Python爬虫知识

一、爬虫

1、概述

网络爬虫，搜索引擎就是爬虫的应用者。

2、爬虫分类

（1）通用爬虫，常见就是搜索引擎，无差别的收集数据，存储，提取关键字，构建索引库，给用户提供搜索接口。

爬取一般流程：

初始化一批URL，将这些url放入到等待爬取队列。

从队列取出这些url，通过dns解析ip，对应ip站点下载HTML页面，保存到本地服务器中，爬取完的url放到已爬取队列。

分析这些网页内容，找出网页里面关心的url连接，继续执行第二步，直到爬取结束。

搜索引擎如何获取一个新网站的url。

新网站主动提交给搜索引擎。

通过其他网站页面中设置的外链。

搜索引擎和dns服务商合作，获取最新收录的网站。

（2）聚焦爬虫

有针对性的编写特定领域数据的爬取程序，针对某些类别数据的采集的爬虫，是面向主题的。

3、robots协议

指定一个robots.txt文件，告诉爬虫引擎什么可以爬取。

这个协议为了让搜索引擎更有效率搜索自己内容，提供了sitemap这样的文件。

这个文件禁止抓取的往往又是可能我们感兴趣的内容，反而泄露了这些地址。。

4、http请求和响应处理

爬虫网页就是通过HTTP协议访问网页，不过通过浏览器访问往往是人的行为，把程序编程人的行为的问题。

Urllib包

from urllib.request import urlopen

response = urlopen('http://www.bing.com')
print(response.closed)

with response:
    print(response.status)
    print(response._method)
    print(response.read())
    print(response.closed)
    print(response.info)
print(response.closed)

使用等，urllib包，使用查询等。

解决useragent问题：

from urllib.request import urlopen,Request

url = 'http://www.bing.com'
ua = 'Mozilla/5.0 (Windows NT 6.2; WOW64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/67.0.3396.99 Safari/537.36'

req = Request(url,headers={'User-agent':ua})
response = urlopen(req,timeout=10)
# print(req)
print(response.closed)

with response:
    print(response.status)
    print(response._method)
    # print(response.read())
    # print(response.closed)
    # # print(response.info)
    print(response.geturl())
print(req.get_header('User-agent'))
print(response.closed)

Chrome浏览器获取useragent

5、parse

from urllib import parse

d = {
    'id':1,
    'name':'tom',
    'url':'http://www.magedu.com'
}

url = 'http://www.magedu.com'
u = parse.urlencode(d)   #url编码
print(u)

print(parse.unquote(u))#解码

6、请求方法

from urllib import parse
import simplejson

base_url = 'http://cn.bing.com/search'

d = {
    'q':'马哥教育'
}
# d = {
#     'id':1,
#     'name':'tom',
#     'url':'http://www.magedu.com'
# }

# url = 'http://www.magedu.com'
u = parse.urlencode(d)   #url编码

# url = '{}?{}'.format(base_url,u)
# print(url)
#
# print(parse.unquote(url))#解码

from urllib.request import urlopen,Request

url = 'http://httpbin.org/post'

ua = 'Mozilla/5.0 (Windows NT 6.2; WOW64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/67.0.3396.99 Safari/537.36'

data = parse.urlencode({'name':'张三，@=/&*','age':'6'})

req = Request(url,headers={
    'User-agent':ua
})

# res = urlopen(req)

with urlopen(req,data= data.encode()) as res:
    text = res.read()
    d = simplejson.loads(text)
    print(d)
    # with open('c:/assets/bing.html','wb+') as f:
        # f.write(res.read())
        # f.flush()

7、爬取豆瓣网

from urllib.request import Request,urlopen
import simplejson
from urllib import parse

ua = 'Mozilla/5.0 (Windows NT 6.2; WOW64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/67.0.3396.99 Safari/537.36'

jurl = 'https://movie.douban.com/j/search_subjects'

d = {
    'type':'movie',
    'tag':'热门',
    'page_limit':10,
    'page_start':10
}

req = Request('{}?{}'.format(jurl,parse.urlencode(d)),headers={
    'User-agent':ua
})

with urlopen(req) as res:
    sub = simplejson.loads(res.read())
    print(len(sub))
    print(sub)

