来源:http://www.51testing.com/
1.背景
我们在爬取网页时一般会使用到三个爬虫库:requests,scrapy,selenium。requests一般用于小型爬虫,scrapy用于构建大的爬虫项目,而selenium主要用来应付负责的页面(复杂js渲染的页面,请求非常难构造,或者构造方式经常变化)。
在我们面对大型爬虫项目时,肯定会优选scrapy框架来开发,但是在解析复杂JS渲染的页面时,又很麻烦。尽管使用selenium浏览器渲染来抓取这样的页面很方便,这种方式下,我们不需要关心页面后台发生了怎样的请求,也不需要分析整个页面的渲染过程,我们只需要关心页面最终结果即可,可见即可爬,但是selenium的效率又太低。
所以,如果可以在scrapy中,集成selenium,让selenium负责复杂页面的爬取,那么这样的爬虫就无敌了,可以爬取任何网站了。
2. 环境
python 3.6.1
系统:win7
IDE:pycharm
安装过chrome浏览器
配置好chromedriver(设置好环境变量)
selenium 3.7.0
scrapy 1.4.0
3.原理分析
3.1. 分析request请求的流程
首先看一下scrapy最新的架构图:
部分流程:
第一:爬虫引擎生成requests请求,送往scheduler调度模块,进入等待队列,等待调度。
第二:scheduler模块开始调度这些requests,出队,发往爬虫引擎。
第三:爬虫引擎将这些requests送到下载中间件(多个,例如加header,代理,自定义等等)进行处理。
第四:处理完之后,送往Downloader模块进行下载。
从这个处理过程来看,突破口就在下载中间件部分,用selenium直接处理掉request请求。
3.2. requests和response中间处理件源码分析
相关代码位置:
源码解析:
代码语言:javascript复制 # 文件:E:MinicondaLibsite-packagesscrapycoredownloadermiddleware.py
"""
Downloader Middleware manager
See documentation in docs/topics/downloader-middleware.rst
"""
import six
from twisted.internet import defer
from scrapy.http import Request, Response
from scrapy.middleware import MiddlewareManager
from scrapy.utils.defer import mustbe_deferred
from scrapy.utils.conf import build_component_list
class DownloaderMiddlewareManager(MiddlewareManager):
component_name = 'downloader middleware'
@classmethod
def _get_mwlist_from_settings(cls, settings):
# 从settings.py或这custom_setting中拿到自定义的Middleware中间件
'''
'DOWNLOADER_MIDDLEWARES': {
'mySpider.middlewares.ProxiesMiddleware': 400,
# SeleniumMiddleware
'mySpider.middlewares.SeleniumMiddleware': 543,
'scrapy.downloadermiddlewares.useragent.UserAgentMiddleware': None,
},
'''
return build_component_list(
settings.getwithbase('DOWNLOADER_MIDDLEWARES'))
# 将所有自定义Middleware中间件的处理函数添加到对应的methods列表中
def _add_middleware(self, mw):
if hasattr(mw, 'process_request'):
self.methods['process_request'].append(mw.process_request)
if hasattr(mw, 'process_response'):
self.methods['process_response'].insert(0, mw.process_response)
if hasattr(mw, 'process_exception'):
self.methods['process_exception'].insert(0, mw.process_exception)
# 整个下载流程
def download(self, download_func, request, spider):
@defer.inlineCallbacks
def process_request(request):
# 处理request请求,依次经过各个自定义Middleware中间件的process_request方法,前面有加入到list中
for method in self.methods['process_request']:
response = yield method(request=request, spider=spider)
assert response is None or isinstance(response, (Response, Request)),
'Middleware %s.process_request must return None, Response or Request, got %s' %
(six.get_method_self(method).__class__.__name__, response.__class__.__name__)
# 这是关键地方
# 如果在某个Middleware中间件的process_request中处理完之后,生成了一个response对象
# 那么会直接将这个response return 出去,跳出循环,不再处理其他的process_request
# 之前我们的header,proxy中间件,都只是加个user-agent,加个proxy,并不做任何return值
# 还需要注意一点:就是这个return的必须是Response对象
# 后面我们构造的HtmlResponse正是Response的子类对象
if response:
defer.returnValue(response)
# 如果在上面的所有process_request中,都没有返回任何Response对象的话
# 最后,会将这个加工过的Request送往download_func,进行下载,返回的就是一个Response对象
# 然后依次经过各个Middleware中间件的process_response方法进行加工,如下
defer.returnValue((yield download_func(request=request,spider=spider)))
@defer.inlineCallbacks
def process_response(response):
assert response is not None, 'Received None in process_response'
if isinstance(response, Request):
defer.returnValue(response)
for method in self.methods['process_response']:
response = yield method(request=request, response=response,
spider=spider)
assert isinstance(response, (Response, Request)),
'Middleware %s.process_response must return Response or Request, got %s' %
(six.get_method_self(method).__class__.__name__, type(response))
