scrapy提供了兩種中間件，下載中間件（Downloader Middleware）和Spider中間件（Spider Middleware）

下載中間件

下載中間件是scrapy提供用于用于在爬蟲過程中可修改Request和Response，用于擴展scrapy的功能堰氓；比如：

1. 可以在請求被Download之前怔蚌，請求頭部加上某些信息嫁佳；
2. 完成請求之后纸淮，回包需要解壓等處理允趟；

如何激活下載中間：

在配置文件settings.py中的DOWNLOADER_MIDDLEWARES中配置鍵值對董栽，鍵為要打開的中間件码倦，值為數字，代表優(yōu)先級锭碳，值越低袁稽，優(yōu)先級越高。
scrapy還有一個內部自帶的下載中間件配置DOWNLOADER_MIDDLEWARES_BASE（不可覆蓋）工禾。scrapy在啟用是會結合DOWNLOADER_MIDDLEWARES_BASE和DOWNLOADER_MIDDLEWARES运提，若要取消scrapy默認打開的中間，可在DOWNLOADER_MIDDLEWARES將該中間的值置為0闻葵。

DOWNLOADER_MIDDLEWARES_BASE = 
{
    'scrapy.downloadermiddlewares.robotstxt.RobotsTxtMiddleware': 100,
    'scrapy.downloadermiddlewares.httpauth.HttpAuthMiddleware': 300,
    'scrapy.downloadermiddlewares.downloadtimeout.DownloadTimeoutMiddleware': 350,
    'scrapy.downloadermiddlewares.defaultheaders.DefaultHeadersMiddleware': 400,
    'scrapy.downloadermiddlewares.useragent.UserAgentMiddleware': 500,
    'scrapy.downloadermiddlewares.retry.RetryMiddleware': 550,
    'scrapy.downloadermiddlewares.ajaxcrawl.AjaxCrawlMiddleware': 560,
    'scrapy.downloadermiddlewares.redirect.MetaRefreshMiddleware': 580,
    'scrapy.downloadermiddlewares.httpcompression.HttpCompressionMiddleware': 590,
    'scrapy.downloadermiddlewares.redirect.RedirectMiddleware': 600,
    'scrapy.downloadermiddlewares.cookies.CookiesMiddleware': 700,
    'scrapy.downloadermiddlewares.httpproxy.HttpProxyMiddleware': 750,
    'scrapy.downloadermiddlewares.stats.DownloaderStats': 850,
    'scrapy.downloadermiddlewares.httpcache.HttpCacheMiddleware': 900,
}

如何編寫一個下載中間件

class Scrapy.downloadermiddleares.DownloaderMiddleware

process_request(request, spider)

當每個Request對象經過下載中間件時會被調用民泵，優(yōu)先級越高的中間件，越先調用槽畔；該方法應該返回以下對象：None/Response對象/Request對象/拋出IgnoreRequest異常栈妆；

1. 返回None：scrapy會繼續(xù)執(zhí)行其他中間件相應的方法；
2. 返回Response對象：scrapy不會再調用其他中間件的process_request方法厢钧，也不會去發(fā)起下載鳞尔，而是直接返回該Response對象；
3. 返回Request對象：scrapy不會再調用其他中間件的process_request()方法早直，而是將其放置調度器待調度下載寥假；
4. 拋出IgnoreRequest異常：已安裝中間件的process_exception()會被調用，如果它們沒有捕獲該異常霞扬，則Request.errback會被調用糕韧；如果再沒被處理枫振，它會被忽略，且不會寫進日志萤彩。

process_response(request, response, spider)

當每個Response經過下載中間件會被調用粪滤，優(yōu)先級越高的中間件，越晚被調用雀扶，與process_request()相反杖小；該方法返回以下對象：Response對象/Request對象/拋出IgnoreRequest異常。

1. 返回Response對象：scrapy會繼續(xù)調用其他中間件的process_response方法愚墓；
2. 返回Request對象：停止中間器調用予权，將其放置到調度器待調度下載；
3. 拋出IgnoreRequest異常：Request.errback會被調用來處理函數转绷，如果沒有處理伟件，它將會被忽略且不會寫進日志。

process_exception(request, exception, spider)

當process_exception()和process_request()拋出異常時會被調用议经，應該返回以下對象：None/Response對象/Request對象斧账；

