Python 爬虫七 Scrapy

Scrapy

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

 

 

Scrapy主要包括了以下组件：

• 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：

1. 引擎从调度器中取出一个链接(URL)用于接下来的抓取
2. 引擎把URL封装成一个请求(Request)传给下载器
3. 下载器把资源下载下来，并封装成应答包(Response)
4. 爬虫解析Response
5. 解析出实体（Item）,则交给实体管道进行进一步的处理
6. 解析出的是链接（URL）,则把URL交给调度器等待抓取

 

一、安装 

Linux：

pip3 install scrapy

Windows:

a. pip3 install wheel
      b. 下载twisted http://www.lfd.uci.edu/~gohlke/pythonlibs/#twisted
      c. 进入下载目录，执行 pip3 install Twisted‑17.1.0‑cp35‑cp35m‑win_amd64.whl
      d. pip3 install scrapy
      e. 下载并安装pywin32：https://sourceforge.net/projects/pywin32/files/

 

二、基本使用

1、创建项目

scrapy startproject 项目名称
    - 在当前目录中创建中创建一个项目文件（类似于Django）

scrapy genspider [-t template] <name> <domain>
    - 穿件爬虫应用
    如：scrapy gensipider -t basic oldboy oldboy.com
          scrapy gensipider -t xmlfeed autohome autohome.com.cn

查看所有命令：scrapy gensipider -l
查看模板命令：scrapy gensipider -d 模板名称

scrapy list
    - 展示爬虫应用列表

scrapy crawl 爬虫应用名称
    - 运行单独爬虫应用

创建实例：

创建项目
shuais-MacBook-Pro:~ dandyzhang$ scrapy startproject scrapy_test

New Scrapy project 'scrapy_test', using template directory '/Library/Frameworks/Python.framework/Versions/3.6/lib/python3.6/site-packages/scrapy/templates/project', created in:
    /Users/dandyzhang/scrapy_test

You can start your first spider with:
    cd scrapy_test
    scrapy genspider example example.com


进入创建的项目
shuais-MacBook-Pro:~ dandyzhang$ cd scrapy_test/

创建爬虫应用1
shuais-MacBook-Pro:scrapy_test dandyzhang$ scrapy genspider chouti chouti.com

Created spider 'chouti' using template 'basic' in module:
  scrapy_test.spiders.chouti

创建爬虫应用2
shuais-MacBook-Pro:scrapy_test dandyzhang$ scrapy genspider cnblogs cnblogs.com
Created spider 'cnblogs' using template 'basic' in module:
  scrapy_test.spiders.cnblogs

 

2、项目结构以及爬虫应用简介

上面的实例，创建好了一个完整的项目：

文件说明：

• scrapy.cfg  项目的主配置信息。（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 

此时，发现之前根据命令创建了2个应用都存储在spiders文件夹内，现在以其中的chouti为例，来撰写第一个爬虫

import scrapy


class ChoutiSpider(scrapy.Spider):
    name = 'chouti'  # 外部scrapy调用的爬虫应用名称
    allowed_domains = ['chouti.com']  # 允许的域名
    start_urls = ['http://dig.chouti.com/']  # 起始url

    def parse(self, response):  # 访问起始url并获取结果后的回调函数
        print(response.text)  # response就是返回结果

查看结果：

 

如果是window用户可能会遇到编码问题：

import sys,os
sys.stdout=io.TextIOWrapper(sys.stdout.buffer,encoding='gb18030')

 

3、小试牛刀

如上，需要在抽屉网中抓去热榜的所有标题，图中的框已经标好，从content-list入手，抓取每一个item中class为part2的share-title

class ChoutiSpider(scrapy.Spider):
    name = 'chouti'
    allowed_domains = ['chouti.com']
    start_urls = ['http://dig.chouti.com/']

    def parse(self, response):
        """
        1.获取想要的内容
        2.如果分页，继续下载内容

        :param response:
        :return:
        """
        # 获取当前页的内容
        item_list = Selector(response=response).xpath('//div[@id="content-list"]/div[@class="item"]')
        # /子标签
        # //起始位置时，是在全局进行查找；非起始位置是在当前标签的子子孙孙内部找
        # ./当前对象下面找

        # 获取index为0的对象中的第一个满足条件的文本
        # obj = item_list[0].xpath('./div[@class="news-content"]//div[@class="part2"]/@share-title').extract_first()
        obj_list = item_list.xpath('./div[@class="news-content"]//div[@class="part2"]/@share-title').extract()
        print(obj_list)  # 获取的结果是列表

如果抓取的是标签的内容而不是属性的话：

obj = item_list[0].xpath('./div[@class="news-content"]//div[@class="show-content"]/text()').extract()

执行命令：

shuais-MacBook-Pro:scrapy_test dandyzhang$ scrapy crawl chouti --nolog

结果：

此时，如果分页内的也需要抓取呢？

首先，先获取以下分页内部的url：

import scrapy
from scrapy.selector import Selector, HtmlXPathSelector
from scrapy.http import Request


class ChoutiSpider(scrapy.Spider):
    name = 'chouti'
    allowed_domains = ['chouti.com']
    start_urls = ['http://dig.chouti.com/']

    def parse(self, response):
        """
        1.获取想要的内容
        2.如果分页，继续下载内容

        :param response:
        :return:
        """
        url_list = Selector(response=response).xpath('//div[@id="dig_lcpage"]//a/@href').extract()
        print(url_list)