if isinstance(response, Request):
defer.returnValue(response)
defer.returnValue(response)
@defer.inlineCallbacks
def process_exception(_failure):
exception = _failure.value
for method in self.methods['process_exception']:
response = yield method(request=request, exception=exception,
spider=spider)
assert response is None or isinstance(response, (Response, Request)),
'Middleware %s.process_exception must return None, Response or Request, got %s' %
(six.get_method_self(method).__class__.__name__, type(response))
if response:
defer.returnValue(response)
defer.returnValue(_failure)
deferred = mustbe_deferred(process_request, request)
deferred.addErrback(process_exception)
deferred.addCallback(process_response)
return deferred
4. 代码
在settings.py中,配置好selenium参数:
代码语言:javascript复制# 文件settings.py中
# ----------- selenium参数配置 -------------
SELENIUM_TIMEOUT = 25 # selenium浏览器的超时时间,单位秒
LOAD_IMAGE = True # 是否下载图片
WINDOW_HEIGHT = 900 # 浏览器窗口大小
WINDOW_WIDTH = 900
在spider中,生成request时,标记哪些请求需要走selenium下载:
代码语言:javascript复制 # 文件mySpider.py中
class mySpider(CrawlSpider):
name = "mySpiderAmazon"
allowed_domains = ['amazon.com']
custom_settings = {
'LOG_LEVEL':'INFO',
'DOWNLOAD_DELAY': 0,
'COOKIES_ENABLED': False, # enabled by default
'DOWNLOADER_MIDDLEWARES': {
# 代理中间件
'mySpider.middlewares.ProxiesMiddleware': 400,
# SeleniumMiddleware 中间件
'mySpider.middlewares.SeleniumMiddleware': 543,
# 将scrapy默认的user-agent中间件关闭
'scrapy.downloadermiddlewares.useragent.UserAgentMiddleware': None,
},
#.....................华丽的分割线.......................
# 生成request时,将是否使用selenium下载的标记,放入到meta中
yield Request(
url = "https://www.amazon.com/",
meta = {'usedSelenium': True, 'dont_redirect': True},
callback = self.parseIndexPage,
errback = self.error
)
在下载中间件middlewares.py中,使用selenium抓取页面(核心部分)
# -*- coding: utf-8 -*-
from selenium import webdriver
from selenium.common.exceptions import TimeoutException
from selenium.webdriver.common.by import By
from selenium.webdriver.support.ui import WebDriverWait
from selenium.webdriver.support import expected_conditions as EC
from selenium.webdriver.common.keys import Keys
from scrapy.http import HtmlResponse
from logging import getLogger
import time
class SeleniumMiddleware():
# 经常需要在pipeline或者中间件中获取settings的属性,可以通过scrapy.crawler.Crawler.settings属性
@classmethod
def from_crawler(cls, crawler):
# 从settings.py中,提取selenium设置参数,初始化类
return cls(timeout=crawler.settings.get('SELENIUM_TIMEOUT'),
isLoadImage=crawler.settings.get('LOAD_IMAGE'),
windowHeight=crawler.settings.get('WINDOW_HEIGHT'),
windowWidth=crawler.settings.get('WINDOW_WIDTH')
)
def __init__(self, timeout=30, isLoadImage=True, windowHeight=None, windowWidth=None):
self.logger = getLogger(__name__)
self.timeout = timeout
self.isLoadImage = isLoadImage
# 定义一个属于这个类的browser,防止每次请求页面时,都会打开一个新的chrome浏览器
# 这样,这个类处理的Request都可以只用这一个browser
self.browser = webdriver.Chrome()
if windowHeight and windowWidth:
self.browser.set_window_size(900, 900)
self.browser.set_page_load_timeout(self.timeout) # 页面加载超时时间
self.wait = WebDriverWait(self.browser, 25) # 指定元素加载超时时间
def process_request(self, request, spider):
'''
用chrome抓取页面
:param request: Request请求对象
:param spider: Spider对象
:return: HtmlResponse响应
'''
# self.logger.debug('chrome is getting page')
print(f"chrome is getting page")
# 依靠meta中的标记,来决定是否需要使用selenium来爬取
usedSelenium = request.meta.get('usedSelenium', False)
if usedSelenium:
try:
self.browser.get(request.url)
# 搜索框是否出现
input = self.wait.until(
EC.presence_of_element_located((By.XPATH, "//div[@class='nav-search-field ']/input"))
)
time.sleep(2)
input.clear()
input.send_keys("iphone 7s")
# 敲enter键, 进行搜索
input.send_keys(Keys.RETURN)
