Python crawler programming framework Scrapy introductory learning tutorial

1. Introduction to Scrapy
Scrapy is an application framework written to crawl website data and extract structured data. It can be used in a series of programs including data mining, information processing or storing historical data.
It was originally designed for page scraping (more specifically, web scraping), but can also be used to obtain data returned by APIs (such as Amazon Associates Web Services) or general web crawlers. Scrapy is widely used and can be used for data mining, monitoring and automated testing
Scrapy uses the Twisted asynchronous network library to handle network communication. The overall structure is roughly as follows

Scrapy mainly includes the following components:

(1) Engine (Scrapy): used to handle data flow processing of the entire system and trigger transactions (core framework)

(2) Scheduler (Scheduler): Used to accept requests from the engine, push them into the queue, and return when the engine requests again. It can be imagined as a URL (the URL or link of the web page that is captured ), which determines the next URL to be crawled and removes duplicate URLs

(3) Downloader: used to download web content and return the web content to the spider (Scrapy downloader is built on twisted, an efficient asynchronous model)

(4) Spiders: Crawlers are mainly used to extract the information they need from specific web pages, which are so-called entities (Items). Users can also extract links from it and let Scrapy continue to crawl the next page

Project Pipeline: Responsible for processing entities extracted from web pages by crawlers. The main function is to persist entities, verify the validity of entities, and remove unnecessary information. When the page is parsed by the crawler, it will be sent to the project pipeline and the data will be processed through several specific sequences.

(5) Downloader Middlewares: A framework located between the Scrapy engine and the downloader. It mainly handles requests and responses between the Scrapy engine and the downloader.

(6) Spider Middlewares: A framework between the Scrapy engine and the crawler. Its main job is to process the spider's response input and request output.

(7) Scheduler Middewares: The middleware between the Scrapy engine and the scheduler, sending requests and responses from the Scrapy engine to the scheduler.

The Scrapy operation process is roughly as follows:

First, the engine takes out a link (URL) from the scheduler for subsequent crawling
The engine encapsulates the URL into a request (Request) and passes it to the downloader. The downloader downloads the resource and encapsulates it into a response package (Response)
Then, the crawler parses Response
If the entity (Item) is parsed, it is handed over to the entity pipeline for further processing.
If the parsed result is a link (URL), the URL will be handed over to the Scheduler to wait for crawling

2. Install Scrapy
Use the following command:

sudo pip install virtualenv #安装虚拟环境工具
virtualenv ENV #创建一个虚拟环境目录
source ./ENV/bin/active #激活虚拟环境
pip install Scrapy
#验证是否安装成功
pip list

#输出如下
cffi (0.8.6)
cryptography (0.6.1)
cssselect (0.9.1)
lxml (3.4.1)
pip (1.5.6)
pycparser (2.10)
pyOpenSSL (0.14)
queuelib (1.2.2)
Scrapy (0.24.4)
setuptools (3.6)
six (1.8.0)
Twisted (14.0.2)
w3lib (1.10.0)
wsgiref (0.1.2)
zope.interface (4.1.1)

For more virtual environment operations, please check my blog post

3. Scrapy Tutorial
Before scraping, you need to create a new Scrapy project. Enter a directory where you want to save the code, and then execute:

$ scrapy startproject tutorial

This command will create a new directory tutorial in the current directory, with the following structure:

.
├── scrapy.cfg
└── tutorial
 ├── __init__.py
 ├── items.py
 ├── pipelines.py
 ├── settings.py
 └── spiders
  └── __init__.py

These files are mainly:

(1) scrapy.cfg: project configuration file
(2) tutorial/: project python module, you will add code here later
(3) tutorial/items.py: project items file
(4) tutorial/pipelines.py: project pipeline file
(5) tutorial/settings.py: project configuration file
(6) tutorial/spiders: directory where spiders are placed

3.1. Define Item
Items is a container that will be loaded with scraped data. It works like a dictionary in python, but it provides more protection, such as padding undefined fields to prevent spelling errors

