Pyspider爬虫教程

Pyspider爬虫教程

一、安装

1、 安装pip

（1）准备工作

yum install –y make gcc-c++ python-devel libxml2-devel libxslt-devel

（2）安装setuptools

https://pypi.python.org/pypi/setuptools/

python setup.py install

（3）安装pip

https://pypi.python.org/pypi/pip

python setup.py install

2、 安装pyspider

（1）安装pyspider及其依赖

pip install pyspider

OR

pip install --allow-all-external pyspider[all]

3、 安装可选库

pip install rsa

4、 phantomjs

下载后复制至/bin/

二、部署pyspider服务器

（1）配置pyspider.conf

配置eagles引擎和生成结果数据库，配置用户名。密码等

（2）运行install.sh安装脚本

（3）/etc/init.d/pyspider start 启动服务即可

三、使用

（1）ip：5000直接打开网页客户端

（2）点击创建脚本

（3）编写脚本，直接调试

（4）选择“running“点击运行

四、爬虫教程（1）—抓取简单的静态页面

静态页面的抓取最简单，获取HTML页面进行标签抽取即可，例子如下：

贵阳晚报新闻文章抓取：http://www.gywb.com.cn/

# Handler类和入口函数
class Handler(BaseHandler):
    crawl_config = {
    }
    @every(minutes=24 * 60)
    def on_start(self):
        for url in urlList:
            self.crawl(url, callback=self.index_page)

self.crawl抓取贵阳晚报首页的url，跳转到回调函数index_page

# 回调函数index_page
# config age:10天不刷新
# response.url:抓取的url
# response.doc(pyquery):获取标签内容，参数：pyquery标签
# 通过又一次抓取url到detail_page
@config(age=10 * 24 * 60 * 60)
    def index_page(self, response):
        for each in response.doc('a[href^="http"]').items():
            for url in urlList:
                if ('%s%s' % (url, 'content/') in each.attr.href) and \ # 字符串连接
                   (not ("#" in each.attr.href)): # 判断#是否在href里面
                        self.crawl(each.attr.href, callback=self.detail_page)

详细页面detail_page：获取文章标题，文章内容，时间等

# config priority:调用优先级
  @config(priority=2)
    def detail_page(self, response):
    # article title
        artTitle = response.doc(artTitleSelector1).text().strip() # 获取文章标题, artTitleSelector1:pyquery标签，如：h1[class="g-content-t text-center"]

完整代码如下：

#!/usr/bin/env python
# -*- encoding: utf-8 -*-
# Created on 2015-05-12 10:41:03
# Project: GYWB
 
 
from pyspider.libs.base_handler import *
import re
 
urlList = [ "http://www.gywb.com.cn/" ]
keyWords = [ #u"贵阳",
             #u"交通",
             u"违章",
             u"交警",
             u"交通管理",
             #u"交通管理局",
             u"交管局" ]
 
# article title
artTitleSelector1 = 'h1[class="g-content-t text-center"]'
artTitleSelector2 = 'div[class="detail_title_yy"] h1'
 
# article content
artContentSelector1 = 'div[class="g-content-c"] p'
artContentSelector2 = 'div[class="detailcon"] p'
 
# publish time
artPubTimeSelector1 = '#pubtime_baidu'
artPubTimeFilter1 = r'[^\d]*'
artPubTimeSelector2 = '.detail_more'
artPubTimeFilter2 = r'[\d\-\:\ ]*'
 
class Handler(BaseHandler):
    crawl_config = {
    }
 
    @every(minutes=24 * 60)
    def on_start(self):
        for url in urlList:
            self.crawl(url, callback=self.index_page)
 
    @config(age=10 * 24 * 60 * 60)
    def index_page(self, response):
        for each in response.doc('a[href^="http"]').items():
            for url in urlList:
                if ('%s%s' % (url, 'content/') in each.attr.href) and \
                   (not ("#" in each.attr.href)):
                        self.crawl(each.attr.href, callback=self.detail_page)
 
    @config(priority=2)
    def detail_page(self, response):
        for each in response.doc('a[href^="http"]').items():
            self.crawl(each.attr.href, callback=self.index_page)
 