# 查看搜索结果是否出现
searchRes = self.wait.until(
EC.presence_of_element_located((By.XPATH, "//div[@id='resultsCol']"))
)
except Exception as e:
# self.logger.debug(f'chrome getting page error, Exception = {e}')
print(f"chrome getting page error, Exception = {e}")
return HtmlResponse(url=request.url, status=500, request=request)
else:
time.sleep(3)
return HtmlResponse(url=request.url,
body=self.browser.page_source,
request=request,
# 最好根据网页的具体编码而定
encoding='utf-8',
status=200)
5. 执行结果
6. 存在的问题
6.1. Spider关闭了,chrome没有退出。
代码语言:javascript复制 2018-04-04 09:26:18 [scrapy.statscollectors] INFO: Dumping Scrapy stats:
{'downloader/response_bytes': 2092766,
'downloader/response_count': 2,
'downloader/response_status_count/200': 2,
'finish_reason': 'finished',
'finish_time': datetime.datetime(2018, 4, 4, 1, 26, 16, 763602),
'log_count/INFO': 7,
'request_depth_max': 1,
'response_received_count': 2,
'scheduler/dequeued': 2,
'scheduler/dequeued/memory': 2,
'scheduler/enqueued': 2,
'scheduler/enqueued/memory': 2,
'start_time': datetime.datetime(2018, 4, 4, 1, 25, 48, 301602)}
2018-04-04 09:26:18 [scrapy.core.engine] INFO: Spider closed (finished)
上面,我们是把browser对象放到了Middleware中间件中,只能做process_request和process_response, 没有说在中间件中介绍如何调用scrapy的close方法。
解决方案:利用信号量的方式,当收到spider_closed信号时,调用browser.quit()
6.2. 当一个项目同时启动多个spider,会共用到Middleware中的selenium,不利于并发。
因为用scrapy selenium的方式,只有部分,甚至是一小部分页面会用到chrome,既然把chrome放到Middleware中有这么多限制,那为什么不能把chrome放到spider里面呢。这样的好处在于:每个spider都有自己的chrome,这样当启动多个spider时,就会有多个chrome,不是所有的spider共用一个chrome,这对我们的并发是有好处的。
解决方案:将chrome的初始化放到spider中,每个spider独占自己的chrome
7. 改进版代码
在settings.py中,配置好selenium参数:
代码语言:javascript复制 # 文件settings.py中
# ----------- selenium参数配置 -------------
SELENIUM_TIMEOUT = 25 # selenium浏览器的超时时间,单位秒
LOAD_IMAGE = True # 是否下载图片
WINDOW_HEIGHT = 900 # 浏览器窗口大小
WINDOW_WIDTH = 900
在spider中,生成request时,标记哪些请求需要走selenium下载:
代码语言:javascript复制 # 文件mySpider.py中
# selenium相关库
from selenium import webdriver
from selenium.webdriver.support.ui import WebDriverWait
# scrapy 信号相关库
from scrapy.utils.project import get_project_settings
# 下面这种方式,即将废弃,所以不用
# from scrapy.xlib.pydispatch import dispatcher
from scrapy import signals
# scrapy最新采用的方案
from pydispatch import dispatcher
class mySpider(CrawlSpider):
name = "mySpiderAmazon"
allowed_domains = ['amazon.com']
custom_settings = {
'LOG_LEVEL':'INFO',
'DOWNLOAD_DELAY': 0,
'COOKIES_ENABLED': False, # enabled by default
'DOWNLOADER_MIDDLEWARES': {
# 代理中间件
'mySpider.middlewares.ProxiesMiddleware': 400,
# SeleniumMiddleware 中间件
'mySpider.middlewares.SeleniumMiddleware': 543,
# 将scrapy默认的user-agent中间件关闭
'scrapy.downloadermiddlewares.useragent.UserAgentMiddleware': None,
},
# 将chrome初始化放到spider中,成为spider中的元素
def __init__(self, timeout=30, isLoadImage=True, windowHeight=None, windowWidth=None):
# 从settings.py中获取设置参数
self.mySetting = get_project_settings()
self.timeout = self.mySetting['SELENIUM_TIMEOUT']
self.isLoadImage = self.mySetting['LOAD_IMAGE']
self.windowHeight = self.mySetting['WINDOW_HEIGHT']
self.windowWidth = self.mySetting['windowWidth']
# 初始化chrome对象
self.browser = webdriver.Chrome()
if self.windowHeight and self.windowWidth:
self.browser.set_window_size(900, 900)
self.browser.set_page_load_timeout(self.timeout) # 页面加载超时时间
self.wait = WebDriverWait(self.browser, 25) # 指定元素加载超时时间
super(mySpider, self).__init__()
# 设置信号量,当收到spider_closed信号时,调用mySpiderCloseHandle方法,关闭chrome
dispatcher.connect(receiver = self.mySpiderCloseHandle,
signal = signals.spider_closed
)
# 信号量处理函数:关闭chrome浏览器
def mySpiderCloseHandle(self, spider):
print(f"mySpiderCloseHandle: enter ")
self.browser.quit()
#.....................华丽的分割线.......................