Declare an Item by creating a scrapy.Item class and defining a class attribute of type scrapy.Field.
We control the site data obtained from dmoz.org by modeling the required items. For example, if we want to obtain the site name, URL and site description, we define the fields of these three attributes. Edit the items.py file in the tutorial directory

from scrapy.item import Item, Field


class DmozItem(Item):
 # define the fields for your item here like:
 name = Field()
 description = Field()
 url = Field()

3.2. Writing Spider
Spider is a user-written class used to crawl information from a domain (or domain group), defining a preliminary list of URLs for downloading, how to follow the links, and how to parse the content of these web pages for Extract items.

To create a Spider, inherit the scrapy.Spider base class and identify three main, mandatory properties:

name：爬虫的识别名，它必须是唯一的，在不同的爬虫中你必须定义不同的名字.
start_urls：包含了Spider在启动时进行爬取的url列表。因此，第一个被获取到的页面将是其中之一。后续的URL则从初始的URL获取到的数据中提取。我们可以利用正则表达式定义和过滤需要进行跟进的链接。
parse()：是spider的一个方法。被调用时，每个初始URL完成下载后生成的 Response 对象将会作为唯一的参数传递给该函数。该方法负责解析返回的数据(response data)，提取数据(生成item)以及生成需要进一步处理的URL的 Request 对象。
这个方法负责解析返回的数据、匹配抓取的数据(解析为 item )并跟踪更多的 URL。
在 /tutorial/tutorial/spiders 目录下创建 dmoz_spider.py

import scrapy

class DmozSpider(scrapy.Spider):
 name = "dmoz"
 allowed_domains = ["dmoz.org"]
 start_urls = [
  "http://www.dmoz.org/Computers/Programming/Languages/Python/Books/",
  "http://www.dmoz.org/Computers/Programming/Languages/Python/Resources/"
 ]

 def parse(self, response):
  filename = response.url.split("/")[-2]
  with open(filename, 'wb') as f:
   f.write(response.body)

3.3. 爬取
当前项目结构

├── scrapy.cfg
└── tutorial
 ├── __init__.py
 ├── items.py
 ├── pipelines.py
 ├── settings.py
 └── spiders
  ├── __init__.py
  └── dmoz_spider.py

到项目根目录, 然后运行命令:

$ scrapy crawl dmoz

运行结果:

2014-12-15 09:30:59+0800 [scrapy] INFO: Scrapy 0.24.4 started (bot: tutorial)
2014-12-15 09:30:59+0800 [scrapy] INFO: Optional features available: ssl, http11
2014-12-15 09:30:59+0800 [scrapy] INFO: Overridden settings: {'NEWSPIDER_MODULE': 'tutorial.spiders', 'SPIDER_MODULES': ['tutorial.spiders'], 'BOT_NAME': 'tutorial'}
2014-12-15 09:30:59+0800 [scrapy] INFO: Enabled extensions: LogStats, TelnetConsole, CloseSpider, WebService, CoreStats, SpiderState
2014-12-15 09:30:59+0800 [scrapy] INFO: Enabled downloader middlewares: HttpAuthMiddleware, DownloadTimeoutMiddleware, UserAgentMiddleware, RetryMiddleware, DefaultHeadersMiddleware, MetaRefreshMiddleware, HttpCompressionMiddleware, RedirectMiddleware, CookiesMiddleware, ChunkedTransferMiddleware, DownloaderStats
2014-12-15 09:30:59+0800 [scrapy] INFO: Enabled spider middlewares: HttpErrorMiddleware, OffsiteMiddleware, RefererMiddleware, UrlLengthMiddleware, DepthMiddleware
2014-12-15 09:30:59+0800 [scrapy] INFO: Enabled item pipelines:
2014-12-15 09:30:59+0800 [dmoz] INFO: Spider opened
2014-12-15 09:30:59+0800 [dmoz] INFO: Crawled 0 pages (at 0 pages/min), scraped 0 items (at 0 items/min)
2014-12-15 09:30:59+0800 [scrapy] DEBUG: Telnet console listening on 127.0.0.1:6023
2014-12-15 09:30:59+0800 [scrapy] DEBUG: Web service listening on 127.0.0.1:6080
2014-12-15 09:31:00+0800 [dmoz] DEBUG: Crawled (200) <GET http://www.dmoz.org/Computers/Programming/Languages/Python/Resources/> (referer: None)
2014-12-15 09:31:00+0800 [dmoz] DEBUG: Crawled (200) <GET http://www.dmoz.org/Computers/Programming/Languages/Python/Books/> (referer: None)
2014-12-15 09:31:00+0800 [dmoz] INFO: Closing spider (finished)
2014-12-15 09:31:00+0800 [dmoz] INFO: Dumping Scrapy stats:
 {'downloader/request_bytes': 516,
  'downloader/request_count': 2,
  'downloader/request_method_count/GET': 2,
  'downloader/response_bytes': 16338,
  'downloader/response_count': 2,
  'downloader/response_status_count/200': 2,
  'finish_reason': 'finished',
  'finish_time': datetime.datetime(2014, 12, 15, 1, 31, 0, 666214),
  'log_count/DEBUG': 4,
  'log_count/INFO': 7,
  'response_received_count': 2,
  'scheduler/dequeued': 2,
  'scheduler/dequeued/memory': 2,
  'scheduler/enqueued': 2,
  'scheduler/enqueued/memory': 2,
  'start_time': datetime.datetime(2014, 12, 15, 1, 30, 59, 533207)}
2014-12-15 09:31:00+0800 [dmoz] INFO: Spider closed (finished)