        # article title
        artTitle = response.doc(artTitleSelector1).text().strip()
        if artTitle == '':
            artTitle = response.doc(artTitleSelector2).text().strip()
        if artTitle == '':
            return None
 
        artContent = response.doc(artContentSelector1).text().strip()
        if artContent == '':
            artContent = response.doc(artContentSelector2).text().strip()
 
        artPubTime = response.doc(artPubTimeSelector1).text().strip()
        if artPubTime != '':
            match = re.match (artPubTimeFilter1, artPubTime)
            if match != None:
                artPubTime = artPubTime[len(match.group()):]
        else:
            artPubTime = response.doc(artPubTimeSelector2).text().strip()
            match = re.match (artPubTimeFilter1, artPubTime)
            if match != None:
                artPubTime = artPubTime[len(match.group()):]
            match = re.search (artPubTimeFilter2, artPubTime)
            if match != None:
                artPubTime = match.group()
 
        artPubTime = artPubTime.strip()
 
        for word in keyWords:
            if word in artContent:
                return {
                    #"url": response.url,
                    #"title": response.doc('title').text(),
                    "title" : artTitle,
                    "time" : artPubTime,
                    "content" : artContent,
                }
        else:
            return None

五、爬虫教程（2）—HTTP请求的页面（如登陆后抓取）

例子：爱卡汽车论坛：http://a.xcar.com.cn/bbs/forum-d-303.html

首先登陆就需要用户名和密码，但是好的网站都需要对用户名和密码进行加密的。所以我们只能模拟登陆方式，获取用户名和密码的加密类型，从而来进行模拟登陆，模拟登陆就需要获取浏览器的cookie

class Handler(BaseHandler):
    crawl_config = {
    }
 
    @every(minutes=24 * 60)
    def on_start(self):
        cookies = getCookies() # 获取cookie
 
        for url in URL_LIST:
            self.crawl(url, cookies = cookies, callback=self.index_page) # 传入cookie模拟登陆

那么怎么样才能获取cookie呢？

（1）获取post提交的data，可以使用Firefox的httpfox插件或者wireshark来对包进行抓取

下面是采用Firefox的httpfox插件进行抓取

从上图可以看出，post_data包含username和password，还有chash和dhash等，当然这些都是进行加密后的。

所以在程序中需要获取post_data

def getPostData(self):
        url = self.login_url.strip() # 登陆的url
        if not re.match(r'^http://', url):
            return None, None
 
        req = urllib2.Request(url)
        resp = urllib2.urlopen(req)
        login_page = resp.read() 
 
        # 获取html表单数据
        doc = HTML.fromstring (login_page)
        post_url = doc.xpath("//form[@name='login' and @id='login']/@action")[0][1:]
        chash = doc.xpath("//input[@name='chash' and @id='chash']/@value")[0]
        dhash = doc.xpath("//input[@name='dhash' and @id='dhash']/@value")[0]
        ehash = doc.xpath("//input[@name='ehash' and @id='ehash']/@value")[0]
        formhash = doc.xpath("//input[@name='formhash']/@value")[0]
        loginsubmit = doc.xpath("//input[@name='loginsubmit']/@value")[0].encode('utf-8')
        cookietime = doc.xpath("//input[@name='cookietime' and @id='cookietime']/@value")[0]
 
        username = self.account # 账户
        password = self.encoded_password # 密码
 
        #组合post_data
        post_data = urllib.urlencode({
            'username' : username,
            'password' : password,
            'chash' : chash,
            'dhash' : dhash,
            'ehash' : ehash,
            'loginsubmit' : loginsubmit,
            'formhash' : formhash,
            'cookietime' : cookietime,
        })
 
        return post_url, post_data

 #将post_data作为参数模拟登陆
 def login(self):
        post_url, post_data = self.getPostData()
        post_url = self.post_url_prefix + post_url
 
        req = urllib2.Request(url = post_url, data = post_data)
        resp = urllib2.urlopen(req)
        return True

# 通过本地浏览器cookie文件获取cookie
# 账号进行md5加密
COOKIES_FILE = '/tmp/pyspider.xcar.%s.cookies' % hashlib.md5(ACCOUNT).hexdigest()
COOKIES_DOMAIN = 'xcar.com.cn'
def getCookies():
    CookiesJar = cookielib.MozillaCookieJar(COOKIES_FILE)
    if not os.path.isfile(COOKIES_FILE):
        CookiesJar.save()
 