8、解决https，ca证书的问题

忽略证书，ssl

from urllib.request import Request,urlopen

from urllib import parse
import ssl

#request = Request('http://www.12306.cn/mormhweb')
request = Request('http://www.baidu.com')
request.add_header('User-agent','Mozilla/5.0 (Windows NT 6.2; WOW64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/67.0.3396.99 Safari/537.36'
)

context = ssl._create_unverified_context() #忽略不可用证书

with urlopen(request,context=context) as res:
print(res._method)
print(res.read())

9、urllib3

pip install urllib3

import urllib3

url = 'http://movie.douban.com'

ua = 'Mozilla/5.0 (Windows NT 6.2; WOW64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/67.0.3396.99 Safari/537.36'

with urllib3.PoolManager() as http:   #连接池管理器
    response = http.request('GET',url,headers={'User-agent':ua})
    print(1,response)
    print(2,type(response))
    print(3,response.status,response.reason)
    print(4,response.headers)
    print(5,response.data)

import urllib3
from urllib.parse import urlencode
from urllib3 import HTTPResponse

url = 'http://movie.douban.com'

ua = 'Mozilla/5.0 (Windows NT 6.2; WOW64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/67.0.3396.99 Safari/537.36'

jurl = 'https://movie.douban.com/j/search_subjects'

d = {
    'type':'movie',
    'tag':'热门',
    'page_limit':10,
    'page_start':10
}

# with urllib3.PoolManager() as http:   #连接池管理器
#     response = http.request('GET',url,headers={'User-agent':ua})   #可以指定请求方法
#     print(1,response)
#     print(2,type(response))
#     print(3,response.status,response.reason)
#     print(4,response.headers)
#     print(5,response.data)

with urllib3.PoolManager() as http:
    response = http.request('GET','{}?{}'.format(jurl,urlencode(d)),headers={'User-agent':ua})
    print(response)
    print(response.status)
    print(response.data)

10、requests库

Requests使用了urllib3.

pip install requests

带会话的方式 session。

会把请求头等信息自动管理。

import urllib3
from urllib.parse import urlencode
from urllib3 import HTTPResponse
import requests

# url = 'http://movie.douban.com'

ua = 'Mozilla/5.0 (Windows NT 6.2; WOW64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/67.0.3396.99 Safari/537.36'

jurl = 'https://movie.douban.com/j/search_subjects'

d = {
    'type':'movie',
    'tag':'热门',
    'page_limit':10,
    'page_start':10
}
# url = '{}?{}'.format(jurl,urlencode(d))

# with urllib3.PoolManager() as http:   #连接池管理器
#     response = http.request('GET',url,headers={'User-agent':ua})   #可以指定请求方法
#     print(1,response)
#     print(2,type(response))
#     print(3,response.status,response.reason)
#     print(4,response.headers)
#     print(5,response.data)

# with urllib3.PoolManager() as http:
#     response = http.request('GET','{}?{}'.format(jurl,urlencode(d)),headers={'User-agent':ua})
#     print(response)
#     print(response.status)
#     print(response.data)

# response = requests.request('GET',url,headers = {'User-agent':ua})
#
# with response:
#     print(response.text)
#     print(response.status_code)
#     print(response.url)
#     print(response.headers)
#     print(response.request)
urls = ['https://www.baidu.com/s?wd=magedu','https://www.baidu.com/s?wd=magedu']

session = requests.Session()
with session:
    for url in urls:
        response = session.get(url,headers={'User-agent':ua})
        with response:
            print(1,response.text)
            print(2,response.status_code)
            print(3,response.url)
            print(4,response.headers)
            print(5,response.request.headers)
            print('--------')
            print(response.cookies)
            print('--------------')
            print(response.cookies)

11、特别注意

个别网站登录的时候cookie，登录的时候要把原来的cookie带回去，然后登录成功后其给你返回一个新的，否则不能进行相关操作。有些时候只是带一些cookie相关的值即可。

反爬措施：对于用户发起的请求来检测上一次是否访问的是我的网站。

在network的referer里面显示上一次访问网站的哪个一页。

Files：上传的文件内容。

路由器的将用户名和密码加密放在请求头里面。

Cert证书。

Requests基本功能：

import requests

ua = 'Mozilla/5.0 (Windows NT 6.2; WOW64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/63.0.3239.26 Safari/537.36 Core/1.63.5514.400 QQBrowser/10.1.1660.400'
url = 'https://dig.chouti.com/login'

data = {
'phone':'8618804928235',
'password':'tana248654',
'oneMonth':'1'
}

r1_urls = 'https://dig.chouti.com'
r1 = requests.get(url=r1_urls,headers={'User-Agent':ua})
# print(r1.text)
r1_cookie = r1.cookies.get_dict()
print('r1',r1.cookies)

response = requests.post(url,data,headers={'User-Agent':ua},cookies=r1_cookie)

print(response.text)
print(response.cookies.get_dict())

r3 = requests.post(url='https://dig.chouti.com/link/vote?linksId=21718341',
cookies={'gpsd':r1_cookie.get('gpsd')},headers={'User-Agent':ua})

print(r3.text)