3.4. 提取Items
3.4.1. 介绍Selector
从网页中提取数据有很多方法。Scrapy使用了一种基于 XPath 或者 CSS 表达式机制： Scrapy Selectors

出XPath表达式的例子及对应的含义:

• /html/head/title: 选择HTML文档中 标签内的
• /html/head/title/text(): 选择
• //td: 选择所有的 元素
• //div[@class="mine"]: 选择所有具有class="mine" 属性的 div 元素

等多强大的功能使用可以查看XPath tutorial

为了方便使用 XPaths，Scrapy 提供 Selector 类， 有四种方法 :

• xpath()：返回selectors列表, 每一个selector表示一个xpath参数表达式选择的节点.
• css() : 返回selectors列表, 每一个selector表示CSS参数表达式选择的节点
• extract()：返回一个unicode字符串，该字符串为XPath选择器返回的数据
• re()： 返回unicode字符串列表，字符串作为参数由正则表达式提取出来

3.4.2. 取出数据


• 首先使用谷歌浏览器开发者工具, 查看网站源码, 来看自己需要取出的数据形式(这种方法比较麻烦), 更简单的方法是直接对感兴趣的东西右键审查元素, 可以直接查看网站源码

在查看网站源码后, 网站信息在第二个

内

<ul class="directory-url" style="margin-left:0;">

 <li><a href="http://www.pearsonhighered.com/educator/academic/product/0,,0130260363,00%2Ben-USS_01DBC.html" class="listinglink">Core Python Programming</a> 
- By Wesley J. Chun; Prentice Hall PTR, 2001, ISBN 0130260363. For experienced developers to improve extant skills; professional level examples. Starts by introducing syntax, objects, error handling, functions, classes, built-ins. [Prentice Hall]
<div class="flag"><a href="/public/flag&#63;cat=Computers%2FProgramming%2FLanguages%2FPython%2FBooks&url=http%3A%2F%2Fwww.pearsonhighered.com%2Feducator%2Facademic%2Fproduct%2F0%2C%2C0130260363%2C00%252Ben-USS_01DBC.html"><img src="/img/flag.png" alt="Python crawler programming framework Scrapy introductory learning tutorial" title="report an issue with this listing"></a></div>
</li>
...省略部分...
</ul>

那么就可以通过一下方式进行提取数据

#通过如下命令选择每个在网站中的 <li> 元素:
sel.xpath('//ul/li')

