[python爬虫] Selenium定向爬取PubMed生物医学摘要信息

实现代码

        实现的代码主要是Selenium通过分析网页DOM结点进行爬取。
        爬取的地址是：http://www.medlive.cn/pubmed/
        在网址中搜索Protein（蛋白质）后，分析网址可发现设置Page=1~20可爬取前1~20页的URL信息。链接如下：
        http://www.medlive.cn/pubmed/pubmed_search.do?q=protein&page=1

# coding=utf-8
""" 
Created on 2015-12-05  Ontology Spider
@author Eastmount CSDN
URL:
  http://www.meddir.cn/cate/736.htm
  http://www.medlive.cn/pubmed/
  http://paper.medlive.cn/literature/1502224
"""

import time          
import re          
import os
import shutil
import sys
import codecs 
from selenium import webdriver      
from selenium.webdriver.common.keys import Keys      
import selenium.webdriver.support.ui as ui      
from selenium.webdriver.common.action_chains import ActionChains  

#Open PhantomJS
driver = webdriver.Firefox()
driver2 = webdriver.PhantomJS(executable_path="G:\phantomjs-1.9.1-windows\phantomjs.exe") 
wait = ui.WebDriverWait(driver,10)

'''
 Load Ontoloty
 去到每个生物本体页面下载摘要信息
 http://paper.medlive.cn/literature/literature_view.php?pmid=26637181
 http://paper.medlive.cn/literature/1526876
'''
def getAbstract(num,title,url):
    try:
        fileName = "E:\\PubMedSpider\\" + str(num) + ".txt"
        #result = open(fileName,"w")
        #Error: 'ascii' codec can't encode character u'\u223c'
        result = codecs.open(fileName,'w','utf-8') 
        result.write("[Title]\r\n")
        result.write(title+"\r\n\r\n")
        result.write("[Astract]\r\n")
        driver2.get(url)
        elem = driver2.find_element_by_xpath("//div[@class='txt']/p")
        #print elem.text
        result.write(elem.text+"\r\n")
    except Exception,e:    
        print 'Error:',e
    finally:
        result.close()
        print 'END\n'

'''
 循环获取搜索页面的URL
 规律 http://www.medlive.cn/pubmed/pubmed_search.do?q=protein&page=1
'''
def getURL():
    page = 1      #跳转的页面总数
    count = 1     #统计所有搜索的生物本体个数    
    while page<=20:
        url_page = "http://www.medlive.cn/pubmed/pubmed_search.do?q=protein&page="+str(page)
        print url_page
        driver.get(url_page)
        elem_url = driver.find_elements_by_xpath("//div[@id='div_data']/div/div/h3/a")
        for url in elem_url:
            num = "%05d" % count
            title = url.text
            url_content = url.get_attribute("href")
            print num
            print title
            print url_content
            #自定义函数获取内容
            getAbstract(num,title,url_content)
            count = count + 1
        else:
            print "Over Page " + str(page) + "\n\n"
        page = page + 1
    else:
        "Over getUrl()\n"
        time.sleep(5)

'''
 主函数预先运行
'''
if __name__ == '__main__':
    '''
    path = "F:\\MedSpider\\"
    if os.path.isfile(path):         #Delete file
        os.remove(path)
    elif os.path.isdir(path):        #Delete dir    
        shutil.rmtree(path, True)    
    os.makedirs(path)                #Create the file directory
    '''
    getURL()
    print "Download has finished."

分析HTML

        1.获取每页Page中的20个关于Protein（蛋白质）的URL链接和标题。其中getURL()函数中的核心代码获取URL如下：
        url = driver.find_elements_by_xpath("//div[@id='div_data']/div/div/h3/a")
        url_content = url.get_attribute("href")
        getAbstract(num,title,url_content)

运行结果

        得到的运行结果如下所示：00001.txt~00400.txt共400个txt文件，每个文件包含标题和摘要，该数据集可简单用于生物医学的本体学习、命名实体识别、本体对齐构建等。