二、HTML解析

通过上面的库，可以拿到HTML内容。

1、Xpath

http://www.qutoric.com/xmlquire/

站点。

路径的遍历，查找到需要的内容。

2、lxml库

解析HTML的库。

https://lxml.de/

安装：

pip install lxml

爬取豆瓣网top10

import urllib3
from urllib.parse import urlencode
from urllib3 import HTTPResponse
import requests
from lxml import etree

# url = 'http://movie.douban.com'

ua = 'Mozilla/5.0 (Windows NT 6.2; WOW64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/67.0.3396.99 Safari/537.36'

jurl = 'https://movie.douban.com/j/search_subjects'

d = {
    'type':'movie',
    'tag':'热门',
    'page_limit':10,
    'page_start':10
}
# urls = ['https://www.baidu.com/s?wd=magedu','https://www.baidu.com/s?wd=magedu']
urls = ['https://movie.douban.com/']

session = requests.Session()
with session:
    for url in urls:
        response = session.get(url,headers={'User-agent':ua})
        with response:
            content = response.text

        html = etree.HTML(content)
        title = html.xpath("//div[@class='billboard-bd']//tr")
        for t in title:
            txt = t.xpath('.//text()')
            print(''.join(map(lambda x:x.strip(),txt)))
            # print(t)

3、beautifulsoup4

4、可以导航的string（navigablestring）

深度优先遍历。

Soup.findall().

Soup.findall(id =’header’)

5、css选择器

Soup.select 正则表达式

Pip install jsonpath.

from concurrent.futures import ThreadPoolExecutor
import threading
import time
from queue import Queue
import logging
import requests
from bs4 import BeautifulSoup

event = threading.Event()
url = 'https://news.enblogs.com'
path = '/n/page/'
ua = 'Mozilla/5.0 (Windows NT 6.2; WOW64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/67.0.3396.99 Safari/537.36'

urls = Queue()
htmls = Queue()
outps = Queue()

def create_urls(start,stop,step=1):
    for i in range(start,stop+1,step):
        url1 = '{}{}{}/'.format(url,path,i)
        urls.put(url1)

def crawler():
    while not event.is_set():
        try:
            url1 = urls.get(True,1)

            response = requests.get(url,headers={'User-agent':ua})
            with response:
                html = response.text
                htmls.put(html)
        except Exception as e:
            print(1,e)

def parse():
    while not event.is_set():
        try:
            html = htmls.get(True,1)
            soup = BeautifulSoup(html,'lxml')
            news = soup.select('h2.news_entry a')

            for n in news:
                txt = n.text
                url1 = url + n.attrs.get('href')
                outps.put((txt,url1))

        except Exception as e:
            print(e)

def save(path):
    with open(path,'a+',encoding='utf-8') as f:
        while not event.is_set():
            try:
                title,url1 = outps.get(True,1)
                f.write('{}{} '.format(title,url1))
                f.flush()
            except Exception as e:
                print(e)

executor = ThreadPoolExecutor(max_workers=10)
executor.submit(create_urls,1,10)
executor.submit(parse)
executor.submit(save,'c:/new.txt')

for i in range(7):
    executor.submit(crawler)

while True:
    cmd = input('>>>')
    if cmd.strip() == 'q':
        event.set()
        executor.shutdown()
        print('close')
        time.sleep()
        break