    CookiesJar.load (COOKIES_FILE)
    CookieProcessor = urllib2.HTTPCookieProcessor(CookiesJar)
    CookieOpener = urllib2.build_opener(CookieProcessor, urllib2.HTTPHandler)
    for item in HTTP_HEADERS:
        CookieOpener.addheaders.append ((item ,HTTP_HEADERS[item]))
    urllib2.install_opener(CookieOpener)
 
    if len(CookiesJar) == 0:
        xc = xcar(ACCOUNT, ENCODED_PASSWORD, LOGIN_URL, POST_URL_PREFIX)
        if xc.login(): # 判断登陆成功，保存cookie
            CookiesJar.save()
        else:
            return None
 
    CookiesDict = {}
    # 选择对本次登陆的cookie
    for cookie in CookiesJar:
        if COOKIES_DOMAIN in cookie.domain:
            CookiesDict[cookie.name] = cookie.value
    return CookiesDict

怎样查看用户名和密码的加密类型？——通过查看js文件

查看登陆login.js和表单信息login.php

发现：username是采用base64加密，password是先采用md5加密，然后再进行base64加密

完整代码如下：

#!/usr/bin/env python
# -*- encoding: utf-8 -*-
# Created on 2015-05-14 17:39:36
# Project: test_xcar
 
 
from pyspider.libs.base_handler import *
from pyspider.libs.response import *
from pyquery import PyQuery
 
import os
import re
import urllib
import urllib2
import cookielib
import lxml.html as HTML
import hashlib
 
 
URL_LIST= [ 'http://a.xcar.com.cn/bbs/forum-d-303.html' ]
 
THREAD_LIST_URL_FILTER = 'bbs/forum-d-303'
THREAD_LIST_URL_REG = r'bbs\/forum-d-303(-\w+)?\.'
 
ACCOUNT = 'ZhangZujian'
# 32-bit MD5 Hash
ENCODED_PASSWORD = 'e3d541408adb57f4b40992202c5018d8'
 
LOGIN_URL = 'http://my.xcar.com.cn/logging.php?action=login'
POST_URL_PREFIX = 'http://my.xcar.com.cn/'
 
THREAD_URL_REG = r'bbs\/thread-\w+-0'
THREAD_URL_HREF_FILTER = 'bbs/thread-'
THREAD_URL_CLASS_LIST = [ 'prev', 'next' ]
 
THREAD_THEME_SELECTOR = 'h2'
POST_ITEM_SELECTOR = '.posts-con > div'
POST_TIME_SELECTOR = '.pt-time > span'
POST_MEMBER_SELECTOR = '.pt-name'
POST_FLOOR_SELECTOR = '.pt-floor > span'
POST_CONTENT_SELECTOR = '.pt-cons'
# THREAD_REPLY_SELECTOR = ''
 
# !!! Notice !!!
# Tasks that share the same account MUST share the same cookies file
COOKIES_FILE = '/tmp/pyspider.xcar.%s.cookies' % hashlib.md5(ACCOUNT).hexdigest()
COOKIES_DOMAIN = 'xcar.com.cn'
# USERAGENT_STR = 'Mozilla/5.0 (iPhone; CPU iPhone OS 8_0 like Mac OS X) AppleWebKit/600.1.4 (KHTML, like Gecko) Version/8.0 Mobile/12A366 Safari/600.1.4'
 
HTTP_HEADERS = {
    'Accept' : 'text/html,application/xhtml+xml,application/xml;q=0.9,image/webp,*/*;q=0.8',
    # 'Accept-Encoding' : 'gzip, deflate, sdch',
    'Accept-Language' : 'zh-CN,zh;q=0.8,en;q=0.6',
    'Connection' : 'keep-alive',
    'DNT' : '1',
    'Host' : 'my.xcar.com.cn',
    'Referer' : 'http://a.xcar.com.cn/',
    'User-Agent' : 'Mozilla/5.0 (Windows NT 6.1; WOW64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/42.0.2311.135 Safari/537.36',
}
 
