇数
    papers_need = 41

    res_unm = int(open_page(driver, theme))

    # 判断所需是否大于总篇数
    papers_need = papers_need if (papers_need <= res_unm) else res_unm
    return driver, papers_need, theme


if __name__ == "\_\_main\_\_":
    # 输入需要搜索的内容
    # theme = input("请输入你要搜索的期刊名称：")
    print('1')
    theme = "中国人口·资源与环境"
    driver, papers_need, theme = webserver(theme)
    print("所需篇数: ", papers_need)
    crawl(driver, papers_need, theme)
    # 关闭浏览器
    driver.close()

踩过的坑

网页加载太慢导致元素查找出错

现在能在网上找到很多很多的学习资源，有免费的也有收费的，当我拿到1套比较全的学习资源之前，我并没着急去看第1节，我而是去审视这套资源是否值得学习，有时候也会去问一些学长的意见，如果可以之后，我会对这套学习资源做1个学习计划，我的学习计划主要包括规划图和学习进度表。

分享给大家这份我薅到的免费视频资料，质量还不错，大家可以跟着学习

网上学习资料一大堆，但如果学到的知识不成体系，遇到问题时只是浅尝辄止，不再深入研究，那么很难做到真正的技术提升。

需要这份系统化学习资料的朋友，可以戳这里无偿获取

一个人可以走的很快，但一群人才能走的更远！不论你是正从事IT行业的老鸟或是对IT行业感兴趣的新人，都欢迎加入我们的的圈子（技术交流、学习资源、职场吐槽、大厂内推、面试辅导），让我们一起学习成长！