三、动态网页处理

很多网站采用的是ajax技术，spa技术。部分内容都是异步加载的，提高用户体验。

1、phantomjs无头浏览器

http://phantomjs.org/

Xml http 与后端服务器建立的连接。

2、selenium

（1）自动化测试工具等，可以直接截图。模仿浏览器的行为等。

from selenium import webdriver
import datetime
import time
import random

driver = webdriver.PhantomJS('c:/assets/phantomjs-2.1.1-windows/bin/phantomjs.exe')

driver.set_window_size(1024,1024)
url = 'https://cn.bing.com/search?q=%E9%A9%AC%E5%93%A5%E6%95%99%E8%82%B2'
driver.get(url)

def savedic():
    try:
      base_dir = 'C:/assets/'
        filename = '{}{:%Y%m%d%H%M%S}{}.png'.format(base_dir,datetime.datetime.now(),random.randint(1,100))
        driver.save_screenshot(filename)
    except Exception as e:
        print(1,e)
# time.sleep(6)
# print('-------')
# savedic()
MAXRETRIES = 5
while MAXRETRIES:
    try:
        ele = driver.find_element_by_id('b_results')
        print(ele)
        print('===========')
        savedic()
        break
    except Exception as e:
        print(e)
        print(type(e))
    time.sleep(1)
    MAXRETRIES -= 1

查找数据等，异步的方式。

（2）下拉框子使用，使用Select。

3、模拟键盘输入

模仿浏览器登录，先找到登录框的id，然后，setkeys。

之后返回登录后的网页。

from selenium import webdriver
from selenium.webdriver.common.keys import Keys
import time
import random
import datetime

driver = webdriver.PhantomJS('c:/assets/phantomjs-2.1.1-windows/bin/phantomjs.exe')

driver.set_window_size(1024,1024)

url = 'https://www.oschina.net/home/login?goto_page=https%3A%2F%2Fwww.oschina.net%2F'

def savedic():
    try:
        base_dir = 'C:/assets/'
        filename = '{}{:%Y%m%d%H%M%S}{}.png'.format(base_dir,datetime.datetime.now(),random.randint(1,100))
        driver.save_screenshot(filename)
    except Exception as e:
        print(1,e)

driver.get(url)
print(driver.current_url,111111111111)
savedic()

email = driver.find_element_by_id('userMail')
passwed = driver.find_element_by_id('userPassword')

email.send_keys('604603701@qq.com')
passwed.send_keys('tana248654')
savedic()
passwed.send_keys(Keys.ENTER)

time.sleep(2)
print(driver.current_url,2222222222)
userinfo = driver.find_element_by_class_name('user-info')
print(userinfo.text)
time.sleep(2)
cookie = driver.get_cookies()
print(cookie)
savedic()

4、页面等待

（1）time.sleep

数据js加载需要一定的时间内。

线程休眠。

设置尝试的次数等

（2）selenium里面的wait

显示等待

from selenium.webdriver.common.by import By
from selenium.webdriver.support.ui import WebDriverWait
from selenium.webdriver.support import expected_conditions as EC

try:
    email = WebDriverWait(driver,10).until(
        EC.presence_of_all_elements_located((By.ID,'userMail'))
    )
    savedic()
finally:
    driver.quit()

隐士的等待

driver.implicitly_wait(10)

总结：

四、scrapy框架

1、安装

Pip install scrapy 可能报错，报错的原因是下载tw开头的文件.whl文件，然后pip安装。

2、使用

scrapy startproject scrapyapp 开启一个项目

scrapy genspider donz_spider dnoz.org 进入spider文件下创建一个新的模块，把要爬取的网站加到url列表中。

scrapy genspider -t basic dbbook douban.com 继承自baseic模板。内容少。

scrapy genspider -t crawl book douban.com 继承自crawl模板，内容多。

-t 后面加的是模板。然后名字和网站

scrapy crawl donz_spider 运行代码，运行时候报错的话pip install pypiwin32

from scrapy.http.response.html import HtmlResponse

response 继承于HTMLResponse。

在item设置中设置要爬取的信息的类例如标题。

在spiders下的文件里面写爬虫的xpath，爬取的队列及爬取内容的匹配。

Middlewares里面是中间件。

Pipelines里面处理函数。

五、scrapy-redis组件

1、scrapy-redis使用

Pip install scrapy_redis

使用redis作为队列需要的配置文件

Setting.py

BOT_NAME = 'scrapyapp'

SPIDER_MODULES = ['scrapyapp.spiders']
NEWSPIDER_MODULE = 'scrapyapp.spiders'

USER_AGENT = 'Mozilla/5.0 (Windows NT 6.2; WOW64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/67.0.3396.99 Safari/537.36'