运行结果：

shuais-MacBook-Pro:scrapy_test dandyzhang$ scrapy crawl chouti --nolog
['/all/hot/recent/2', '/all/hot/recent/3', '/all/hot/recent/4', '/all/hot/recent/5', '/all/hot/recent/6', '/all/hot/recent/7', '/all/hot/recent/8', '/all/hot/recent/9', '/all/hot/recent/10', '/all/hot/recent/2']

此时需要先拼接url，然后抓取数据：

# -*- coding: utf-8 -*-
import scrapy
from scrapy.selector import Selector, HtmlXPathSelector
from scrapy.http import Request  # 这里导入了一个Request，用来迭代


class ChoutiSpider(scrapy.Spider):
    name = 'chouti'
    allowed_domains = ['chouti.com']
    start_urls = ['http://dig.chouti.com/']

    def parse(self, response):
        """
        1.获取想要的内容
        2.如果分页，继续下载内容

        :param response:
        :return:
        """
        item_list = Selector(response=response).xpath('//div[@id="content-list"]/div[@class="item"]')

        obj_list = item_list.xpath('./div[@class="news-content"]//div[@class="part2"]/@share-title').extract()
        print(obj_list)

        url_list = Selector(response=response).xpath('//div[@id="dig_lcpage"]//a/@href').extract()
        for url in url_list:
            url = 'http://dig.chouti.com' + url
            yield Request(url=url)  # 迭代处理

这里可以在settings配置文件内设置下钻的深度：

DEPTH_LIMIT = 2

可以发现产生来了多个列表文件：

a、Request是一个封装用户请求的类，在回调函数中yield该对象表示继续访问

b、HtmlXpathSelector用于结构化HTML代码并提供选择器功能

 

 

4、选择器

from scrapy.selector import Selector, HtmlXPathSelector  # 一个是新版本，一个是旧版本，后面会被取消
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)  # 对象
# print(hxs)
# hxs = Selector(response=response).xpath('//a')  # 取全局内所有a标签
# print(hxs)
# hxs = Selector(response=response).xpath('//a[2]')  # 取全局内index为2的a标签
# print(hxs)
# hxs = Selector(response=response).xpath('//a[@id]')  # 取全局所有有id属性的a标签
# print(hxs)
# hxs = Selector(response=response).xpath('//a[@id="i1"]')  # 取全局所有id="i1"的a标签
# print(hxs)
# hxs = Selector(response=response).xpath('//a[@href="link.html"][@id="i1"]')  # 取全局所有href为link.html并且id为i1的a标签
# print(hxs)
# hxs = Selector(response=response).xpath('//a[contains(@href, "link")]')  # 取全局所有href有link字符串的a标签
# print(hxs)
# hxs = Selector(response=response).xpath('//a[starts-with(@href, "link")]')  # 取全局所有href以link字符串开头的a标签
# print(hxs)
# hxs = Selector(response=response).xpath('//a[re:test(@id, "i\d+")]')  # 正则 取全局所有a标签，id属性是i+数字的
# print(hxs)
# hxs = Selector(response=response).xpath('//a[re:test(@id, "i\d+")]/text()').extract()  # 正则 取全局所有a标签，id属性是i+数字的 内部的值
# print(hxs)
# hxs = Selector(response=response).xpath('//a[re:test(@id, "i\d+")]/@href').extract()  # 正则 取全局所有a标签，id属性是i+数字的 href属性值
# 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)

抽屉点赞：

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 = {}  # 发送过请求的集合

    def start_requests(self):  # 继承Spider，Spider内部先执行的是start_requests方法
        url = 'http://dig.chouti.com/'
        # return [Request(url=url, callback=self.login)]
        yield Request(url=url, callback=self.login)  # 爬取网页，指定回调函数；其实Request默认的callback是parse，
        # 这也解释了为什么新建的爬虫应用内部都是def parse(self, response):方法。可以像这样重写start_requests方法，指定callback

    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  # 指定回调函数
        )
        yield req

    def check_login(self, response):
        req = Request(
            url='http://dig.chouti.com/',
            method='GET',
            callback=self.show,  # 定义callback
            cookies=self.cookie_dict,
            dont_filter=True  # 不被去重过滤
        )
        yield req

    def show(self, response):
        # print(response)
        hxs = HtmlXPathSelector(response)  # 实例化标签对象
        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()  # 获取id
            yield Request(  # 点赞
                url='http://dig.chouti.com/link/vote?linksId=%s' %(link_id,),
                method='POST',
                cookies=self.cookie_dict,
                callback=self.do_favor
            )
        # 获取分页的网址
        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:  # 加密key请求在已请求的列表中，则pass
                pass
            else:  # 如果没有发送请求，继续发送
                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)  # 打印以下点赞之后的返回值

 