1. 如果返回None：scrapy會繼續(xù)調用其他中間件的process_exception()；
2. 如果返回Response對象：中間件鏈的process_response()開始啟動煞肾，不會繼續(xù)調用其他中間件的process_exception()咧织；
3. 如果返回Request對象：停止中間器的process_exception()方法調用，將其放置到調度器待調度下載籍救。

from_crawler(cls, crawler)

如果存在該函數习绢，from_crawler會被調用使用crawler來創(chuàng)建中間器對象，必須返回一個中間器對象蝙昙，通過這種方式闪萄，可以訪問到crawler的所有核心部件，如settings奇颠、signals等败去。

scray提供的一些下載中間件

以下講述的是一些常用的下載中間件，更多的下載中間件請查看文檔和代碼

HttpProxyMiddleware

scrapy.downloadermiddlewares.httpproxy.HttpProxyMiddleware

用以設置代理服務器烈拒，通過設定Request.meta['proxy']來設置代理圆裕，會從環(huán)境變量http_proxy、https_proxy荆几、no_proxy依次獲认抛薄；我們以http://httpbin.org/ip的返回來測試下：

#shell 命令
export http_proxy='http://193.112.216.55:1234'

# -*- coding: utf-8 -*-
import scrapy

class ProxySpider(scrapy.Spider):
    name = 'proxy'
    allowed_domains = ['httpbin.org']
    start_urls = ['http://httpbin.org/ip']

    def parse(self, response):
        print(response.text)

運行scrapy crawl proxy --nolog吨铸，獲得以下結果：

{"origin":"111.231.115.150, 193.112.216.55"}

返回了我們設置的代理地址IP行拢。

UserAgentMiddleware

scrapy.downloadermiddlewares.useragent.UserAgentMiddleware

通過配置項USER_AGENT設置用戶代理；我們以http://httpbin.org/headers的返回來看看測試下：

settings.py
#...
#UserAgentMiddleware默認打開
USER_AGENT = 'Mozilla/5.0 (Macintosh; Intel Mac OS X 10_13_4) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/66.0.3359.181 Safari/537.36'
#...

# -*- coding: utf-8 -*-
import scrapy

class UserAgentSpider(scrapy.Spider):
    name = 'user_agent'
    allowed_domains = ['httpbin.org']
    start_urls = ['http://httpbin.org/headers']

    def parse(self, response):
        print(response.text)

運行scrapy crawl user_agent --nolog诞吱，獲得以下結果：

{"headers":{"Accept":"text/html,application/xhtml+xml,application/xml;q=0.9,*/*;q=0.8","Accept-Encoding":"gzip,deflate","Accept-Language":"en","Connection":"close","Host":"httpbin.org","User-Agent":"Mozilla/5.0 (Macintosh; Intel Mac OS X 10_13_4) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/66.0.3359.181 Safari/537.36"}}

返回了我們設置的用戶代理剂陡。

使用隨機用戶代理與IP代理

某些網站會通過檢測訪問IP和User-agent來進行反爬蟲狈涮，如果檢測到來自同一IP的大量請求，會判斷該IP正在進行爬蟲鸭栖，故而拒絕請求。有些網站也會檢測User-Agent握巢。我們可以使用多個代理IP和不同的User-Agent來對網站數據進行爬取晕鹊，避免被封禁IP。
我們可以通過繼承HttpProxyMiddleware和UserAgentMiddleware并修改來使得scrapy使用proxy和user-agent按我們的想法來運行暴浦。HttpProxyMiddleware和UserAgentMiddleware見httpproxy.py和useragent.py

代碼如下：

#middlewares.py
# -*- coding: utf-8 -*-

# Define here the models for your spider middleware
#
# See documentation in:
# https://doc.scrapy.org/en/latest/topics/spider-middleware.html

from scrapy import signals
from scrapy.downloadermiddlewares.httpproxy import HttpProxyMiddleware
from scrapy.exceptions import NotConfigured
from collections import defaultdict
from urllib.parse import urlparse
from faker import Faker                 #引入Faker溅话，pip install faker下載
import random

class RandomHttpProxyMiddleware(HttpProxyMiddleware):

    def __init__(self, auth_encoding='latin-1', proxy_list = None):
        if not proxy_list:
            raise NotConfigured
        self.proxies = defaultdict(list)
        for proxy in proxy_list:
            parse = urlparse(proxy)
            self.proxies[parse.scheme].append(proxy)        #生成dict，鍵為協(xié)議歌焦，值為代理ip列表