ROBOTSTXT_OBEY = False
DOWNLOAD_DELAY = 1
COOKIES_ENABLED = False

SCHEDULER = "scrapy_redis.scheduler.Scheduler"
DUPEFILTER_CLASS = "scrapy_redis.dupefilter.RFPDupeFilter"

ITEM_PIPELINES = {      # redis数据库连接相关
    'scrapyapp.pipelines.ScrapyappPipeline': 300,
    'scrapy_redis.pipelines.RedisPipeline': 543,
}

REDIS_HOST = '192.168.118.130'
REDIS_PORT = 6379

# LOG_LEVEL = 'DEBUG'

Spiders 下面的爬虫文件.py

# -*- coding: utf-8 -*-
import scrapy
from scrapy.linkextractors import LinkExtractor
from scrapy.spiders import CrawlSpider, Rule
from scrapy_redis.spiders import RedisCrawlSpider
from ..items import MovieItem

class MoviecommentSpider(RedisCrawlSpider):
    name = 'moviecomment'
    allowed_domains = ['douban.com']
    # start_urls = ['http://douban.com/']
    redis_key = 'moviecomment1:start_urls'

    rules = (
        Rule(LinkExtractor(allow=r'start=d+'), callback='parse_item', follow=False),
    )

    def parse_item(self, response):
        # i = {}
        #i['domain_id'] = response.xpath('//input[@id="sid"]/@value').extract()
        #i['name'] = response.xpath('//div[@id="name"]').extract()
        #i['description'] = response.xpath('//div[@id="description"]').extract()
        # return i
        comment = '//div[@class="comment-item"]//span[@class="short"]/text()'
        reviews = response.xpath(comment).extract()
        for review in reviews:
            item = MovieItem()
            item['comment'] = review.strip()
            yield item

Item.py

import scrapy

class MovieItem(scrapy.Item):
    # define the fields for your item here like:
    # name = scrapy.Field()
    comment = scrapy.Field()

redis数据中要设置一个key值和movecomment.py 中的redis_key = 'moviecomment1:start_urls' 设置value及初始的url值。

完成后数据库会存储响应的值

可以在redis-cli 后面加上 –ra

2、分析

（1）jieba分词

Pip install jieba

（2）stopword停用词

数据清洗：把脏数据洗掉，检测出并除去数据中无效或者无关的数据，例如空值，非法值的检测，重复数据检测等。

（3）词云

Pip install wordcloud

from redis import Redis
import json
import jieba

redis = Redis()
stopwords = set()
with open('', encoding='gbk') as f:
    for line in f:
        print(line.rstrip(' ').encode())
        stopwords.add(line.rstrip(' '))
print(len(stopwords))
print(stopwords)
items = redis.lrange('dbreview:items', 0, -1)
print(type(items))

words = {}
for item in items:
    val = json.loads(item)['review']
    for word in jieba.cut(val):
        words[word] = words.get(word, 0) + 1
print(len(words))
print(sorted(words.items(), key=lambda x: x[1], reverse=True))

分词代码测试

六、scrapy项目

1、知识回顾

2、爬取技术网站

praise_nums = response.xpath("//span[contains(@class, 'vote-post-up')]/text()").extract()
fav_nums = response.xpath("//span[contains(@class, 'bookmark-btn')]/text()").extract()
# match_re = re.match(".*(d+).*", fav_nums)

class的值有多个的时候，使用container进行选取。

from scrapy.http import Request #找到的url传递给下一级
from urllib import parse

#提取下一页并交给scrapy下载
next_url = response.xpath('//div[@class="navigation margin-20"]/a[4]/@href').extract()
if next_url:
yield Request(url=parse.urljoin(response.url, next_url), callback=self.parse)

（1）图片处理及存储：

pip install pillow

IMAGES_URLS_FIELD = "front_image_url"
project_dir = os.path.abspath(os.path.dirname(__file__))
IMAGES_STORE = os.path.join(project_dir, 'images')

（2）写入到本地文件：

class JsonWithEncodingPipeline(object):
    def __init__(self):
        self.file = codecs.open('article.json', 'w', encoding='utf-8')

    def process_item(self, item, spider):
        lines = json.dumps(dict(item), ensure_ascii=False) + " "
        self.file.write(lines)
        return item

    def spider_closed(self, spider):
        self.file.close()

scrapy自带的JsonItemExporter

（3）导出功能，还有csv文件等

class JsonItemExporterPipeline(object):
    '''
    调用scrapy的JsonItemExporter
    '''
    def __init__(self):
        self.file = open('articleexport.json', 'wb')
        self.exporter = JsonItemExporter(self.file, encoding="utf-8", ensure_ascii=False)
        self.exporter.start_exporting()

    def close_spider(self, spider):
        self.exporter.finish_exporting()
        self.file.close()

    def process_item(self, item, spider):
        self.exporter.export_item(item)
        return item