处理Cookie：

import scrapy
from scrapy.http.response.html import HtmlResponse
from scrapy.http import Request
from scrapy.http.cookies import CookieJar


class ChoutiSpider(scrapy.Spider):
    name = "chouti"
    allowed_domains = ["chouti.com"]
    start_urls = (
        'http://www.chouti.com/',
    )

    def start_requests(self):
        url = 'http://dig.chouti.com/'
        yield Request(url=url, callback=self.login, meta={'cookiejar': True})  # 如此设置cookiejar，可以自动获取cookie

    def login(self, response):
        print(response.headers.getlist('Set-Cookie'))
        req = Request(
            url='http://dig.chouti.com/login',
            method='POST',
            headers={'Content-Type': 'application/x-www-form-urlencoded; charset=UTF-8'},
            body='phone=8613121758648&password=woshiniba&oneMonth=1',
            callback=self.check_login,
            meta={'cookiejar': True}
        )
        yield req

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

注意：settings.py中设置DEPTH_LIMIT = 1来指定“递归”的层数。

 

这里对于上面的代码简单解释下，基础流程：

首先最初创建的爬虫应用的源码：继承了Spider类，该类内部有一个start_requests方法，这是爬虫执行的起始函数，如果start_urls不为空，爬取此url。即上图的yield Request(url, dont_filter=True)、这也可以解释为什么继续爬取分页的url时，写的是yield  Request(url)。此时大家也许不明白，即start_urls不为空，为什么会执行parse函数呢？其实在开始执行的yield Request中有一个默认参数是callback=parse，所以初始化的爬虫应用的流程就一目了然了。

现在解释下点赞的爬虫，前面提到继承了Spider类，第一个执行的是start_requests，此时既然继承了父类Spider，就可以对此类进行重写，已经知道了其实位置是start_requests，毫无疑问重写此方法，内部指定url（外部的start_urls删除），执行爬虫则调用Request方法，指定callback函数，这样根据callback也就形成了一个串行爬虫链。另外要提到的一点yield都知道是一个生成器，在Scrapy内部，spider内部调度yield Request只是其中的一部分，用来爬虫。另外一部分也是通过yield调用来做持久化的，即对于爬取的数据的处理跟保存。下面会讲到这部分，这里先提一下。

 

5、格式化处理

之前的实例只是一些简单的处理，所以在parse方法中直接处理。如果想要获取更多的数据处理，则可以利用Scrapy的items将数据格式化，然后统一交由pipelines来处理。

回到最原始的parse代码，抓取以下热榜标题跟链接

chouti.py

import scrapy
from scrapy.selector import HtmlXPathSelector, Selector
from ..items import ScrapyTestItem

class ChoutiSpider(scrapy.Spider):
    name = 'chouti'
    allowed_domains = ['chouti.com']
    start_urls = ['http://dig.chouti.com/']

    def parse(self, response):
        item_list = Selector(response=response).xpath('//div[@id="content-list"]/div[@class="item"]')

        for item in item_list:
            t = item.xpath('./div[@class="news-content"]//div[@class="part1"]/a/text()').extract()
            h = item.xpath('./div[@class="news-content"]//div[@class="part1"]/a/@href').extract()
            item_obj = ScrapyTestItem(title=t, href=h)   # 调用Item
            yield item_obj  # 这里指向了另一个调度器，持久化调度器

items.py

import scrapy


class ScrapyTestItem(scrapy.Item):
    # define the fields for your item here like: 定义要抓取保存的字段
    title = scrapy.Field()
    href = scrapy.Field()

pipelines.py

class ScrapyTestPipeline(object):
    def process_item(self, item, spider):
        print(item, spider)
        return item

这里需要注意的是跟Django一样，需要注册以下：

在settings文件里面找到下面这段话，去掉注释，其中300代表优先级，稍后进行这个数字的测试。

ITEM_PIPELINES = {
   'scrapy_test.pipelines.ScrapyTestPipeline': 300,
}

此时执行爬虫：

语法没写好，抓到2个href了，不要在意这些细节。

此时，了解了yield的另一个功能，当yield Item_obj是就会调度pipelines进行持久化，当然上面我们只是打印了以下结果，可以看到item对应的是字段，spider是爬虫应用函数方法。

所以对于不同的要求可以直接在pipelines里面写到：

class ScrapyTestPipeline1(object):
    def process_item(self, item, spider):
        print('step 1 输出到屏幕')
        return item


class ScrapyTestPipeline2(object):
    def process_item(self, item, spider):
        print('step 2 保存到文件')
        return item


class ScrapyTestPipeline3(object):
    def process_item(self, item, spider):
        print('step 3 保存到数据库')
        return item

注册以下：

ITEM_PIPELINES = {
    'scrapy_test.pipelines.ScrapyTestPipeline1': 100,
    'scrapy_test.pipelines.ScrapyTestPipeline2': 200,
    'scrapy_test.pipelines.ScrapyTestPipeline3': 300,
}

执行结果、注意顺序

假设step3的类没有注册，就只会执行step1 & step2。

那么、如果想在执行到某一个pipeline类终止怎么办？

from scrapy.exceptions import DropItem  # 导入DropItem

class ScrapyTestPipeline1(object):
    def process_item(self, item, spider):
        print('step 1 输出到屏幕')
        raise DropItem()


class ScrapyTestPipeline2(object):
    def process_item(self, item, spider):
        print('step 2 保存到文件')
        return item


class ScrapyTestPipeline3(object):
    def process_item(self, item, spider):
        print('step 3 保存到数据库')
        return item