 
class xcar(object):
    def __init__(self, account, encoded_password, login_url, post_url_prefix):
        self.account = account
        self.encoded_password = encoded_password
        self.login_url = login_url
        self.post_url_prefix = post_url_prefix
 
 
    def login(self):
        post_url, post_data = self.getPostData()
        post_url = self.post_url_prefix + post_url
 
        req = urllib2.Request(url = post_url, data = post_data)
        resp = urllib2.urlopen(req)
        return True
 
 
    def getPostData(self):
        url = self.login_url.strip()
        if not re.match(r'^http://', url):
            return None, None
 
        req = urllib2.Request(url)
        resp = urllib2.urlopen(req)
        login_page = resp.read()
 
        doc = HTML.fromstring (login_page)
        post_url = doc.xpath("//form[@name='login' and @id='login']/@action")[0][1:]
        chash = doc.xpath("//input[@name='chash' and @id='chash']/@value")[0]
        dhash = doc.xpath("//input[@name='dhash' and @id='dhash']/@value")[0]
        ehash = doc.xpath("//input[@name='ehash' and @id='ehash']/@value")[0]
        formhash = doc.xpath("//input[@name='formhash']/@value")[0]
        loginsubmit = doc.xpath("//input[@name='loginsubmit']/@value")[0].encode('utf-8')
        cookietime = doc.xpath("//input[@name='cookietime' and @id='cookietime']/@value")[0]
 
        username = self.account
        password = self.encoded_password
 
        post_data = urllib.urlencode({
            'username' : username,
            'password' : password,
            'chash' : chash,
            'dhash' : dhash,
            'ehash' : ehash,
            'loginsubmit' : loginsubmit,
            'formhash' : formhash,
            'cookietime' : cookietime,
        })
 
        return post_url, post_data
 
 
def getCookies():
    CookiesJar = cookielib.MozillaCookieJar(COOKIES_FILE)
    if not os.path.isfile(COOKIES_FILE):
        CookiesJar.save()
 
    CookiesJar.load (COOKIES_FILE)
    CookieProcessor = urllib2.HTTPCookieProcessor(CookiesJar)
    CookieOpener = urllib2.build_opener(CookieProcessor, urllib2.HTTPHandler)
    for item in HTTP_HEADERS:
        CookieOpener.addheaders.append ((item ,HTTP_HEADERS[item]))
    urllib2.install_opener(CookieOpener)
 
    if len(CookiesJar) == 0:
        xc = xcar(ACCOUNT, ENCODED_PASSWORD, LOGIN_URL, POST_URL_PREFIX)
        if xc.login():
            CookiesJar.save()
        else:
            return None
 
    CookiesDict = {}
    for cookie in CookiesJar:
        if COOKIES_DOMAIN in cookie.domain:
            CookiesDict[cookie.name] = cookie.value
    return CookiesDict
 
 
class Handler(BaseHandler):
    crawl_config = {
    }
 
    @every(minutes=24 * 60)
    def on_start(self):
        cookies = getCookies()
 
        for url in URL_LIST:
            self.crawl(url, cookies = cookies, callback=self.index_page)
 
    @config(age=10 * 24 * 60 * 60)
    def index_page(self, response):
        cookies = getCookies()
 
        for each in response.doc('a[href*="%s"]' % THREAD_URL_HREF_FILTER).items():
            if re.search(THREAD_URL_REG, each.attr.href) and \
                '#' not in each.attr.href:
                self.crawl(each.attr.href, cookies = cookies, callback=self.detail_page)
 
        for each in response.doc('a[href*="%s"]' % THREAD_LIST_URL_FILTER).items():
            if re.search(THREAD_LIST_URL_REG, each.attr.href) and \
                '#' not in each.attr.href:
                self.crawl(each.attr.href, cookies = cookies, callback=self.index_page)
 
