Spider_lxml
xpath工具(解析)
- xpath
- 在XML文档中查找信息的语言,同样适用于HTML文档的检索
- xpath辅助工具
- Chrome插件 :XPath Helper
- 打开 :Ctrl + Shift + X
- 关闭 :Ctrl + Shift + X
- Firefox插件 :XPath checker
- XPath表达式编辑工具 :XML quire
- Chrome插件 :XPath Helper
xpath匹配规则
- 匹配演示
- 查找bookstore下所有节点:
/bookstore
- 查找所有的book节点:
//book
- 查找所有book下的title节点中,lang属性为"en"的节点
//book/title[@lang="en"]
- 查找bookstore下的第2个book节点下的title节点:
/bookstore/book[2]/title/text()
- 查找bookstore下所有节点:
- 选取节点
- / :从根节点开始选取
- //:从整个文档中查找节点
//price 、 /bookstore/book//price
- @ :选取某个节点的属性
//title[@lang="en"]
- @的使用
- 匹配多路径
- 符号:
|
- 获取所有book节点下的 title节点和price节点
//book/title | //book/price
- 符号:
- 函数
contains()
:匹配一个属性值中包含某些字符串的节点//title[contains(@lang,"e")]
text()
//title[contains(@lang,"e")]/text()
lxml库及xpath使用
lxml库 :HTML/XML解析库
- 安装
python -m pip install lxml
conda install lxml
- 使用流程
- 导模块
from lxml import etree
- 利用lxml库的etree模块创建解析对象
parseHtml = etree.HTML(html)
- 解析对象调用xpath工具定位节点信息
r_list = parseHtml.xpath('xpath表达式')
只要调用了xpath,结果一定是列表
- 如何获取节点对象的内容
节点对象.text
- 导模块
from lxml import etree
html = """"""
# 构造解析对象
parseHtml = etree.HTML(html)
# 利用解析对象调用xpath匹配
r1 = parseHtml.xpath('//a/@href')
#print(r1)
# 获取 /
r2 = parseHtml.xpath('//a[@id="channel"]/@href')
#print(r2)
# 获取非 /
r3 = parseHtml.xpath('//ul[@id="nav"]//a/@href')
#print(r3)
# 获取所有 a 节点的文本内容
r4 = parseHtml.xpath('//a/text()')
#print(r4)
# 获取 图片、军事 ...
r5 = parseHtml.xpath('//ul[@id="nav"]//a')
for i in r5:
print(i.text)
案例1 :抓取百度贴吧帖子里面所有的图片
- 目标:抓取指定贴吧所有图片
- 思路:
- 获取贴吧主页URL,下一页:找URL规律
- 获取1页中每个帖子的URL
- 对每个帖子URL发请求,获取帖子里图片URL
- 对图片URL发请求,以wb方式写入本地文件
- 步骤
- 获取贴吧主页URL
http://tieba.baidu.com/f? + 查询参数
- 找到页面中所有帖子的URL
- src : 完整链接
- href : 和主URL进行拼接
- /p/5926064184
- http://tieba.baidu.com/p/5926064184
- xpath匹配链接:
写法1:
//div[@class="col2_right j_threadlist_li_right"]/div/div/a/@href
- 写法2(推荐):
//div[@class="t_con cleafix"]/div/div/div/a/@href
- 找每个帖子中图片URL
- Xpath匹配:
//img[@class="BDE_Image"]/@src
- Xpath匹配:
- 代码实现
- 获取贴吧主页URL
'''百度贴吧图片抓取'''
import requests
from lxml import etree
class TiebaImageSpider:
def __init__(self):
self.headers = {"User-Agent": "Mozilla/5.0"}
self.baseurl = 'https://tieba.baidu.com'
self.pageurl = 'https://tieba.baidu.com/f?'
# 获取所有帖子URL列表
def get_page_url(self, params):
res = requests.get(self.pageurl, params=params, headers=self.headers)
res.encoding = 'utf-8'
print(res.url)
html = res.text
parse_html = etree.HTML(html)
t_list = parse_html.xpath('//div[contains(@class,"t_con")]/div/div/div/a/@href')
print(t_list)
for t_link in t_list:
t_link = self.baseurl + t_link
self.get_image_url(t_link)
# 获取帖子中图片URL列表
def get_image_url(self, t_link):
res = requests.get(t_link, headers=self.headers)
res.encoding = 'utf-8'
html = res.text
parse_html = etree.HTML(html)
img_list = parse_html.xpath('//img[@class="BDE_Image"]/@src')
for img_link in img_list:
self.write_image(img_link)
# 保存到本地
def write_image(self, img_link):
res = requests.get(img_link, headers=self.headers)
res.encoding = 'utf-8'
html = res.content
filename = img_link[-12:]
with open(filename, 'wb') as f:
f.write(html)
print("%s下载成功" % filename)
# 主函数
def main(self):
name = input("请输入贴吧名字:")
begin = int(input("请输入起始页:"))
end = int(input("请输入中止页:"))
for n in range(begin, end + 1):
pn = (n - 1) * 50
params = {'kw': name,
'pn': str(pn)}
self.get_page_url(params)
# s = '//div[@class="t_con cleafix"]/div/div/div/a/@href'
# s = '//*[@id="thread_list"]/li/div/div[2]/div[1]/div[1]/a/@href'
# s2 = '//cc/div/img[@class="BDE_Image"]/@src'
if __name__ == '__main__':
spider = TiebaImageSpider()
spider.main()
案例2:糗事百科-xpath
- 目标 :用户昵称、段子内容、好笑数、评论数
- 步骤
- 找URL
https://www.qiushibaike.com/8hr/page/1/
- xpath匹配
- 基准xpath:
//li[contains(@id,"qiushi")]
- 标题内容:``
- 用户昵称:
.//span[@class="recmd-name"]
- 好笑数量:
./div/div/div/span[1]
- 评论数量:
./div/div/div/span[4]
- 基准xpath:
- 找URL
import requests
from lxml import etree
import pymongo
class QiuShiSpider:
def __init__(self):
self.url = "https://www.qiushibaike.com/"
self.headers = {"User-Agent": "Mozilla/5.0"}
self.conn = pymongo.MongoClient("localhost", 27017)
self.db = self.conn.spiderdb
self.set = self.db.qiushi
def get_page(self):
res = requests.get(self.url, headers=self.headers)
res.encoding = 'utf-8'
html = res.text
self.parse_page(html)
def parse_page(self, html):
parse_html = etree.HTML(html)
# 匹配每个段子列表
base_list = parse_html.xpath('//li[contains(@id,"qiushi")]')
# 遍历每个段子的借点对象
for base in base_list:
# 标题
title = base.xpath('./div/a')[0].text
# 用户昵称
username = base.xpath('.//span[@class="recmd-name"]')[0].text
# 好笑数量
try:
laugh_num = base.xpath('./div/div/div/span[1]')[0].text
# 或:laugh_num = base.xpath('./div/div/div/span')[0].text
except IndexError:
laugh_num = 0
# 评论数量
try:
com_num = base.xpath('./div/div/div/span[4]')[0].text
# com_num = base.xpath('./div/div/div/span')[3].text
except IndexError:
com_num = 0
d = {
"title": title,
"username": username,
"laugh_num": laugh_num,
"com_num": com_num
}
self.set.insert(d)
print("存入数据库成功!")
if __name__ == "__main__":
spider = QiuShiSpider()
spider.get_page()
posted @
2019-05-12 17:33 黑洞频率 阅读(
...) 评论(
...) 编辑 收藏