那spider参数是干嘛用的呢？

假设，抓取的name是chouti的时候，不让其继续执行后续的：

from scrapy.exceptions import DropItem


class ScrapyTestPipeline1(object):
    def process_item(self, item, spider):
        print('step 1 输出到屏幕')
        if spider.name == 'chouti':
            raise DropItem()
        return item

class ScrapyTestPipeline2(object):
    def process_item(self, item, spider):
        print('step 2 保存到文件')
        return item


class ScrapyTestPipeline3(object):
    def process_item(self, item, spider):
        print('step 3 保存到数据库')
        return item

pipelines更多：

假设需要将数据写入文件，首先想到的方法一定是

class ScrapyTestPipeline(object):
    def process_item(self, item, spider):
        with open('***', 'a+') as f:
            f.write('***')
        print('step 2 保存到文件')
        return item

但是这样会在一次爬虫中频繁的打开文件，浪费IO

此时引入另外的方法

from scrapy.exceptions import DropItem

class CustomPipeline(object):
    def __init__(self,v):  # v就是类方法返回的参数val
        self.value = v
        print(self.value)

    def process_item(self, item, spider):
        # 操作并进行持久化

        # return表示会被后续的pipeline继续处理
        print('****操作****')
        return item

        # 表示将item丢弃，不会被后续pipeline处理
        # raise DropItem()


    @classmethod
    def from_crawler(cls, crawler):
        """
        初始化时候，用于创建pipeline对象
        :param crawler: 
        :return: 
        """
        val = crawler.settings.get('MYPATH')  # 类方法获取配置文件参数
        print(val)
        return cls(val)

    def open_spider(self,spider):
        """
        爬虫开始执行时，调用
        :param spider: 
        :return: 
        """
        print('000000')

    def close_spider(self,spider):
        """
        爬虫关闭时，被调用
        :param spider: 
        :return: 
        """
        print('111111')

此时在setting中配置以下文件地址就可以了：

MYPATH = '***path***'

settings参数必须全部大写，小写测试失败，未抓取到。

执行结果

所以以后，可以在from_crawler里面通过参数定义文件名，setting文件设置文件路径，然后打开文件，中间对文件句柄进行追加，一次打开，一次关闭，避免重复操作。

 

 

6、中间件

自动化里面Django blog其实已经讲过了中间件的一个大致流程，其实在scrapy中中间件的核心依然是同样的。

上图是Django中中间件的一个基本概念图，而在scrapy中则是：

爬虫中间件

class SpiderMiddleware(object):

    def process_spider_input(self,response, spider):
        """
        下载完成，执行，然后交给parse处理（默认有start_urls时，parse时默认的callback函数）
        :param response: 
        :param spider: 
        :return: 
        """
        pass

    def process_spider_output(self,response, result, spider):
        """
        spider处理完成，返回时调用
        :param response:
        :param result:
        :param spider:
        :return: 必须返回包含 Request 或 Item 对象的可迭代对象(iterable)
        """
        return result

    def process_spider_exception(self,response, exception, spider):
        """
        异常调用
        :param response:
        :param exception:
        :param spider:
        :return: None,继续交给后续中间件处理异常；含 Response 或 Item 的可迭代对象(iterable)，交给调度器或pipeline
        """
        return None


    def process_start_requests(self,start_requests, spider):
        """
        爬虫启动时调用
        :param start_requests:
        :param spider:
        :return: 包含 Request 对象的可迭代对象、丢给调度器分配下载或者是解析文本
        """
        return start_requests

首先爬虫引擎启动全局，到spider的start_urls抓取数据返回start_request，放到任务调度器里面,下载器去任务调度器抓取任务执行。

 

 

下载器中间件

class DownMiddleware1(object):
    def process_request(self, request, spider):
        """
        请求需要被下载时，经过所有下载器中间件的process_request调用
        :param request: 
        :param spider: 
        :return:  
            None,继续后续中间件去下载；
            Response对象，停止process_request的执行，开始执行process_response
            Request对象，停止中间件的执行，将Request重新调度器
            raise IgnoreRequest异常，停止process_request的执行，开始执行process_exception
        """
        pass



    def process_response(self, request, response, spider):
        """
        spider处理完成，返回时调用
        :param response:
        :param result:
        :param spider:
        :return: 
            Response 对象：转交给其他中间件process_response
            Request 对象：停止中间件，request会被重新调度下载
            raise IgnoreRequest 异常：调用Request.errback
        """
        print('response1')
        return response

    def process_exception(self, request, exception, spider):
        """
        当下载处理器(download handler)或 process_request() (下载中间件)抛出异常
        :param response:
        :param exception:
        :param spider:
        :return: 
            None：继续交给后续中间件处理异常；
            Response对象：停止后续process_exception方法
            Request对象：停止中间件，request将会被重新调用下载
        """
        return None

 