    @config(priority=2)
    def detail_page(self, response):
        cookies = getCookies()
 
        if '#' not in response.url:
            for each in response.doc(POST_ITEM_SELECTOR).items():
                floorNo = each(POST_FLOOR_SELECTOR).text()
                url = '%s#%s' % (response.url, floorNo)
                self.crawl(url, cookies = cookies, callback=self.detail_page)
            return None
        else:
            floorNo = response.url[response.url.find('#')+1:]
 
            for each in response.doc(POST_ITEM_SELECTOR).items():
                if each(POST_FLOOR_SELECTOR).text() == floorNo:
                    theme = response.doc(THREAD_THEME_SELECTOR).text()
                    time = each(POST_TIME_SELECTOR).text()
                    member = each(POST_MEMBER_SELECTOR).text()
                    content = each(POST_CONTENT_SELECTOR).text()
                    return {
                        # "url" : response.url,
                        # "title" : response.doc('title').text(),
                        'theme' : theme,
                        'floor' : floorNo,
                        'time' : time,
                        'member' : member,
                        'content' : content,
                    }

六、爬虫教程（3）使用Phantomjs渲染带js的页面（视频链接抓取）

（1）安装Phantomjs

（2）其实和爬取网页的不同点在于，传入fetch_tpy = 'js'

#!/usr/bin/env python
# -*- encoding: utf-8 -*-
# Created on 2015-03-20 09:46:20
# Project: fly_spider
import re
import time
#from pyspider.database.mysql.mysqldb import SQL
from pyspider.libs.base_handler import *
from pyquery import PyQuery as pq
class Handler(BaseHandler):
    headers= {
    "Accept":"text/html,application/xhtml+xml,application/xml;q=0.9,image/webp,*/*;q=0.8",
    "Accept-Encoding":"gzip, deflate, sdch",
    "Accept-Language":"zh-CN,zh;q=0.8",
    "Cache-Control":"max-age=0",
    "Connection":"keep-alive",
    "User-Agent":"Mozilla/5.0 (Windows NT 6.1; WOW64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/41.0.2272.101 Safari/537.36"
    }
    crawl_config = {
        "headers" : headers,
        "timeout" : 100
    }
    @every(minutes= 1)
    def on_start(self):
        self.crawl('http://www.zhanqi.tv/games',callback=self.index_page)
    @config(age=10 * 24 * 60 * 60)
    def index_page(self, response):
        print(response)
        for each in response.doc('a[href^="http://www.zhanqi.tv/games/"]').items():
            if re.match("http://www.zhanqi.tv/games/\w+", each.attr.href, re.U):
                self.crawl(each.attr.href, 
                fetch_type='js', # fetch_type参数                   
                js_script=""" # JavaScript
                function() {
                    setTimeout(window.scrollTo(0,document.body.scrollHeight), 5000);         
                }
                """,callback=self.list_page)  
    @config(age=1*60*60, priority=2)                
    def list_page(self, response):
        for each in response.doc('.active > div.live-list-tabc > ul#hotList.clearfix > li > a').items():
            if re.match("http://www.zhanqi.tv/\w+", each.attr.href, re.U):
                self.crawl(each.attr.href,
                fetch_type='js',                   
                js_script="""
                function() {
                    setTimeout(window.scrollTo(0,document.body.scrollHeight), 5000);        
                }
                """,callback=self.detail_page)    
    @config(age=1*60*60, priority=2)
    def detail_page(self, response):
        for each in response.doc('.video-flash-cont').items():
            d = pq(each)
            print(d.html())
        return {
            "url": response.url,
            "author":response.doc('.meat > span').text(),
            "title":response.doc('.title-name').text(),
            "game-name":response.doc('span > .game-name').text(),
            "users2":response.doc('div.live-anchor-info.clearfix > div.sub-anchor-info > div.clearfix > div.meat-info > span.num.dv.js-onlines-panel > span.dv.js-onlines-txt > span').text(),
            "flash-cont":d.html(),
            "picture":response.doc('.active > img').text(),
            }

七、附录

怎样判断获取pyquery的结果是自己想要的？

通过参看元素的命令行，可以查看结果

八、参考

官方教程：http://docs.pyspider.org/en/latest/tutorial/

pyspider 爬虫教程（一）：HTML 和 CSS 选择器：http://segmentfault.com/a/1190000002477863

pyspider 爬虫教程（二）：AJAX 和 HTTP：http://segmentfault.com/a/1190000002477870

pyspider 爬虫教程（三）：使用 PhantomJS 渲染带 JS 的页面：http://segmentfault.com/a/1190000002477913