《python网络数据采集》笔记1

第一部分-创建爬虫

1.urllib

       1)urllib.request

       request.urlopen(url)

       request.urlretrieve 可以根据文件的 URL 下载文件

       2)urllib.parse

       3)urllib.error

      

             

2.异常：

       try...except...else...

       常见异常：HTTPError,AttributeError,

      

3.BeautifulSoup

       from bs4 import BeautifulSoup

       bsObj=BeautifulSoup(html,'lxml')

       1)

       print(bsObj.text)

       print(bsObj.html)

       print(bsObj.p.a)

       ...

       2)

       findAll(tag, attributes, recursive, text, limit, keywords)     #返回一个ResultSet

       find(tag, attributes, recursive, text, keywords)   #返回一个Tag

       借助它们,你可以通过标签的不同属性轻松地过滤 HTML 页面，查找需要的标签组或单个标签   

       例：

       .findAll({"h1","h2","h3","h4","h5","h6"})

       .findAll("span", {"class":{"green", "red"}})

       .findAll(id="text")   同 .findAll("", {"id":"text"})

       .findAll(src=True) 有src属性的标签

       3)

       get_text() 会把你正在处理的 HTML 文档中所有的标签都清除，然后返回一个只包含文字的str

       4)返回类型NavigatorString

       .children （所有子标签）

       .next_sibling（      下一个兄弟标签 ）.next_siblings（所有之后的兄弟标签）

       .previous_sibling（上一个兄弟标签）.previous（所有之前的兄弟标签）

       .parent （直接父标签 ）.parents（所有父标签）、

       5）

       .attrs      获取标签所有属性(dict)

       .attrs['src']      获取src值    

       6)正则表达式

       7)lambda表达式

       #获取有两个属性的标签：

       bsObj.findAll(lambda tag: len(tag.attrs) == 2)

      

4.Scrapy

       //TODO

5.JSON

       把 JSON 转换成字典，

       JSON 数组转换成列表，

       JSON 字符串转换成 Python 字符串。

       常用函数：loads,get

6.存储数据

       1)下载

              from urllib.request import urlretrieve

              urlretrieve(resourceLocation,fileName)

       2）CSV（Comma-Separated Values）

              import csv

              csvFile=open("test.csv","w+")

              try:

                     writer=csv.writer(csvFile)

                     writer.writerow(('青山隐隐水迢迢 秋尽江南草未凋','24桥明月夜'))

                     for i in range(1,5):

                            writer.writerow((i,i+2,i*2))

              finally:

                     csvFile.close()

       3）MySQL

              import pymysql

              #获取连接 获取光标

              conn=pymysql.connect(host='localhost',user='root',passwd=None)

              cur=conn.cursor()

              #执行SQL语句

              cur.execute('use ssm01')

              cur.execute('select * from user')

              print(cur.fetchone())#获取一条数据

              #关闭资源

              cur.close()

              coon.close()

       4）Email

       //TODO

7.读取文档

       1)读取txt

       from urllib.request import urlopen    

       txt=urlopen('http://www.pythonscraping.com/pages/warandpeace/chapter1.txt')

       print(txt.read())

      

       2)读取csv

       #从网上直接把文件读成一个字符串，然后转换成一个 StringIO 对象，使它具有文件的属性。

       from urllib.request import urlopen

       from io import StringIO

       import csv

       data = urlopen('http://pythonscraping.com/files/MontyPythonAlbums.csv').read().decode('utf-8')

       dataFile=StringIO(data)

       csvFile=csv.reader(dataFile)

       for row in csvFile:

              print(row)

      

       3）读取PDF 

       #PDFMiner3K

       #把任意 PDF 读成字符串，然后用 StringIO 转换成文件对象

       from urllib.request import urlopen

       from pdfminer.pdfinterp import PDFResourceManager, process_pdf

       from pdfminer.converter import TextConverter

       from pdfminer.layout import LAParams

       from io import StringIO

       def readPDF(pdfFile):

              rsrcmgr = PDFResourceManager()

              retstr = StringIO()

              laparams = LAParams()

              device = TextConverter(rsrcmgr, retstr, laparams=laparams)

              process_pdf(rsrcmgr, device, pdfFile)

              device.close()

              content = retstr.getvalue()

              retstr.close()

              return content

       pdfFile = urlopen("http://pythonscraping.com/pages/warandpeace/chapter1.pdf")

       outputString = readPDF(pdfFile)

       print(outputString)

       pdfFile.close()

////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// 

例

3-1.网络数据采集

       #从 http://oreilly.com 开始，然后随机地从一个外链跳到另一个外链。

       from urllib.request import urlopen

       from bs4 import BeautifulSoup

       import re

       import datetime

       import random

       pages = set()

       random.seed(datetime.datetime.now())

       # 获取页面所有内链的列表

       def getInternalLinks(bsObj, includeUrl):

              internalLinks = []

              # 找出所有以"/"开头的链接

              for link in bsObj.findAll("a", href=re.compile("^(/|.*"+includeUrl+")")):

                     if link.attrs['href'] is not None:

                            if link.attrs['href'] not in internalLinks:

                                   internalLinks.append(link.attrs['href'])

              return internalLinks

       # 获取页面所有外链的列表

       def getExternalLinks(bsObj, excludeUrl):

              externalLinks = []