7、自定制命令

a、在spiders同级创建任意目录，如：commands

b、在其中创建 crawlall.py 文件 （此处文件名就是自定义的命令）-- 如果创建多个文件，其实是相当于创建了多个命令

from scrapy.commands import ScrapyCommand
from scrapy.utils.project import get_project_settings


    class Command(ScrapyCommand):

        requires_project = True

        def syntax(self):
            return '[options]'

        def short_desc(self):
            return 'Runs all of the spiders'

        def run(self, args, opts):
            spider_list = self.crawler_process.spiders.list()  # 去spiders文件夹下获取所有的爬虫文件
            for name in spider_list:
                self.crawler_process.crawl(name, **opts.__dict__)  # 为所有的爬虫创建任务
            self.crawler_process.start()  # 并发的开始执行

crawlall.py

c、在settings.py 中添加配置 COMMANDS_MODULE = '项目名称.目录名称'

d、在项目目录执行命令：scrapy crawlall 

PS：scrapy的源码，建议从run开始着手看。

单个爬虫：

import sys
from scrapy.cmdline import execute

if __name__ == '__main__':
    execute(["scrapy","github","--nolog"])

View Code

 

8、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（跟Django的信号很相似）

from scrapy import signals


class MyExtension(object):
    def __init__(self, value):
        self.value = value

    @classmethod
    def from_crawler(cls, crawler):
        val = crawler.settings.get('MMMM')
        ext = cls(val)

        crawler.signals.connect(ext.openn, signal=signals.spider_opened)
        crawler.signals.connect(ext.closee, signal=signals.spider_closed)

        return ext

    def openn(self, spider):
        print('open')

    def closee(self, spider):
        print('close')

"""
Scrapy signals

These signals are documented in docs/topics/signals.rst. Please don't add new
signals here without documenting them there.
"""

engine_started = object()
engine_stopped = object()
spider_opened = object()
spider_idle = object()
spider_closed = object()
spider_error = object()
request_scheduled = object()
request_dropped = object()
response_received = object()
response_downloaded = object()
item_scraped = object()
item_dropped = object()

# for backwards compatibility
stats_spider_opened = spider_opened
stats_spider_closing = spider_closed
stats_spider_closed = spider_closed

item_passed = item_scraped

request_received = request_scheduled

爬虫的所有信号

跟pipelines一样，需要注册类在settings文件里。

 

 

 

9、避免重复访问

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

class RepeatUrl:
    def __init__(self):
        self.visited_url = set()

    @classmethod
    def from_settings(cls, settings):
        """
        初始化时，调用
        :param settings: 
        :return: 
        """
        return cls()

    def request_seen(self, request):
        """
        检测当前请求是否已经被访问过
        :param request: 
        :return: True表示已经访问过；False表示未访问过
        """
        if request.url in self.visited_url:
            return True
        self.visited_url.add(request.url)
        return False

    def open(self):
        """
        开始爬去请求时，调用
        :return: 
        """
        print('open replication')

    def close(self, reason):
        """
        结束爬虫爬取时，调用
        :param reason: 
        :return: 
        """
        print('close replication')

    def log(self, request, spider):
        """
        记录日志
        :param request: 
        :param spider: 
        :return: 
        """
        print('repeat', request.url)

自定义URL去重操作

自定义url去重

 

 

10、settings其他设置

# -*- coding: utf-8 -*-

# Scrapy settings for step8_king project
#
# For simplicity, this file contains only settings considered important or
# commonly used. You can find more settings consulting the documentation:
#
#     http://doc.scrapy.org/en/latest/topics/settings.html
#     http://scrapy.readthedocs.org/en/latest/topics/downloader-middleware.html
#     http://scrapy.readthedocs.org/en/latest/topics/spider-middleware.html

# 1. 爬虫名称
BOT_NAME = 'scrapy_test'

# 2. 爬虫应用路径
SPIDER_MODULES = ['scrapy_test.spiders']
NEWSPIDER_MODULE = 'scrapy_test.spiders'

# Crawl responsibly by identifying yourself (and your website) on the user-agent
# 3. 客户端 user-agent请求头 通用配置，也可以在Request内部配置
# USER_AGENT = 'scrapy_test (+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（反爬虫，所有的爬虫都是延迟2秒）
# 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,]


# 10. 默认请求头
# Override the default request headers:
# 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 = {
#    'scrapy_test.pipelines.JsonPipeline': 700,
#    'scrapy_test.pipelines.FilePipeline': 500,
# }



# 12. 自定义扩展，基于信号进行调用
# Enable or disable extensions
# See http://scrapy.readthedocs.org/en/latest/topics/extensions.html
# EXTENSIONS = {
#     # 'scrapy_test.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  queue
# SCHEDULER = 'scrapy.core.scheduler.Scheduler'
# from scrapy.core.scheduler import Scheduler


# 16. 访问URL去重
# DUPEFILTER_CLASS = 'scrapy_test.duplication.RepeatUrl'


# Enable and configure the AutoThrottle extension (disabled by default)
# See http://doc.scrapy.org/en/latest/topics/autothrottle.html

