Scrapy突破反爬虫的限制

随机切换UserAgent

https://github.com/hellysmile/fake-useragent

scrapy使用fake-useragent

在全局配置文件中禁用掉默认的UA,将其设置为None即可
settings.py

DOWNLOADER_MIDDLEWARES = {
    ...
    'scrapy.contrib.downloadermiddleware.useragent.UserAgentMiddleware': None,
}

在中间件中编写自己的middleware
middlewares.py

class RandomUserAgentMiddleware(object):
    def __init__(self, crawler):
        super(RandomUserAgentMiddleware, self).__init__()
        self.ua = UserAgent()
        self.ua_type = crawler.settings.get('RANDOM_UA_TYPE', 'random')

    @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())

将自己写的middleware配置进settings中

DOWNLOADER_MIDDLEWARES = {
    'myproject.middlewares.CustomDownloaderMiddleware': 543,
    'scrapy.contrib.downloadermiddleware.useragent.UserAgentMiddleware': None,
}

随机切换IP

https://github.com/scrapy-plugins/scrapy-crawlera
爬取西刺IP代理网站获取IP

from scrapy.linkextractors import LinkExtractor
from scrapy.spiders import CrawlSpider, Rule
from w3lib.html import remove_tags

from ArticleSpider.items import ProxyIpItemLoader, ProxyIpItem


class ProxyIpSpider(CrawlSpider):
    name = 'proxy'
    allowed_domains = ['www.xicidaili.com']
    start_urls = ['http://www.xicidaili.com']

    rules = (
        Rule(LinkExtractor(allow=('nn/\d*')), callback='parse_detail', follow=True),
    )

    def parse_detail(self, response):
        ip_list = response.css('#ip_list tr')
        for ipdata in ip_list[1:]:
            item_loader = ProxyIpItemLoader(item=ProxyIpItem(), response=response)
            data = ipdata.css('td')
            item_loader.add_value('ip', data[1].css('td::text').extract_first())
            item_loader.add_value('port', data[2].css('td::text').extract_first())
            item_loader.add_value('addr', self.get_addr(data[3]))
            item_loader.add_value('ishidden', data[4].css('td::text').extract_first())
            item_loader.add_value('type', data[5].css('td::text').extract_first())
            item_loader.add_value('speed', data[6].css('div::attr(title)').extract_first())
            item_loader.add_value('conn_time', data[7].css('div::attr(title)').extract_first())
            item_loader.add_value('live_time', data[8].css('td::text').extract_first())
            item_loader.add_value('check_time', data[9].css('td::text').extract_first())

            proxy_ip_item = item_loader.load_item()

            yield proxy_ip_item


    def get_addr(self, value):
        if value.css('a'):
            return remove_tags(value.extract()).strip()
        else:
            return "未知"

对数据进行简单处理

class ProxyIpItemLoader(ItemLoader):
    default_output_processor = TakeFirst()

def live_time(value):
    '''
    分钟, 小时, 天
    统一转换成分钟
    '''
    if '分钟' in value:
        return int(value.split('分钟')[0])
    elif '小时' in value:
        value = value.split('小时')[0]
        return int(value) * 60
    elif '天' in value:
        value = value.split('天')[0]
        return int(value) * 60 * 24

def ishidden_to_int(value):
    if '高匿' in value:
        return int(1)
    else:
        return int(0)

def check_time(value):
    return datetime.datetime.strptime(value, "%y-%m-%d %H:%M")

class ProxyIpItem(scrapy.Item):
    '''
    {'addr': '陕西西安',
     'check_time': '12-12-31 18:52',
     'conn_time': '0.82秒',
     'ip': '113.133.160.203',
     'ishidden': '高匿',
     'live_time': '1分钟',
     'port': '6675',
     'speed': '3.595秒',
     'type': 'socks4/5'
     }
    '''
    ip = scrapy.Field()
    port = scrapy.Field()
    addr = scrapy.Field(
        input_processor = MapCompose(remove_tags, lambda x:x.strip())
    )
    ishidden = scrapy.Field(
        input_processor=MapCompose(ishidden_to_int)
    )
    type = scrapy.Field()
    speed = scrapy.Field()
    conn_time = scrapy.Field()
    live_time = scrapy.Field(
        input_processor = MapCompose(live_time)
    )
    check_time = scrapy.Field()

    def get_insert_sql(self):
        insert_sql = """
                    insert into proxy_ip(ip, port, addr, ishidden, type, speed, conn_time, live_time, check_time)
                    VALUES (%s, %s, %s, %s,%s,%s, %s, %s, %s)
                """
        params = (self["ip"], self["port"], self["addr"],
                  self["ishidden"], self["type"],self["speed"],
                  self["conn_time"], self["live_time"], self["check_time"])
        return insert_sql, params

在pipeline中进行数据的再次清洗,抛弃所有的特殊端口的item,并数据进行保存
在中间件中创建切换IP的中间件,在主配置文件中启用这个中间件
IP是否可用,只需要请求百度即可

验证码识别

没必要自己写一个验证码识别代码
可以使用云打码平台进行验证码识别
http://www.yundama.com/
需要分别注册一个普通用户和一个开发者账号

下载pythonhttp版本
http://www.yundama.com/apidoc/YDM_SDK.html#DLL
解压后里面有一个3.x的文件,打开后进行配置

# 用户名(普通用户)
username    = 'username'

# 密码(普通用户)
password    = 'password'                            

# 软件ＩＤ，开发者分成必要参数。登录开发者后台【我的软件】获得！
appid       = 1                                     

# 软件密钥，开发者分成必要参数。登录开发者后台【我的软件】获得！
appkey      = '22cc5376925e9387a23cf797cb9ba745'    

# 图片文件
filename    = 'getimage.jpg'                        

# 验证码类型，# 例：1004表示4位字母数字，不同类型收费不同。请准确填写，否则影响识别率。在此查询所有类型 http://www.yundama.com/price.html
codetype    = 1004

# 超时时间，秒
timeout     = 60