              # 找出所有以"http"或"www"开头且不包含当前URL的链接

              for link in bsObj.findAll("a",href=re.compile("^(http|www)((?!"+excludeUrl+").)*$")):

                     if link.attrs['href'] is not None:

                            if link.attrs['href'] not in externalLinks:

                                   externalLinks.append(link.attrs['href'])

              return externalLinks

       def splitAddress(address):

              addressParts = address.replace("http://", "").split("/")

              return addressParts

       def getRandomExternalLink(startingPage):

              html = urlopen(startingPage)

              bsObj = BeautifulSoup(html,'lxml')

              externalLinks = getExternalLinks(bsObj, splitAddress(startingPage)[0])

              if len(externalLinks) == 0:

                     internalLinks = getInternalLinks(startingPage)

                     return getNextExternalLink(internalLinks[random.randint(0,len(internalLinks)-1)])

              else:

                     return externalLinks[random.randint(0, len(externalLinks)-1)]

       def followExternalOnly(startingSite):

              externalLink = getRandomExternalLink("http://oreilly.com")

              print("随机外链是："+externalLink)

              followExternalOnly(externalLink)

       followExternalOnly("http://oreilly.com")

5-1.JSON

       import json

       jsonString='{

              "arrayOfNums":[{"number":0},{"number":1},{"number":2}],

              "arrayOfFruits":[{"fruit":"apple"},{"fruit":"banana"},{"fruit":"pear"}]

              }'

       jsonObj=json.loads(jsonString)

       print(jsonObj.get("arrayOfFruits")[2].get("fruit"))

6-1.把 http://pythonscraping.com 的所有图片下载下来

       from urllib.request import urlretrieve

       from urllib.request import urlopen

       from bs4 import BeautifulSoup

       def pageSrc(url):

              html=urlopen(url)

              bsObj=BeautifulSoup(html,'lxml')

              srcList=bsObj.findAll("img",src=True)

              urlList=[]

              for i in srcList:

                     urlList.append(i['src'])

              return urlList

       def getInternalLinks(bsObj,includeUrl):

                     internalLinks = []

                     # 找出所有以"/"开头的链接

                     for link in bsObj.findAll("a", href=re.compile("^(/|.*"+includeUrl+")")):

                            if link.attrs['href'] is not None:

                                   if link.attrs['href'] not in internalLinks:

                                          internalLinks.append(link.attrs['href'])

                     return internalLinks

       def allimgs(url):

              #找到该页面所有的img src

              srcset=set()

              for i in pageSrc(url):

                     if i not in srcset:

                            print(i)

                            srcset.add(i)

                            name=i.split('/').pop()

                            urlretrieve(i,name)

              #找到该页面的所有内链

              html=urlopen(url)

              bsObj=BeautifulSoup(html,'lxml')

              for i in getInternalLinks(bsObj,url):

                     newUrl=url+i

                     for j in pageSrc(newUrl):

                            if j not in srcset:

                                   srcset.add(i)

                                   print(j)

                                   name=j.split('/').pop()

                                   urlretrieve(j,name)

       url="http://pythonscraping.com"

       allimgs(url)

6-2.存储到CSV

       #获取 HTML 表格并写入 CSV 文件

       import csv

       from urllib.request import urlopen

       from bs4 import BeautifulSoup

       html = urlopen("http://en.wikipedia.org/wiki/Comparison_of_text_editors")

       bsObj = BeautifulSoup(html,'lxml')

       # 主对比表格是当前页面上的第一个表格

       table = bsObj.findAll("table",{"class":"wikitable"})[0]

       rows = table.findAll("tr")

       csvFile = open("editors.csv", 'wt', newline='',encoding='utf-8')

       writer = csv.writer(csvFile)

       try:

              for row in rows:

                     csvRow = []

                     for cell in row.findAll(['td', 'th']):

                            csvRow.append(cell.get_text()[:-1])

                     print(csvRow)

                     writer.writerow(csvRow)

       finally:

              csvFile.close()

6-3.存储到mysql

       #存储维基百科数据

       from urllib.request import urlopen

       from bs4 import BeautifulSoup

       import re

       import datetime

       import random

       import pymysql

       conn = pymysql.connect(host='127.0.0.1',user='root', passwd=None, charset='utf8')

       cur = conn.cursor()

       cur.execute("USE ssm01")

       cur.execute("CREATE TABLE pages(title varchar(200),content varchar(3000))")

       random.seed(datetime.datetime.now())

       #存储到数据库

       def store(title, content):

              cur.execute("INSERT INTO pages (title, content) VALUES ("%s","%s")", (title, content))

              cur.connection.commit()

       #找到数据 存储到数据库

       def getLinks(articleUrl):

              html = urlopen("http://en.wikipedia.org"+articleUrl)

              bsObj = BeautifulSoup(html,'lxml')

              title = bsObj.find("h1").get_text()

              content = bsObj.find("div", {"id":"mw-content-text"}).find("p").get_text()

              store(title, content)

              return bsObj.find("div", {"id":"bodyContent"}).findAll("a",href=re.compile("^(/wiki/)((?!:).)*$"))

       links = getLinks("/wiki/Kevin_Bacon")

       try:

              while len(links) > 0:

                     newArticle = links[random.randint(0, len(links)-1)].attrs["href"]

                     print(newArticle)

                     links = getLinks(newArticle)

       finally:

              cur.close()

              conn.close()