"""
17. 自动限速算法
    from scrapy.contrib.throttle import AutoThrottle
    自动限速设置
    1. 获取最小延迟 DOWNLOAD_DELAY
    2. 获取最大延迟 AUTOTHROTTLE_MAX_DELAY
    3. 设置初始下载延迟 AUTOTHROTTLE_START_DELAY
    4. 当请求下载完成后，获取其"连接"时间 latency，即：请求连接到接受到响应头之间的时间
    5. 用于计算的... AUTOTHROTTLE_TARGET_CONCURRENCY
    target_delay = latency / self.target_concurrency
    new_delay = (slot.delay + target_delay) / 2.0 # 表示上一次的延迟时间
    new_delay = max(target_delay, new_delay)
    new_delay = min(max(self.mindelay, new_delay), self.maxdelay)
    slot.delay = new_delay
"""

# 开始自动限速
# AUTOTHROTTLE_ENABLED = True
# The initial download delay
# 初始下载延迟
# AUTOTHROTTLE_START_DELAY = 5
# The maximum download delay to be set in case of high latencies
# 最大下载延迟
# AUTOTHROTTLE_MAX_DELAY = 10
# 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


"""
18. 启用缓存
    目的用于将已经发送的请求或相应缓存下来，以便以后使用
    
    from scrapy.downloadermiddlewares.httpcache import HttpCacheMiddleware
    from scrapy.extensions.httpcache import DummyPolicy
    from scrapy.extensions.httpcache import FilesystemCacheStorage
"""
# 是否启用缓存策略
# HTTPCACHE_ENABLED = True

# 缓存策略：所有请求均缓存，下次在请求直接访问原来的缓存即可
# HTTPCACHE_POLICY = "scrapy.extensions.httpcache.DummyPolicy"
# 缓存策略：根据Http响应头：Cache-Control、Last-Modified 等进行缓存的策略
# HTTPCACHE_POLICY = "scrapy.extensions.httpcache.RFC2616Policy"

# 缓存超时时间
# HTTPCACHE_EXPIRATION_SECS = 0

# 缓存保存路径
# HTTPCACHE_DIR = 'httpcache'

# 缓存忽略的Http状态码
# HTTPCACHE_IGNORE_HTTP_CODES = []

# 缓存存储的插件
# HTTPCACHE_STORAGE = 'scrapy.extensions.httpcache.FilesystemCacheStorage'


"""
19. 代理，需要在环境变量中设置
    from scrapy.contrib.downloadermiddleware.httpproxy import HttpProxyMiddleware
    
    方式一：使用默认，key不可以修改
        os.environ
        {
            http_proxy:http://root:woshiniba@192.168.11.11:9999/
            https_proxy:http://192.168.11.11:9999/
        }
    方式二：使用自定义下载中间件
    
    def to_bytes(text, encoding=None, errors='strict'):
        if isinstance(text, bytes):
            return text
        if not isinstance(text, six.string_types):
            raise TypeError('to_bytes must receive a unicode, str or bytes '
                            'object, got %s' % type(text).__name__)
        if encoding is None:
            encoding = 'utf-8'
        return text.encode(encoding, errors)
        
    class ProxyMiddleware(object):
        def process_request(self, request, spider):
            PROXIES = [
                {'ip_port': '111.11.228.75:80', 'user_pass': ''},
                {'ip_port': '120.198.243.22:80', 'user_pass': ''},
                {'ip_port': '111.8.60.9:8123', 'user_pass': ''},
                {'ip_port': '101.71.27.120:80', 'user_pass': ''},
                {'ip_port': '122.96.59.104:80', 'user_pass': ''},
                {'ip_port': '122.224.249.122:8088', 'user_pass': ''},
            ]
            proxy = random.choice(PROXIES)
            if proxy['user_pass'] is not None:
                request.meta['proxy'] = to_bytes（"http://%s" % proxy['ip_port']）
                encoded_user_pass = base64.encodestring(to_bytes(proxy['user_pass']))
                request.headers['Proxy-Authorization'] = to_bytes('Basic ' + encoded_user_pass)
                print "**************ProxyMiddleware have pass************" + proxy['ip_port']
            else:
                print "**************ProxyMiddleware no pass************" + proxy['ip_port']
                request.meta['proxy'] = to_bytes("http://%s" % proxy['ip_port'])
    
    DOWNLOADER_MIDDLEWARES = {
       'step8_king.middlewares.ProxyMiddleware': 500,
    }
    
"""

"""
20. Https访问
    Https访问时有两种情况：
    1. 要爬取网站使用的可信任证书(默认支持)
        DOWNLOADER_HTTPCLIENTFACTORY = "scrapy.core.downloader.webclient.ScrapyHTTPClientFactory"
        DOWNLOADER_CLIENTCONTEXTFACTORY = "scrapy.core.downloader.contextfactory.ScrapyClientContextFactory"
        
    2. 要爬取网站使用的自定义证书
        DOWNLOADER_HTTPCLIENTFACTORY = "scrapy.core.downloader.webclient.ScrapyHTTPClientFactory"
        DOWNLOADER_CLIENTCONTEXTFACTORY = "scrapy_test.https.MySSLFactory"
        