#网站描述:
sel.xpath('//ul/li/text()').extract()

#网站标题:
sel.xpath('//ul/li/a/text()').extract()

#网站链接:
sel.xpath('//ul/li/a/@href').extract()

如前所述，每个 xpath() 调用返回一个 selectors 列表，所以我们可以结合 xpath() 去挖掘更深的节点。我们将会用到这些特性，所以:

for sel in response.xpath('//ul/li')
 title = sel.xpath('a/text()').extract()
 link = sel.xpath('a/@href').extract()
 desc = sel.xpath('text()').extract()
 print title, link, desc

在已有的爬虫文件中修改代码

import scrapy

class DmozSpider(scrapy.Spider):
 name = "dmoz"
 allowed_domains = ["dmoz.org"]
 start_urls = [
  "http://www.dmoz.org/Computers/Programming/Languages/Python/Books/",
  "http://www.dmoz.org/Computers/Programming/Languages/Python/Resources/"
 ]

 def parse(self, response):
  for sel in response.xpath('//ul/li'):
   title = sel.xpath('a/text()').extract()
   link = sel.xpath('a/@href').extract()
   desc = sel.xpath('text()').extract()
   print title, link, desc

3.4.3. 使用item
Item对象是自定义的python字典,可以使用标准的字典语法来获取到其每个字段的值(字段即是我们之前用Field赋值的属性)

>>> item = DmozItem()
>>> item['title'] = 'Example title'
>>> item['title']
'Example title'

一般来说，Spider将会将爬取到的数据以 Item 对象返回, 最后修改爬虫类，使用 Item 来保存数据，代码如下

from scrapy.spider import Spider
from scrapy.selector import Selector
from tutorial.items import DmozItem


class DmozSpider(Spider):
 name = "dmoz"
 allowed_domains = ["dmoz.org"]
 start_urls = [
  "http://www.dmoz.org/Computers/Programming/Languages/Python/Books/",
  "http://www.dmoz.org/Computers/Programming/Languages/Python/Resources/",
 ]

 def parse(self, response):
  sel = Selector(response)
  sites = sel.xpath('//ul[@class="directory-url"]/li')
  items = []

  for site in sites:
   item = DmozItem()
   item['name'] = site.xpath('a/text()').extract()
   item['url'] = site.xpath('a/@href').extract()
   item['description'] = site.xpath('text()').re('-\s[^\n]*\\r')
   items.append(item)
  return items

3.5. 使用Item Pipeline
当Item在Spider中被收集之后，它将会被传递到Item Pipeline，一些组件会按照一定的顺序执行对Item的处理。
每个item pipeline组件(有时称之为ItemPipeline)是实现了简单方法的Python类。他们接收到Item并通过它执行一些行为，同时也决定此Item是否继续通过pipeline，或是被丢弃而不再进行处理。
以下是item pipeline的一些典型应用：

• 清理HTML数据
• 验证爬取的数据(检查item包含某些字段)
• 查重(并丢弃)
• 将爬取结果保存，如保存到数据库、XML、JSON等文件中

编写你自己的item pipeline很简单，每个item pipeline组件是一个独立的Python类，同时必须实现以下方法:

（1）process_item(item, spider)  #每个item pipeline组件都需要调用该方法，这个方法必须返回一个 Item (或任何继承类)对象，或是抛出 DropItem异常，被丢弃的item将不会被之后的pipeline组件所处理。

#参数:

item: 由 parse 方法返回的 Item 对象(Item对象)

spider: 抓取到这个 Item 对象对应的爬虫对象(Spider对象)

（2）open_spider(spider)  #当spider被开启时，这个方法被调用。

#参数:

spider : (Spider object) – 被开启的spider

（3）close_spider(spider)  #当spider被关闭时，这个方法被调用，可以再爬虫关闭后进行相应的数据处理。

#参数:

spider : (Spider object) – 被关闭的spider

为JSON文件编写一个items

from scrapy.exceptions import DropItem

class TutorialPipeline(object):

