爬虫之Scripy
Scrapy是一个为了爬取网站数据，提取结构性数据而编写的应用框架。 其可以应用在数据挖掘，信息处理或存储历史数据等一系列的程序中。
其最初是为了页面抓取 (更确切来说, 网络抓取 )所设计的， 也可以应用在获取API所返回的数据(例如 Amazon Associates Web Services ) 或者通用的网络爬虫。Scrapy用途广泛，可以用于数据挖掘、监测和自动化测试。
Scrapy 使用了 Twisted异步网络库来处理网络通讯。整体架构大致如下
- 引擎(Scrapy)
用来处理整个系统的数据流处理, 触发事务(框架核心) - 调度器(Scheduler)
用来接受引擎发过来的请求, 压入队列中, 并在引擎再次请求的时候返回. 可以想像成一个URL（抓取网页的网址或者说是链接）的优先队列, 由它来决定下一个要抓取的网址是什么, 同时去除重复的网址 - 下载器(Downloader)
用于下载网页内容, 并将网页内容返回给蜘蛛(Scrapy下载器是建立在twisted这个高效的异步模型上的) - 爬虫(Spiders)
爬虫是主要干活的, 用于从特定的网页中提取自己需要的信息, 即所谓的实体(Item)。用户也可以从中提取出链接,让Scrapy继续抓取下一个页面 - 项目管道(Pipeline)
负责处理爬虫从网页中抽取的实体，主要的功能是持久化实体、验证实体的有效性、清除不需要的信息。当页面被爬虫解析后，将被发送到项目管道，并经过几个特定的次序处理数据。 - 下载器中间件(Downloader Middlewares)
位于Scrapy引擎和下载器之间的框架，主要是处理Scrapy引擎与下载器之间的请求及响应。 - 爬虫中间件(Spider Middlewares)
介于Scrapy引擎和爬虫之间的框架，主要工作是处理蜘蛛的响应输入和请求输出。 - 调度中间件(Scheduler Middewares)
介于Scrapy引擎和调度之间的中间件，从Scrapy引擎发送到调度的请求和响应。
Scrapy运行流程大概如下：
- 引擎从调度器中取出一个链接(URL)用于接下来的抓取
- 引擎把URL封装成一个请求(Request)传给下载器
- 下载器把资源下载下来，并封装成应答包(Response)
- 爬虫解析Response
- 解析出实体（Item）,则交给实体管道进行进一步的处理
- 解析出的是链接（URL）,则把URL交给调度器等待抓取
安装：
#scrapy 的一些依赖：pywin32、pyOpenSSL、Twisted、lxml 、zope.interface。（安装的时候，注意看报错信息） #安装wheel pip3 install wheel-i http://pypi.douban.com/simple --trusted-host pypi.douban.com #安装这个依赖包，才有安装上Twisted pip3 install Incremental -i http://pypi.douban.com/simple --trusted-host pypi.douban.com #再pip3安装Twisted，但是还是安装不成功，会报错。（解决其它依赖问题） pip3 install Twisted -i http://pypi.douban.com/simple --trusted-host pypi.douban.com #再进入软件存放目录，再安装就可以成功啦。 pip3 install Twisted-17.1.0-cp35-cp35m-win32.whl #安装scrapy pip3 install scrapy -i http://pypi.douban.com/simple --trusted-host pypi.douban.com #pywin32 下载：https://sourceforge.net/projects/pywin32/files/
创建：
#创建项目 scrapy startproject xiaohuar #进入项目 cd xiaohuar #创建爬虫应用 scrapy genspider xiaohuar xiaohar.com #运行爬虫 scrapy crawl chouti --nolog
目录：
project_name/ scrapy.cfg project_name/ __init__.py items.py pipelines.py settings.py spiders/ __init__.py
解释：
- scrapy.cfg 项目的配置信息，主要为Scrapy命令行工具提供一个基础的配置信息。（真正爬虫相关的配置信息在settings.py文件中）
- items.py 设置数据存储模板，用于结构化数据，如：Django的Model
- pipelines 数据处理行为，如：一般结构化的数据持久化
- settings.py 配置文件，如：递归的层数、并发数，延迟下载等
- spiders 爬虫目录，如：创建文件，编写爬虫规则
注意：一般创建爬虫文件时，以网站域名命名
选择器：
#!/usr/bin/env python # -*- coding:utf-8 -*- from scrapy.selector import Selector, HtmlXPathSelector from scrapy.http import HtmlResponse html = """<!DOCTYPE html> <html> <head lang="en"> <meta charset="UTF-8"> <title></title> </head> <body> <ul> <li class="item-"><a id='i1' href="link.html">first item</a></li> <li class="item-0"><a id='i2' href="llink.html">first item</a></li> <li class="item-1"><a href="llink2.html">second item<span>vv</span></a></li> </ul> <div><a href="llink2.html">second item</a></div> </body> </html> """ response = HtmlResponse(url='http://example.com', body=html,encoding='utf-8') # hxs = HtmlXPathSelector(response) # print(hxs) # hxs = Selector(response=response).xpath('//a') # print(hxs) # hxs = Selector(response=response).xpath('//a[2]') # print(hxs) # hxs = Selector(response=response).xpath('//a[@id]') # print(hxs) # hxs = Selector(response=response).xpath('//a[@id="i1"]') # print(hxs) # hxs = Selector(response=response).xpath('//a[@href="link.html"][@id="i1"]') # print(hxs) # hxs = Selector(response=response).xpath('//a[contains(@href, "link")]') # print(hxs) # hxs = Selector(response=response).xpath('//a[starts-with(@href, "link")]') # print(hxs) # hxs = Selector(response=response).xpath('//a[re:test(@id, "i\d+")]') # print(hxs) # hxs = Selector(response=response).xpath('//a[re:test(@id, "i\d+")]/text()').extract() # print(hxs) # hxs = Selector(response=response).xpath('//a[re:test(@id, "i\d+")]/@href').extract() # print(hxs) # hxs = Selector(response=response).xpath('/html/body/ul/li/a/@href').extract() # print(hxs) # hxs = Selector(response=response).xpath('//body/ul/li/a/@href').extract_first() # print(hxs) # ul_list = Selector(response=response).xpath('//body/ul/li') # for item in ul_list: # v = item.xpath('./a/span') # # 或 # # v = item.xpath('a/span') # # 或 # # v = item.xpath('*/a/span') # print(v)
自定义扩展：
自定义扩展时，利用信号在指定位置注册制定操作
from scrapy import signals class MyExtension(object): def __init__(self, value): self.value = value @classmethod def from_crawler(cls, crawler): val = crawler.settings.getint('MMMM') ext = cls(val) crawler.signals.connect(ext.spider_opened, signal=signals.spider_opened) crawler.signals.connect(ext.spider_closed, signal=signals.spider_closed) return ext def spider_opened(self, spider): print('open') def spider_closed(self, spider): print('close')
自定义去重复：
scrapy默认使用 scrapy.dupefilter.RFPDupeFilter 进行去重，相关配置有：
DUPEFILTER_CLASS = 'scrapy.dupefilter.RFPDupeFilter' DUPEFILTER_DEBUG = False JOBDIR = "保存范文记录的日志路径，如：/root/" # 最终路径为 /root/requests.seen
自定义：
#偶合性低，给url去重使用 class RepeatFilter(object): def __init__(self): self.visited_set = set() @classmethod def from_settings(cls, settings): return cls() def request_seen(self, request): if request.url in self.visited_set:#先看当前url在不在visited_set return True self.visited_set.add(request.url) #如果不在就加进去 return False def open(self): # 每次开始的时候都会调用 # print('open') pass def close(self, reason): #每次结束的时候都会调用 # print('close') pass def log(self, request, spider):#每次捕捉到重复的url都会写在log里面 # print('log....') pass
settings：