        # https.py
        from scrapy.core.downloader.contextfactory import ScrapyClientContextFactory
        from twisted.internet.ssl import (optionsForClientTLS, CertificateOptions, PrivateCertificate)
        
        class MySSLFactory(ScrapyClientContextFactory):
            def getCertificateOptions(self):
                from OpenSSL import crypto
                v1 = crypto.load_privatekey(crypto.FILETYPE_PEM, open('/Users/wupeiqi/client.key.unsecure', mode='r').read())
                v2 = crypto.load_certificate(crypto.FILETYPE_PEM, open('/Users/wupeiqi/client.pem', mode='r').read())
                return CertificateOptions(
                    privateKey=v1,  # pKey对象
                    certificate=v2,  # X509对象
                    verify=False,
                    method=getattr(self, 'method', getattr(self, '_ssl_method', None))
                )
    其他：
        相关类
            scrapy.core.downloader.handlers.http.HttpDownloadHandler
            scrapy.core.downloader.webclient.ScrapyHTTPClientFactory
            scrapy.core.downloader.contextfactory.ScrapyClientContextFactory
        相关配置
            DOWNLOADER_HTTPCLIENTFACTORY
            DOWNLOADER_CLIENTCONTEXTFACTORY

"""



"""
21. 爬虫中间件
    class SpiderMiddleware(object):

        def process_spider_input(self,response, spider):
            '''
            下载完成，执行，然后交给parse处理
            :param response: 
            :param spider: 
            :return: 
            '''
            pass
    
        def process_spider_output(self,response, result, spider):
            '''
            spider处理完成，返回时调用
            :param response:
            :param result:
            :param spider:
            :return: 必须返回包含 Request 或 Item 对象的可迭代对象(iterable)
            '''
            return result
    
        def process_spider_exception(self,response, exception, spider):
            '''
            异常调用
            :param response:
            :param exception:
            :param spider:
            :return: None,继续交给后续中间件处理异常；含 Response 或 Item 的可迭代对象(iterable)，交给调度器或pipeline
            '''
            return None
    
    
        def process_start_requests(self,start_requests, spider):
            '''
            爬虫启动时调用
            :param start_requests:
            :param spider:
            :return: 包含 Request 对象的可迭代对象
            '''
            return start_requests
    
    内置爬虫中间件：
        'scrapy.contrib.spidermiddleware.httperror.HttpErrorMiddleware': 50,
        'scrapy.contrib.spidermiddleware.offsite.OffsiteMiddleware': 500,
        'scrapy.contrib.spidermiddleware.referer.RefererMiddleware': 700,
        'scrapy.contrib.spidermiddleware.urllength.UrlLengthMiddleware': 800,
        'scrapy.contrib.spidermiddleware.depth.DepthMiddleware': 900,

"""
# from scrapy.contrib.spidermiddleware.referer import RefererMiddleware
# Enable or disable spider middlewares
# See http://scrapy.readthedocs.org/en/latest/topics/spider-middleware.html
SPIDER_MIDDLEWARES = {
   # 'scrapy_test.middlewares.SpiderMiddleware': 543,
}


"""
22. 下载中间件
    class DownMiddleware1(object):
        def process_request(self, request, spider):
            '''
            请求需要被下载时，经过所有下载器中间件的process_request调用
            :param request:
            :param spider:
            :return:
                None,继续后续中间件去下载；
                Response对象，停止process_request的执行，开始执行process_response
                Request对象，停止中间件的执行，将Request重新调度器
                raise IgnoreRequest异常，停止process_request的执行，开始执行process_exception
            '''
            pass
    
    
    
        def process_response(self, request, response, spider):
            '''
            spider处理完成，返回时调用
            :param response:
            :param result:
            :param spider:
            :return:
                Response 对象：转交给其他中间件process_response
                Request 对象：停止中间件，request会被重新调度下载
                raise IgnoreRequest 异常：调用Request.errback
            '''
            print('response1')
            return response
    
        def process_exception(self, request, exception, spider):
            '''
            当下载处理器(download handler)或 process_request() (下载中间件)抛出异常
            :param response:
            :param exception:
            :param spider:
            :return:
                None：继续交给后续中间件处理异常；
                Response对象：停止后续process_exception方法
                Request对象：停止中间件，request将会被重新调用下载
            '''
            return None

    
    默认下载中间件
    {
        'scrapy.contrib.downloadermiddleware.robotstxt.RobotsTxtMiddleware': 100,
        'scrapy.contrib.downloadermiddleware.httpauth.HttpAuthMiddleware': 300,
        'scrapy.contrib.downloadermiddleware.downloadtimeout.DownloadTimeoutMiddleware': 350,
        'scrapy.contrib.downloadermiddleware.useragent.UserAgentMiddleware': 400,
        'scrapy.contrib.downloadermiddleware.retry.RetryMiddleware': 500,
        'scrapy.contrib.downloadermiddleware.defaultheaders.DefaultHeadersMiddleware': 550,
        'scrapy.contrib.downloadermiddleware.redirect.MetaRefreshMiddleware': 580,
        'scrapy.contrib.downloadermiddleware.httpcompression.HttpCompressionMiddleware': 590,
        'scrapy.contrib.downloadermiddleware.redirect.RedirectMiddleware': 600,
        'scrapy.contrib.downloadermiddleware.cookies.CookiesMiddleware': 700,
        'scrapy.contrib.downloadermiddleware.httpproxy.HttpProxyMiddleware': 750,
        'scrapy.contrib.downloadermiddleware.chunked.ChunkedTransferMiddleware': 830,
        'scrapy.contrib.downloadermiddleware.stats.DownloaderStats': 850,
        'scrapy.contrib.downloadermiddleware.httpcache.HttpCacheMiddleware': 900,
    }

