bp4解析库
pip3 install beautifulsoup4 # 安装bs4
pip3 install lxml # 下载lxml解析器
html_doc = """The Dormouse's story $37
Once upon a time there were three little sisters; and their names were Elsie, Lacie and Tillie; and they lived at the bottom of a well.
...
""" # 从bs4中导入BeautifulSoup from bs4 import BeautifulSoup # 调用BeautifulSoup实例化得到一个soup对象 # 参数一: 解析文本 # 参数二: # 参数二: 解析器(html.parser、lxml...) soup = BeautifulSoup(html_doc, 'lxml') print(soup) print('*' * 100) print(type(soup)) print('*' * 100) # 文档美化 html = soup.prettify() print(html)
遍历文档树:
1、直接使用
2、获取标签的名称
3、获取标签的属性
4、获取标签的内容
5、嵌套选择
6、子节点、子孙节点
7、父节点、祖先节点
8、兄弟节点
# 1、直接使用 print(soup.p) # 查找第一个p标签 print(soup.a) # 查找第一个a标签 # 2、获取标签的名称 print(soup.head.name) # 获取head标签的名称 # 3、获取标签的属性 print(soup.a.attrs) # 获取a标签中的所有属性 print(soup.a.attrs['href']) # 获取a标签中的href属性 # 4、获取标签的内容 print(soup.p.text) # $37 # 5、嵌套选择 print(soup.html.head) # 6、子节点、子孙节点 print(soup.body.children) # body所有子节点,返回的是迭代器对象 print(list(soup.body.children)) # 强转成列表类型 print(soup.body.descendants) # 子孙节点 print(list(soup.body.descendants)) # 子孙节点 # 7、父节点、祖先节点 print(soup.p.parent) # 获取p标签的父亲节点 # 返回的是生成器对象 print(soup.p.parents) # 获取p标签所有的祖先节点 print(list(soup.p.parents)) # 8、兄弟节点 # 找下一个兄弟 print(soup.p.next_sibling) # 找下面所有的兄弟,返回的是生成器 print(soup.p.next_siblings) print(list(soup.p.next_siblings)) # 找上一个兄弟 print(soup.a.previous_sibling) # 找到第一个a标签的上一个兄弟节点 # 找到a标签上面的所有兄弟节点 print(soup.a.previous_siblings) # 返回的是生成器 print(list(soup.a.previous_siblings))
搜索文档树:
find() 找一个
find_all() 找多个
标签查找与属性查找:
标签:
name 属性匹配
attrs 属性查找匹配
text 文本匹配
- 字符串过滤器
字符串全局匹配
- 正则过滤器
re模块匹配
- 列表过滤器
列表内的数据匹配
- bool过滤器
True匹配
- 方法过滤器
用于一些要的属性以及不需要的属性查找。
属性:
- class_
- id
from bs4 import BeautifulSoup soup = BeautifulSoup(html_doc, 'lxml') # # 字符串过滤器 # # name # p_tag = soup.find(name='p') # print(p_tag) # 根据文本p查找某个标签 # # 找到所有标签名为p的节点 # tag_s1 = soup.find_all(name='p') # print(tag_s1) # # # # attrs # # 查找第一个class为sister的节点 # p = soup.find(attrs={"class": "sister"}) # print(p) # # 查找所有class为sister的节点 # tag_s2 = soup.find_all(attrs={"class": "sister"}) # print(tag_s2) # # # # text # text = soup.find(text="$37") # print(text) # # # # 配合使用: # # 找到一个id为link2、文本为Lacie的a标签 # a_tag = soup.find(name="a", attrs={"id": "link2"}, text="Lacie") # print(a_tag) # # 正则过滤器 # import re # # name # p_tag = soup.find(name=re.compile('p')) # print(p_tag) # 列表过滤器 # import re # # name # tags = soup.find_all(name=['p', 'a', re.compile('html')]) # print(tags) # - bool过滤器 # True匹配 # 找到有id的p标签 # p = soup.find(name='p', attrs={"id": True}) # print(p) # 方法过滤器 # 匹配标签名为a、属性有id没有class的标签 # def have_id_class(tag): # if tag.name == 'a' and tag.has_attr('id') and tag.has_attr('class'): # return tag # # tag = soup.find(name=have_id_class) # print(tag)
爬取豌豆荚游戏主页
主页:
图标地址、下载次数、大小、详情页地址
详情页:
游戏名、图标名、好评率、评论数、小编点评、简介、网友评论、1-5张截图链接地址、下载地址
https://www.wandoujia.com/wdjweb/api/category/morecatId=6001&subCatId=0&page=1&ctoken=FRsWKgWBqMBZLdxLaK4iem9B
https://www.wandoujia.com/wdjweb/api/category/morecatId=6001&subCatId=0&page=2&ctoken=FRsWKgWBqMBZLdxLaK4iem9B
https://www.wandoujia.com/wdjweb/api/category/morecatId=6001&subCatId=0&page=3&ctoken=FRsWKgWBqMBZLdxLaK4iem9B
import requests from bs4 import BeautifulSoup # 1、发送请求 def get_page(url): response = requests.get(url) return response # 2、开始解析 # 解析主页 def parse_index(data): soup = BeautifulSoup(data, 'lxml') # 获取所有app的li标签 app_list = soup.find_all(name='li', attrs={"class": "card"}) for app in app_list: # print('tank *' * 1000) # print(app) # 图标地址 img = app.find(name='img').attrs['data-original'] print(img) # 下载次数 down_num = app.find(name='span', attrs={"class": "install-count"}).text print(down_num) import re # 大小 size = soup.find(name='span', text=re.compile("\d+MB")).text print(size) # 详情页地址 detail_url = soup.find(name='a', attrs={"class": "detail-check-btn"}).attrs['href'] print(detail_url) def main(): for line in range(1, 33): url = f"https://www.wandoujia.com/wdjweb/api/category/more?catId=6001&subCatId=0&page={line}&ctoken=FRsWKgWBqMBZLdxLaK4iem9B" # 1、往app接口发送请求 response = get_page(url) # print(response.text) print('*' * 1000) # 反序列化为字典 data = response.json() # 获取接口中app标签数据 app_li = data['data']['content'] # print(app_li) # 2、解析app标签数据 parse_index(app_li) if __name__ == '__main__': main()