（4）数据库插入操作

class MysqlPipeline(object):
    def __init__(self):
        self.conn = MySQLdb.connect('192.168.118.131', 'wang', 'wang', 'scrapy_jobbole', charset='utf8', use_unicode=True)
        self.cursor = self.conn.cursor()

    def process_item(self, item, spider):
        insert_sql = """
        insert into jobbole_article(title, url, create_date, fav_nums)
        values (%s, %s, %s, %s)
        """
        self.cursor.execute(insert_sql, (item['title'], item['url'], item['create_date'], item['fav_nums']))
        self.conn.commit()

（5）scrapy提供的异步方法

import MySQLdb
import MySQLdb.cursors
from twisted.enterprise import adbapi

class MysqlTwistedPipeline(object):
    def __init__(self, dbpool):
        self.dbpool = dbpool

    @classmethod
    def from_settings(cls, settings):
        dbparms = dict(
            host=settings['MYSQL_HOST'],
            db=settings['MYSQL_DBNAME'],
            user=settings['MYSQL_USER'],
            password = settings['MYSQL_PASSWORD'],
            charset='utf8',
            cursorclass = MySQLdb.cursors.DictCursor,
            use_unicode = True
        )

        dbpool = adbapi.ConnectionPool('MySQLdb', **dbparms)
        return cls(dbpool)

    def process_item(self, item, spider):
        '''
        异步操作
        :param item:
        :param spider:
        :return:
        '''
        query = self.dbpool.runInteraction(self.do_insert, item)
        query.addErrback(self.handle_error)

    def handle_error(self, failure):
        '''
        处理插入的异常
        :param failure:
        :return:
        '''
        print(failure)

    def do_insert(self, cursor, item):
        '''
        执行具体插入
        :param cursor:
        :param item:
        :return:
        '''
        insert_sql = """
        insert into jobbole_article(title, url, create_date, fav_nums)
        values (%s, %s, %s, %s)
      """
        cursor.execute(insert_sql, (item['title'], item['url'], item['create_date'], item['fav_nums']))

（5）将django的model集成到scrapy

Scrapy-djangoitem

（6）改变超多的xpath和css，使用itemloader

# 通过itemloader加载item
item_loader = ArticleItemLoader(item=ArticleItem(), response=response)
# item_loader.add_css()
item_loader.add_xpath('title', '//div[@class="entry-header"]/h1/text()')

可以在item里面的field里面选择，

class ArticleItem(scrapy.Item):
    title = scrapy.Field(
        input_processor=MapCompose(add_jobbole)
    )
    create_date = scrapy.Field(
        input_processor=MapCompose(add_time)
    )

自定义输出：

class ArticleItemLoader(ItemLoader):
# 自定义item loader
default_output_processor = TakeFirst()

pipeline后面的数值是优先级的问题

七、反爬虫策略

1、修改settings和middlewares文件

Setting里面设置一个user-agent-list的列表。

Middlewares里面设置

class RandomUserAgentMiddlware(object):
    '''
    随机更换user-agent
    '''
    def __init__(self, crawler):
        super(RandomUserAgentMiddlware, self).__init__()
        self.user_agent_list = crawler.settings.get("user_agent_list", [])

    @classmethod
    def from_crawler(cls, crawler):
        return cls(crawler)

    def process_request(self, request, spider):
        request.headers.setdefault('User-Agent', random())

2、随意更换user-agent 的库

>pip install fake-useragent

from fake_useragent import UserAgent

class RandomUserAgentMiddlware(object):
    '''
    随机更换user-agent
    '''
    def __init__(self, crawler):
        super(RandomUserAgentMiddlware, self).__init__()
        # self.user_agent_list = crawler.settings.get("user_agent_list", [])
        self.ua = UserAgent()
    @classmethod
    def from_crawler(cls, crawler):
        return cls(crawler)

    def process_request(self, request, spider):
        request.headers.setdefault('User-Agent', self.ua.random)

class RandomUserAgentMiddlware(object):
    '''
    随机更换user-agent
    '''
    def __init__(self, crawler):
        super(RandomUserAgentMiddlware, self).__init__()
        # self.user_agent_list = crawler.settings.get("user_agent_list", [])
        self.ua = UserAgent()
        self.ua_type = crawler.settings.get("RANDOM_UA_TYPE", "random") 配置项
    @classmethod
    def from_crawler(cls, crawler):
        return cls(crawler)

    def process_request(self, request, spider):
        def get_ua():
            return getattr(self.ua, self.ua_type)
        request.headers.setdefault('User-Agent', get_ua())