# 1. 爬虫名称 # BOT_NAME = 'step8_king' # 2. 爬虫应用路径 # SPIDER_MODULES = ['step8_king.spiders'] # NEWSPIDER_MODULE = 'step8_king.spiders' # Crawl responsibly by identifying yourself (and your website) on the user-agent # 3. 客户端 user-agent请求头 # USER_AGENT = 'step8_king (+http://www.yourdomain.com)' # Obey robots.txt rules # 4. 禁止爬虫配置 # ROBOTSTXT_OBEY = False # Configure maximum concurrent requests performed by Scrapy (default: 16) # 5. 并发请求数 # CONCURRENT_REQUESTS = 4 # Configure a delay for requests for the same website (default: 0) # See http://scrapy.readthedocs.org/en/latest/topics/settings.html#download-delay # See also autothrottle settings and docs # 6. 延迟下载秒数 # DOWNLOAD_DELAY = 2 # The download delay setting will honor only one of: # 7. 单域名访问并发数，并且延迟下次秒数也应用在每个域名 # CONCURRENT_REQUESTS_PER_DOMAIN = 2 # 单IP访问并发数，如果有值则忽略：CONCURRENT_REQUESTS_PER_DOMAIN，并且延迟下次秒数也应用在每个IP # CONCURRENT_REQUESTS_PER_IP = 3 # Disable cookies (enabled by default) # 8. 是否支持cookie，cookiejar进行操作cookie # COOKIES_ENABLED = True # COOKIES_DEBUG = True # Disable Telnet Console (enabled by default) # 9. Telnet用于查看当前爬虫的信息，操作爬虫等... # 使用telnet ip port ，然后通过命令操作 # TELNETCONSOLE_ENABLED = True # TELNETCONSOLE_HOST = '127.0.0.1' # TELNETCONSOLE_PORT = [6023,] # Override the default request headers: # 10. 默认请求头 # DEFAULT_REQUEST_HEADERS = { # 'Accept': 'text/html,application/xhtml+xml,application/xml;q=0.9,*/*;q=0.8', # 'Accept-Language': 'en', # } # Configure item pipelines # See http://scrapy.readthedocs.org/en/latest/topics/item-pipeline.html # 11. 定义pipeline处理请求 # ITEM_PIPELINES = { # 'step8_king.pipelines.CustomPipeline': 500, # } # Enable or disable extensions # See http://scrapy.readthedocs.org/en/latest/topics/extensions.html # 12. 自定义扩展，基于信号进行调用 # EXTENSIONS = { # # 'step8_king.extensions.MyExtension': 500, # } # 13. 爬虫允许的最大深度，可以通过meta查看当前深度；0表示无深度 # DEPTH_LIMIT = 3 # 14. 爬取时，0表示深度优先Lifo(默认)；1表示广度优先FiFo # 后进先出，深度优先 # DEPTH_PRIORITY = 0 # SCHEDULER_DISK_QUEUE = 'scrapy.squeue.PickleLifoDiskQueue' # SCHEDULER_MEMORY_QUEUE = 'scrapy.squeue.LifoMemoryQueue' # 先进先出，广度优先 # DEPTH_PRIORITY = 1 # SCHEDULER_DISK_QUEUE = 'scrapy.squeue.PickleFifoDiskQueue' # SCHEDULER_MEMORY_QUEUE = 'scrapy.squeue.FifoMemoryQueue' # 15. 调度器队列 # SCHEDULER = 'scrapy.core.scheduler.Scheduler' # from scrapy.core.scheduler import Scheduler # 16. 访问URL去重 # DUPEFILTER_CLASS = 'step8_king.duplication.RepeatUrl' # Enable and configure the AutoThrottle extension (disabled by default) # See http://doc.scrapy.org/en/latest/topics/autothrottle.html # 开始自动限速 # AUTOTHROTTLE_ENABLED = True # The initial download delay # 初始下载延迟 # AUTOTHROTTLE_START_DELAY = 10 # The maximum download delay to be set in case of high latencies # 最大下载延迟 # AUTOTHROTTLE_MAX_DELAY = 60 # The average number of requests Scrapy should be sending in parallel to each remote server # 平均每秒并发数 # AUTOTHROTTLE_TARGET_CONCURRENCY = 1.0 # Enable showing throttling stats for every response received: # 是否显示 # AUTOTHROTTLE_DEBUG = True # Enable and configure HTTP caching (disabled by default) # See http://scrapy.readthedocs.org/en/latest/topics/downloader-middleware.html#httpcache-middleware-settings # HTTPCACHE_ENABLED = True # HTTPCACHE_EXPIRATION_SECS = 0 # HTTPCACHE_DIR = 'httpcache' # HTTPCACHE_IGNORE_HTTP_CODES = [] # HTTPCACHE_STORAGE = 'scrapy.extensions.httpcache.FilesystemCacheStorage' # Enable or disable spider middlewares # See http://scrapy.readthedocs.org/en/latest/topics/spider-middleware.html # 爬虫中间件 SPIDER_MIDDLEWARES = { 'step8_king.middlewares.MyCustomSpiderMiddleware': 543, } # Enable or disable downloader middlewares # See http://scrapy.readthedocs.org/en/latest/topics/downloader-middleware.html # 下载中间件 DOWNLOADER_MIDDLEWARES = { # 'step8_king.middlewares.MyCustomDownloaderMiddleware': 500, }
自定义pipline
一个简单的爬虫：
#!/usr/bin/env python # -*- coding:utf-8 -*- import scrapy from scrapy.http import Request from scrapy.selector import HtmlXPathSelector import re import urllib import os class XiaoHuarSpider(scrapy.spiders.Spider): name = "xiaohuar" allowed_domains = ["xiaohuar.com"] start_urls = [ "http://www.xiaohuar.com/list-1-1.html", ] def parse(self, response): # 分析页面 # 找到页面中符合规则的内容（校花图片），保存 # 找到所有的a标签，再访问其他a标签，一层一层的搞下去 hxs = HtmlXPathSelector(response) # 如果url是 http://www.xiaohuar.com/list-1-\d+.html if re.match('http://www.xiaohuar.com/list-1-\d+.html', response.url): items = hxs.select('//div[@class="item_list infinite_scroll"]/div') for i in range(len(items)): src = hxs.select('//div[@class="item_list infinite_scroll"]/div[%d]//div[@class="img"]/a/img/@src' % i).extract() name = hxs.select('//div[@class="item_list infinite_scroll"]/div[%d]//div[@class="img"]/span/text()' % i).extract() school = hxs.select('//div[@class="item_list infinite_scroll"]/div[%d]//div[@class="img"]/div[@class="btns"]/a/text()' % i).extract() if src: ab_src = "http://www.xiaohuar.com" + src[0] file_name = "%s_%s.jpg" % (school[0].encode('utf-8'), name[0].encode('utf-8')) file_path = os.path.join("/Users/wupeiqi/PycharmProjects/beauty/pic", file_name) urllib.urlretrieve(ab_src, file_path) # 获取所有的url，继续访问，并在其中寻找相同的url all_urls = hxs.select('//a/@href').extract() for url in all_urls: if url.startswith('http://www.xiaohuar.com/list-1-'): yield Request(url, callback=self.parse)
以上代码将符合规则的页面中的图片保存在指定目录，并且在HTML源码中找到所有的其他 a 标签的href属性，从而“递归”的执行下去，直到所有的页面都被访问过为止。以上代码之所以可以进行“递归”的访问相关URL，关键在于parse方法使用了 yield Request对象。
注：可以修改settings.py 中的配置文件，以此来指定“递归”的层数，如： DEPTH_LIMIT = 1
获取相应的cookie：
def parse(self, response): from scrapy.http.cookies import CookieJar cookieJar = CookieJar() cookieJar.extract_cookies(response, response.request) print(cookieJar._cookies)
格式化处理：
上述实例只是简单的图片处理，所以在parse方法中直接处理。如果对于想要获取更多的数据（获取页面的价格、商品名称、QQ等），则可以利用Scrapy的items将数据格式化，然后统一交由pipelines来处理。
import scrapy class JieYiCaiItem(scrapy.Item): company = scrapy.Field() title = scrapy.Field() qq = scrapy.Field() info = scrapy.Field() more = scrapy.Field()
上述定义模板，以后对于从请求的源码中获取的数据同意按照此结构来获取，所以在spider中需要有一下操作：