    @classmethod
    def from_crawler(cls, crawler):
        if not crawler.settings.get('HTTP_PROXY_LIST'):
            raise NotConfigured

        http_proxy_list = crawler.settings.get('HTTP_PROXY_LIST')   #從配置文件中讀取
        auth_encoding = crawler.settings.get('HTTPPROXY_AUTH_ENCODING', 'latin-1')

        return cls(auth_encoding, http_proxy_list)

    def _set_proxy(self, request, scheme):
        proxy = random.choice(self.proxies[scheme])     #隨機抽取選中協(xié)議的IP
        request.meta['proxy'] = proxy
        
class RandomUserAgentMiddleware(object):

    def __init__(self):
        self.faker = Faker(local='zh_CN')
        self.user_agent = ''

    @classmethod
    def from_crawler(cls, crawler):
        o = cls()
        crawler.signals.connect(o.spider_opened, signal=signals.spider_opened)
        return o

    def spider_opened(self, spider):
        self.user_agent = getattr(spider, 'user_agent',self.user_agent)

    def process_request(self, request, spider):
        self.user_agent = self.faker.user_agent()       #獲得隨機user_agent
        request.headers.setdefault(b'User-Agent', self.user_agent)

#settings.py
#...
DOWNLOADER_MIDDLEWARES = {
    'newproject.middlewares.RandomHttpProxyMiddleware': 543,
    'newproject.middlewares.RandomUserAgentMiddleware': 550,
    'scrapy.downloadermiddlewares.httpproxy.HttpProxyMiddleware':None,
    'scrapy.downloadermiddlewares.useragent.UserAgentMiddleware':None,
}
HTTP_PROXY_LIST = [
    'http://193.112.216.55:1234',
    'http://118.24.172.34:1234',
]
#...

#anything.py
# -*- coding: utf-8 -*-
import scrapy
import json
import pprint
class AnythingSpider(scrapy.Spider):
    name = 'anything'
    allowed_domains = ['httpbin.org']
    start_urls = ['http://httpbin.org/anything']

    def parse(self, response):
        ret = json.loads(response.text)
        pprint.pprint(ret)

上面引入了faker庫飞几，該庫是用來偽造數據的庫，十分方便独撇。我們通過訪問http://httpbin.org/anything來得到我們的請求內容屑墨；如下：

#scrapy crawl anything --nolog
{'args': {},
 'data': '',
 'files': {},
 'form': {},
 'headers': {'Accept': 'text/html,application/xhtml+xml,application/xml;q=0.9,*/*;q=0.8',
             'Accept-Encoding': 'gzip,deflate',
             'Accept-Language': 'en',
             'Cache-Control': 'max-age=259200',
             'Connection': 'close',
             'Host': 'httpbin.org',
             'User-Agent': 'Opera/8.85.(Windows NT 5.2; sc-IT) Presto/2.9.177 '
                           'Version/10.00'},
 'json': None,
 'method': 'GET',
 'origin': '193.112.216.55',
 'url': 'http://httpbin.org/anything'}
 
 #scrapy crawl anything --nolog
 {'args': {},
 'data': '',
 'files': {},
 'form': {},
 'headers': {'Accept': 'text/html,application/xhtml+xml,application/xml;q=0.9,*/*;q=0.8',
             'Accept-Encoding': 'gzip,deflate',
             'Accept-Language': 'en',
             'Cache-Control': 'max-age=259200',
             'Connection': 'close',
             'Host': 'httpbin.org',
             'User-Agent': 'Mozilla/5.0 (Macintosh; PPC Mac OS X 10_12_3) '
                           'AppleWebKit/5342 (KHTML, like Gecko) '
                           'Chrome/40.0.810.0 Safari/5342'},
 'json': None,
 'method': 'GET',
 'origin': '118.24.172.34',
 'url': 'http://httpbin.org/anything'}

可以看到，我們的spider通過下載中間件纷铣，不斷的更改了IP和User-Agent卵史。

總結

本篇講述了什么是下載中間件以及如何自定義和啟用下載中間件，最后實踐了自定義下載中間件搜立。后面將會學習另一個中間件：Spider中間件以躯。