随机选取一个user-agent

3、代理ip

普通ip代理

request.meta['proxy'] = "http://61.135.217.7:80" #ip 代理

（1）直接设置普通ip

（2）首先爬取某代理网站的代理ip存入到数据库中，然后从数据库中找到数据，放到middlewares里面进行ip代理。

import requests
from scrapy.selector import Selector
import MySQLdb
import threading
from fake_useragent import UserAgent

conn = MySQLdb.connect(host='127.0.0.1', user='root', passwd='centos', db='test', charset='utf8')
cour = conn.cursor()

ua = UserAgent()

def crawl_ips():
    headers = {
        'User-Agent': 'Mozilla/5.0 (Windows NT 6.2; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/69.0.3497.100 Safari/537.36',
    }
    for i in range(3):
        re = requests.get('http://www.xicidaili.com/wt/{0}'.format(i), headers=headers)

    seletor = Selector(text=re.text)
    all_trs = seletor.css('#ip_list tr')
    ip_list = []
    for tr in all_trs:
        speed_strs = tr.css(".bar::attr(title)").extract()
        if speed_strs:
            speed_str = speed_strs[0]

        all_texts = tr.css('td::text').extract()
        if all_texts:
            ip = all_texts[0]
            port = all_texts[1]
            proxy_type = all_texts[5]
            ip_list.append((ip, port, proxy_type, speed_str.split('秒')[0]))

        for ip_info in ip_list:
            cour.execute(
                "insert xici_ip_list(ip, port, speed, proxy_type) VALUES('{0}', '{1}', '{2}', '{3}')".format(
                    ip_info[0], ip_info[1], ip_info[3], ip_info[2])

            )
            conn.commit()
            print('数据库写入完成')

# crawl_ips()

class GetIP(object):
    def delete_ip(self, ip):
        delete_sql = """
        delete from xici_ip_list where ip='{0}'
        """.format(ip)
        cour.execute(delete_sql)
        conn.commit()
        return True

    def judge_ip(self, ip, port):
        http_url = 'http://ww.baidu.com'
        proxy_url = 'http://{}:{}'.format(ip, port)
        try:
            proxy_dict = {
                'http': proxy_url
            }
            response = requests.get(http_url, proxies=proxy_dict)
        except Exception as e:
            print('invalid ip and port')
            self.delete_ip(ip)
            return False
        else:
            code = response.status_code
            if code >= 200 and code < 300:
                print('eddective ip')
                return True
            else:
                print('invalid ip and port')
                self.delete_ip(ip)
                return False

    def get_random_ip(self):
        # 从数据库中随机获取一个ip
        sql = """
        SELECT ip, port FROM xici_ip_list
        ORDER BY RAND()
        LIMIT 1
        """
        result = cour.execute(sql)
        for ip_info in cour.fetchall():
            ip = ip_info[0]
            port = ip_info[1]
            judge_ip = self.judge_ip(ip, port)
            if judge_ip:
                return "http://{0}:{1}".format(ip, port)
            else:
                return self.get_random_ip()

# t = threading.Thread(target=crawl_ips)
# t.start()

get_ip = GetIP()

get_ip.get_random_ip()

class RandomProxyMiddleware(object):
    #动态设计ip代理
    def process_request(self, request, spider):
        get_ip = GetIP()
        request.meta['proxy'] = get_ip.get_random_ip() #ip 代理

（3）插件化scrapy-proxies

https://github.com/aivarsk/scrapy-proxies/blob/master/scrapy_proxies

（4）scrapy-crawlera

收费版本

（5）tor洋葱网络

https://github.com/aivarsk/scrapy-proxies/blob/master/scrapy_proxies

稳定版本

八、验证码识别

1、验证码识别方法

编码实现tesseract-cor

在线打码

http://www.yundama.com/

人工打码