爬取豌豆荚数据
import requests from bs4 import BeautifulSoup # 1、发送请求 def get_page(url): response = requests.get(url) return response # 2、开始解析 # 解析详情页 def parse_detail(text): soup = BeautifulSoup(text, 'lxml') # print(soup) # app名称 name = soup.find(name="span", attrs={"class": "title"}).text # print(name) # 好评率 love = soup.find(name='span', attrs={"class": "love"}).text # print(love) # 评论数 commit_num = soup.find(name='a', attrs={"class": "comment-open"}).text # print(commit_num) # 小编点评 commit_content = soup.find(name='div', attrs={"class": "con"}).text # print(commit_content) # app下载链接 download_url = soup.find(name='a', attrs={"class": "normal-dl-btn"}).attrs['href'] # print(download_url) print( f''' ============= tank ============== app名称:{name} 好评率: {love} 评论数: {commit_num} 小编点评: {commit_content} app下载链接: {download_url} ============= end ============== ''' ) # 解析主页 def parse_index(data): soup = BeautifulSoup(data, 'lxml') # 获取所有app的li标签 app_list = soup.find_all(name='li', attrs={"class": "card"}) for app in app_list: # print(app) # print('tank' * 1000) # print('tank *' * 1000) # print(app) # 图标地址 # 获取第一个img标签中的data-original属性 img = app.find(name='img').attrs['data-original'] print(img) # 下载次数 # 获取class为install-count的span标签中的文本 down_num = app.find(name='span', attrs={"class": "install-count"}).text print(down_num) import re # 大小 # 根据文本正则获取到文本中包含 数字 + MB(\d+代表数字)的span标签中的文本 size = soup.find(name='span', text=re.compile("\d+MB")).text print(size) # 详情页地址 # 获取class为detail-check-btn的a标签中的href属性 # detail_url = soup.find(name='a', attrs={"class": "name"}).attrs['href'] # print(detail_url) # 详情页地址 detail_url = app.find(name='a').attrs['href'] print(detail_url) # 3、往app详情页发送请求 response = get_page(detail_url) # 4、解析app详情页 parse_detail(response.text) def main(): for line in range(1, 33): url = f"https://www.wandoujia.com/wdjweb/api/category/more?catId=6001&subCatId=0&page={line}&ctoken=FRsWKgWBqMBZLdxLaK4iem9B" # 1、往app接口发送请求 response = get_page(url) # print(response.text) print('*' * 1000) # 反序列化为字典 data = response.json() # 获取接口中app标签数据 app_li = data['data']['content'] # print(app_li) # 2、解析app标签数据 parse_index(app_li) if __name__ == '__main__': main()
pymongo使用
from pymongo import MongoClient # 1、链接mongoDB客户端 # 参数1: mongoDB的ip地址 # 参数2: mongoDB的端口号 默认:27017 client = MongoClient('localhost', 27017) # print(client) # 2、进入tank_db库,没有则创建 print(client['tank_db']) # 3、创建集合 print(client['tank_db']['people']) # 4、给tank_db库插入数据 # 1.插入一条 data1 = { 'name': 'tank', 'age': 18, 'sex': 'male' } client['tank_db']['people'].insert(data1) # 2.插入多条 # data1 = { # 'name': 'tank', # 'age': 18, # 'sex': 'male' # } # data2 = { # 'name': '戚志云', # 'age': 84, # 'sex': 'female' # } # data3 = { # 'name': '沈金金', # 'age': 73, # 'sex': 'male' # } # client['tank_db']['people'].insert([data1, data2, data3]) # # # 5、查数据 # # 查看所有数据 data_s = client['tank_db']['people'].find() # print(data_s) ## # 需要循环打印所有数据 # for data in data_s: # print(data) # # # 查看一条数据 # data = client['tank_db']['people'].find_one() # print(data) # 官方推荐使用 # 插入一条insert_one # client['tank_db']['people'].insert_one() # 插入多条insert_many # client['tank_db']['people'].insert_many()