"""
# from scrapy.contrib.downloadermiddleware.httpauth import HttpAuthMiddleware
# Enable or disable downloader middlewares
# See http://scrapy.readthedocs.org/en/latest/topics/downloader-middleware.html
# DOWNLOADER_MIDDLEWARES = {
#    'scrapy_test.middlewares.DownMiddleware1': 100,
#    'scrapy_test.middlewares.DownMiddleware2': 500,
# }

settings.py

 

11、模拟scrapy框架

#!/usr/bin/env python
# -*- coding:utf-8 -*-
from twisted.web.client import getPage, defer
from twisted.internet import reactor
import queue


class Response(object):
    def __init__(self, body, request):
        self.body = body
        self.request = request
        self.url = request.url

    @property
    def text(self):
        return self.body.decode('utf-8')


class Request(object):
    def __init__(self, url, callback=None):
        self.url = url
        self.callback = callback


class Scheduler(object):
    def __init__(self, engine):
        self.q = queue.Queue()
        self.engine = engine

    def enqueue_request(self, request):
        self.q.put(request)

    def next_request(self):
        try:
            req = self.q.get(block=False)
        except Exception as e:
            req = None

        return req

    def size(self):
        return self.q.qsize()


class ExecutionEngine(object):
    def __init__(self):
        self._closewait = None
        self.running = True
        self.start_requests = None
        self.scheduler = Scheduler(self)

        self.inprogress = set()

    def check_empty(self, response):
        if not self.running:
            self._closewait.callback('......')

    def _next_request(self):
        while self.start_requests:
            try:
                request = next(self.start_requests)
            except StopIteration:
                self.start_requests = None
            else:
                self.scheduler.enqueue_request(request)

        while len(self.inprogress) < 5 and self.scheduler.size() > 0:  # 最大并发数为5

            request = self.scheduler.next_request()
            if not request:
                break

            self.inprogress.add(request)
            d = getPage(bytes(request.url, encoding='utf-8'))
            d.addBoth(self._handle_downloader_output, request)
            d.addBoth(lambda x, req: self.inprogress.remove(req), request)
            d.addBoth(lambda x: self._next_request())

        if len(self.inprogress) == 0 and self.scheduler.size() == 0:
            self._closewait.callback(None)

    def _handle_downloader_output(self, body, request):
        """
        获取内容，执行回调函数，并且把回调函数中的返回值获取，并添加到队列中
        :param response: 
        :param request: 
        :return: 
        """
        import types

        response = Response(body, request)
        func = request.callback or self.spider.parse
        gen = func(response)
        if isinstance(gen, types.GeneratorType):
            for req in gen:
                self.scheduler.enqueue_request(req)

    @defer.inlineCallbacks
    def start(self):
        self._closewait = defer.Deferred()
        yield self._closewait

    def open_spider(self, spider, start_requests):
        self.start_requests = start_requests
        self.spider = spider
        reactor.callLater(0, self._next_request)


class Crawler(object):
    def __init__(self, spidercls):
        self.spidercls = spidercls

        self.spider = None
        self.engine = None

    @defer.inlineCallbacks
    def crawl(self):
        self.engine = ExecutionEngine()
        self.spider = self.spidercls()
        start_requests = iter(self.spider.start_requests())
        start_requests = iter(start_requests)
        self.engine.open_spider(self.spider, start_requests)
        yield self.engine.start()


class CrawlerProcess(object):
    def __init__(self):
        self._active = set()
        self.crawlers = set()

    def crawl(self, spidercls, *args, **kwargs):
        crawler = Crawler(spidercls)

        self.crawlers.add(crawler)
        d = crawler.crawl(*args, **kwargs)
        self._active.add(d)
        return d

    def start(self):
        dl = defer.DeferredList(self._active)
        dl.addBoth(self._stop_reactor)
        reactor.run()

    def _stop_reactor(self, _=None):
        reactor.stop()


class Spider(object):
    def start_requests(self):
        for url in self.start_urls:
            yield Request(url)


class ChoutiSpider(Spider):
    name = "chouti"
    start_urls = [
        'http://dig.chouti.com/',
    ]

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


class CnblogsSpider(Spider):
    name = "cnblogs"
    start_urls = [
        'http://www.cnblogs.com/',
    ]

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


if __name__ == '__main__':

    spider_cls_list = [ChoutiSpider, CnblogsSpider]

    crawler_process = CrawlerProcess()
    for spider_cls in spider_cls_list:
        crawler_process.crawl(spider_cls)

    crawler_process.start()

模拟scrapy框架

 

参见文档：http://www.cnblogs.com/wupeiqi/articles/6229292.html

 

来源：oschina

链接：https://my.oschina.net/u/4323462/blog/3887161