import scrapy import hashlib from beauty.items import JieYiCaiItem from scrapy.http import Request from scrapy.selector import HtmlXPathSelector from scrapy.spiders import CrawlSpider, Rule from scrapy.linkextractors import LinkExtractor class JieYiCaiSpider(scrapy.spiders.Spider): count = 0 url_set = set() name = "jieyicai" domain = 'http://www.jieyicai.com' allowed_domains = ["jieyicai.com"] start_urls = [ "http://www.jieyicai.com", ] rules = [ #下面是符合规则的网址,但是不抓取内容,只是提取该页的链接(这里网址是虚构的,实际使用时请替换) #Rule(SgmlLinkExtractor(allow=(r'http://test_url/test?page_index=\d+'))), #下面是符合规则的网址,提取内容,(这里网址是虚构的,实际使用时请替换) #Rule(LinkExtractor(allow=(r'http://www.jieyicai.com/Product/Detail.aspx?pid=\d+')), callback="parse"), ] def parse(self, response): md5_obj = hashlib.md5() md5_obj.update(response.url) md5_url = md5_obj.hexdigest() if md5_url in JieYiCaiSpider.url_set: pass else: JieYiCaiSpider.url_set.add(md5_url) hxs = HtmlXPathSelector(response) if response.url.startswith('http://www.jieyicai.com/Product/Detail.aspx'): item = JieYiCaiItem() item['company'] = hxs.select('//span[@class="username g-fs-14"]/text()').extract() item['qq'] = hxs.select('//span[@class="g-left bor1qq"]/a/@href').re('.*uin=(?P<qq>\d*)&') item['info'] = hxs.select('//div[@class="padd20 bor1 comard"]/text()').extract() item['more'] = hxs.select('//li[@class="style4"]/a/@href').extract() item['title'] = hxs.select('//div[@class="g-left prodetail-text"]/h2/text()').extract() yield item current_page_urls = hxs.select('//a/@href').extract() for i in range(len(current_page_urls)): url = current_page_urls[i] if url.startswith('/'): url_ab = JieYiCaiSpider.domain + url yield Request(url_ab, callback=self.parse)
此处代码的关键在于：
- 将获取的数据封装在了Item对象中
- yield Item对象 （一旦parse中执行yield Item对象，则自动将该对象交个pipelines的类来处理）
# -*- coding: utf-8 -*- # Define your item pipelines here # # Don't forget to add your pipeline to the ITEM_PIPELINES setting # See: http://doc.scrapy.org/en/latest/topics/item-pipeline.html import json from twisted.enterprise import adbapi import MySQLdb.cursors import re mobile_re = re.compile(r'(13[0-9]|15[012356789]|17[678]|18[0-9]|14[57])[0-9]{8}') phone_re = re.compile(r'(\d+-\d+|\d+)') class JsonPipeline(object): def __init__(self): self.file = open('/Users/wupeiqi/PycharmProjects/beauty/beauty/jieyicai.json', 'wb') def process_item(self, item, spider): line = "%s %s\n" % (item['company'][0].encode('utf-8'), item['title'][0].encode('utf-8')) self.file.write(line) return item class DBPipeline(object): def __init__(self): self.db_pool = adbapi.ConnectionPool('MySQLdb', db='DbCenter', user='root', passwd='123', cursorclass=MySQLdb.cursors.DictCursor, use_unicode=True) def process_item(self, item, spider): query = self.db_pool.runInteraction(self._conditional_insert, item) query.addErrback(self.handle_error) return item def _conditional_insert(self, tx, item): tx.execute("select nid from company where company = %s", (item['company'][0], )) result = tx.fetchone() if result: pass else: phone_obj = phone_re.search(item['info'][0].strip()) phone = phone_obj.group() if phone_obj else ' ' mobile_obj = mobile_re.search(item['info'][1].strip()) mobile = mobile_obj.group() if mobile_obj else ' ' values = ( item['company'][0], item['qq'][0], phone, mobile, item['info'][2].strip(), item['more'][0]) tx.execute("insert into company(company,qq,phone,mobile,address,more) values(%s,%s,%s,%s,%s,%s)", values) def handle_error(self, e): print 'error',e
上述中的pipelines中有多个类，到底Scapy会自动执行那个？哈哈哈哈，当然需要先配置了，不然Scapy就蒙逼了。。。
在settings.py中做如下配置：
ITEM_PIPELINES = { 'beauty.pipelines.DBPipeline': 300, 'beauty.pipelines.JsonPipeline': 100, } # 每行后面的整型值，确定了他们运行的顺序，item按数字从低到高的顺序，通过pipeline，通常将这些数字定义在0-1000范围内。
一个小蜘蛛：
import scrapy from scrapy.selector import HtmlXPathSelector from scrapy.http.request import Request from scrapy.http.cookies import CookieJar from scrapy import FormRequest class ChouTiSpider(scrapy.Spider): # 爬虫应用的名称，通过此名称启动爬虫命令 name = "chouti" # 允许的域名 allowed_domains = ["chouti.com"] cookie_dict = {} has_request_set = {} def start_requests(self): url = 'http://dig.chouti.com/' # return [Request(url=url, callback=self.login)] yield Request(url=url, callback=self.login) def login(self, response): cookie_jar = CookieJar() cookie_jar.extract_cookies(response, response.request) for k, v in cookie_jar._cookies.items(): for i, j in v.items(): for m, n in j.items(): self.cookie_dict[m] = n.value req = Request( url='http://dig.chouti.com/login', method='POST', headers={'Content-Type': 'application/x-www-form-urlencoded; charset=UTF-8'}, body='phone=8615131255089&password=pppppppp&oneMonth=1', cookies=self.cookie_dict, callback=self.check_login ) yield req def check_login(self, response): req = Request( url='http://dig.chouti.com/', method='GET', callback=self.show, cookies=self.cookie_dict, dont_filter=True ) yield req def show(self, response): # print(response) hxs = HtmlXPathSelector(response) news_list = hxs.select('//div[@id="content-list"]/div[@class="item"]') for new in news_list: # temp = new.xpath('div/div[@class="part2"]/@share-linkid').extract() link_id = new.xpath('*/div[@class="part2"]/@share-linkid').extract_first() yield Request( url='http://dig.chouti.com/link/vote?linksId=%s' %(link_id,), method='POST', cookies=self.cookie_dict, callback=self.do_favor ) page_list = hxs.select('//div[@id="dig_lcpage"]//a[re:test(@href, "/all/hot/recent/\d+")]/@href').extract() for page in page_list: page_url = 'http://dig.chouti.com%s' % page import hashlib hash = hashlib.md5() hash.update(bytes(page_url,encoding='utf-8')) key = hash.hexdigest() if key in self.has_request_set: pass else: self.has_request_set[key] = page_url yield Request( url=page_url, method='GET', callback=self.show ) def do_favor(self, response): print(response.text)