 # put all words in lowercase
 words_to_filter = ['politics', 'religion']

 def process_item(self, item, spider):
  for word in self.words_to_filter:
   if word in unicode(item['description']).lower():
    raise DropItem("Contains forbidden word: %s" % word)
  else:
   return item

在 settings.py 中设置ITEM_PIPELINES激活item pipeline，其默认为[]

ITEM_PIPELINES = {'tutorial.pipelines.FilterWordsPipeline': 1}

3.6. 存储数据
使用下面的命令存储为json文件格式

scrapy crawl dmoz -o items.json

4.示例
4.1最简单的spider（默认的Spider）
用实例属性start_urls中的URL构造Request对象
框架负责执行request
将request返回的response对象传递给parse方法做分析

简化后的源码：

class Spider(object_ref):
 """Base class for scrapy spiders. All spiders must inherit from this
 class.
 """
 
 name = None
 
 def __init__(self, name=None, **kwargs):
  if name is not None:
   self.name = name
  elif not getattr(self, 'name', None):
   raise ValueError("%s must have a name" % type(self).__name__)
  self.__dict__.update(kwargs)
  if not hasattr(self, 'start_urls'):
   self.start_urls = []
 
 def start_requests(self):
  for url in self.start_urls:
   yield self.make_requests_from_url(url)
 
 def make_requests_from_url(self, url):
  return Request(url, dont_filter=True)
 
 def parse(self, response):
  raise NotImplementedError
 
 
BaseSpider = create_deprecated_class('BaseSpider', Spider)

一个回调函数返回多个request的例子

import scrapyfrom myproject.items import MyItemclass MySpider(scrapy.Spider):
 name = 'example.com'
 allowed_domains = ['example.com']
 start_urls = [
  'http://www.example.com/1.html',
  'http://www.example.com/2.html',
  'http://www.example.com/3.html',
 ]
 
 def parse(self, response):
  sel = scrapy.Selector(response)
  for h3 in response.xpath('//h3').extract():
   yield MyItem(title=h3)
 
  for url in response.xpath('//a/@href').extract():
   yield scrapy.Request(url, callback=self.parse)

构造一个Request对象只需两个参数： URL和回调函数

4.2CrawlSpider
通常我们需要在spider中决定：哪些网页上的链接需要跟进， 哪些网页到此为止，无需跟进里面的链接。CrawlSpider为我们提供了有用的抽象——Rule，使这类爬取任务变得简单。你只需在rule中告诉scrapy,哪些是需要跟进的。
回忆一下我们爬行mininova网站的spider.

class MininovaSpider(CrawlSpider):
 name = 'mininova'
 allowed_domains = ['mininova.org']
 start_urls = ['http://www.mininova.org/yesterday']
 rules = [Rule(LinkExtractor(allow=['/tor/\d+']), 'parse_torrent')]
 
 def parse_torrent(self, response):
  torrent = TorrentItem()
  torrent['url'] = response.url
  torrent['name'] = response.xpath("//h1/text()").extract()
  torrent['description'] = response.xpath("//div[@id='description']").extract()
  torrent['size'] = response.xpath("//div[@id='specifications']/p[2]/text()[2]").extract()
  return torrent

上面代码中 rules的含义是：匹配/tor/\d+的URL返回的内容，交给parse_torrent处理，并且不再跟进response上的URL。
官方文档中也有个例子：

 rules = (
  # 提取匹配 'category.php' (但不匹配 'subsection.php') 的链接并跟进链接(没有callback意味着follow默认为True)
  Rule(LinkExtractor(allow=('category\.php', ), deny=('subsection\.php', ))),
 
  # 提取匹配 'item.php' 的链接并使用spider的parse_item方法进行分析
  Rule(LinkExtractor(allow=('item\.php', )), callback='parse_item'),
 )

除了Spider和CrawlSpider外，还有XMLFeedSpider, CSVFeedSpider, SitemapSpider