# 生成request时,将是否使用selenium下载的标记,放入到meta中
yield Request(
url = "https://www.amazon.com/",
meta = {'usedSelenium': True, 'dont_redirect': True},
callback = self.parseIndexPage,
errback = self.error
)
在下载中间件middlewares.py中,使用selenium抓取页面
代码语言:javascript复制 # -*- coding: utf-8 -*-
from selenium import webdriver
from selenium.common.exceptions import TimeoutException
from selenium.webdriver.common.by import By
from selenium.webdriver.support.ui import WebDriverWait
from selenium.webdriver.support import expected_conditions as EC
from selenium.webdriver.common.keys import Keys
from scrapy.http import HtmlResponse
from logging import getLogger
import time
class SeleniumMiddleware():
# Middleware中会传递进来一个spider,这就是我们的spider对象,从中可以获取__init__时的chrome相关元素
def process_request(self, request, spider):
'''
用chrome抓取页面
:param request: Request请求对象
:param spider: Spider对象
:return: HtmlResponse响应
'''
print(f"chrome is getting page")
# 依靠meta中的标记,来决定是否需要使用selenium来爬取
usedSelenium = request.meta.get('usedSelenium', False)
if usedSelenium:
try:
spider.browser.get(request.url)
# 搜索框是否出现
input = spider.wait.until(
EC.presence_of_element_located((By.XPATH, "//div[@class='nav-search-field ']/input"))
)
time.sleep(2)
input.clear()
input.send_keys("iphone 7s")
# 敲enter键, 进行搜索
input.send_keys(Keys.RETURN)
# 查看搜索结果是否出现
searchRes = spider.wait.until(
EC.presence_of_element_located((By.XPATH, "//div[@id='resultsCol']"))
)
except Exception as e:
print(f"chrome getting page error, Exception = {e}")
return HtmlResponse(url=request.url, status=500, request=request)
else:
time.sleep(3)
# 页面爬取成功,构造一个成功的Response对象(HtmlResponse是它的子类)
return HtmlResponse(url=request.url,
body=spider.browser.page_source,
request=request,
# 最好根据网页的具体编码而定
encoding='utf-8',
status=200)
运行结果(spider结束,执行mySpiderCloseHandle关闭chrome浏览器):
代码语言:javascript复制['categorySelectorAmazon1.pipelines.MongoPipeline']
2018-04-04 11:56:21 [scrapy.core.engine] INFO: Spider opened
2018-04-04 11:56:21 [scrapy.extensions.logstats] INFO: Crawled 0 pages (at 0 pages/min), scraped 0 items (at 0 items/min)
chrome is getting page
parseProductDetail url = https://www.amazon.com/, status = 200, meta = {'usedSelenium': True, 'dont_redirect': True, 'download_timeout': 25.0, 'proxy': 'http://H37XPSB6V57VU96D:CAB31DAEB9313CE5@proxy.abuyun.com:9020', 'depth': 0}
chrome is getting page
2018-04-04 11:56:54 [scrapy.core.engine] INFO: Closing spider (finished)
mySpiderCloseHandle: enter
2018-04-04 11:56:59 [scrapy.statscollectors] INFO: Dumping Scrapy stats:
{'downloader/response_bytes': 1938619,
'downloader/response_count': 2,
'downloader/response_status_count/200': 2,
'finish_reason': 'finished',
'finish_time': datetime.datetime(2018, 4, 4, 3, 56, 54, 301602),
'log_count/INFO': 7,
'request_depth_max': 1,
'response_received_count': 2,
'scheduler/dequeued': 2,
'scheduler/dequeued/memory': 2,
'scheduler/enqueued': 2,
'scheduler/enqueued/memory': 2,
'start_time': datetime.datetime(2018, 4, 4, 3, 56, 21, 642602)}
2018-04-04 11:56:59 [scrapy.core.engine] INFO: Spider closed (finished)