爬蟲框架:scrapy
- 一 介紹
- 二 安裝
- 三 命令列工具
- 四 專案結構以及爬蟲應用簡介
- 五 Spiders
- 六 Selectors
- 七 Items
- 八 Item Pipeline
- 九 Dowloader Middeware
- 十 Spider Middleware
- 十一 自定義擴充套件
- 十二 settings.py
一 介紹
Scrapy一個開源和協作的框架,其最初是為了頁面抓取 (更確切來說, 網路抓取 )所設計的,使用它可以以快速、簡單、可擴充套件的方式從網站中提取所需的資料。但目前Scrapy的用途十分廣泛,可用於如資料探勘、監測和自動化測試等領域,也可以應用在獲取API所返回的資料(例如 Amazon Associates Web Services ) 或者通用的網路爬蟲。
Scrapy 是基於twisted框架開發而來,twisted是一個流行的事件驅動的python網路框架。因此Scrapy使用了一種非阻塞(又名非同步)的程式碼來實現併發。整體架構大致如下
The data flow in Scrapy is controlled by the execution engine, and goes like this:
- TheEnginegets the initial Requests to crawl from theSpider.
- TheEngineschedules the Requests in theSchedulerand asks for the next Requests to crawl.
- TheSchedulerreturns the next Requests to theEngine.
- TheEnginesends the Requests to theDownloader, passing through theDownloader Middlewares(see
process_request()
). - Once the page finishes downloading theDownloadergenerates a Response (with that page) and sends it to the Engine, passing through theDownloader Middlewares
process_response()
). - TheEnginereceives the Response from theDownloaderand sends it to theSpiderfor processing, passing through theSpider Middleware(see
process_spider_input()
). - TheSpiderprocesses the Response and returns scraped items and new Requests (to follow) to theEngine, passing through theSpider Middleware(see
process_spider_output()
). - TheEnginesends processed items toItem Pipelines, then send processed Requests to theSchedulerand asks for possible next Requests to crawl.
- The process repeats (from step 1) until there are no more requests from theScheduler.
Components:
- 引擎(EGINE)
引擎負責控制系統所有元件之間的資料流,並在某些動作發生時觸發事件。有關詳細資訊,請參見上面的資料流部分。
- 排程器(SCHEDULER)
用來接受引擎發過來的請求, 壓入佇列中, 並在引擎再次請求的時候返回. 可以想像成一個URL的優先順序佇列, 由它來決定下一個要抓取的網址是什麼, 同時去除重複的網址 - 下載器(DOWLOADER)
用於下載網頁內容, 並將網頁內容返回給EGINE,下載器是建立在twisted這個高效的非同步模型上的 - 爬蟲(SPIDERS)
SPIDERS是開發人員自定義的類,用來解析responses,並且提取items,或者傳送新的請求 - 專案管道(ITEM PIPLINES)
在items被提取後負責處理它們,主要包括清理、驗證、持久化(比如存到資料庫)等操作 - 下載器中介軟體(Downloader Middlewares)
位於Scrapy引擎和下載器之間,主要用來處理從EGINE傳到DOWLOADER的請求request,已經從DOWNLOADER傳到EGINE的響應response,你可用該中介軟體做以下幾件事- process a request just before it is sent to the Downloader (i.e. right before Scrapy sends the request to the website);
- change received response before passing it to a spider;
- send a new Request instead of passing received response to a spider;
- pass response to a spider without fetching a web page;
- silently drop some requests.
- 爬蟲中介軟體(Spider Middlewares)
位於EGINE和SPIDERS之間,主要工作是處理SPIDERS的輸入(即responses)和輸出(即requests)
官網連結:https://docs.scrapy.org/en/latest/topics/architecture.html
二 安裝
#Windows平臺 1、pip3 install wheel #安裝後,便支援通過wheel檔案安裝軟體,wheel檔案官網:https://www.lfd.uci.edu/~gohlke/pythonlibs 3、pip3 install lxml 4、pip3 install pyopenssl 5、下載並安裝pywin32:https://sourceforge.net/projects/pywin32/files/pywin32/ 6、下載twisted的wheel檔案:http://www.lfd.uci.edu/~gohlke/pythonlibs/#twisted 7、執行pip3 install 下載目錄\Twisted-17.9.0-cp36-cp36m-win_amd64.whl 8、pip3 install scrapy #Linux平臺 1、pip3 install scrapy
三 命令列工具
#1 檢視幫助 scrapy -h scrapy <command> -h #2 有兩種命令:其中Project-only必須切到專案資料夾下才能執行,而Global的命令則不需要 Global commands: startproject #建立專案 genspider #建立爬蟲程式 settings #如果是在專案目錄下,則得到的是該專案的配置 runspider #執行一個獨立的python檔案,不必建立專案 shell #scrapy shell url地址 在互動式除錯,如選擇器規則正確與否 fetch #獨立於程單純地爬取一個頁面,可以拿到請求頭 view #下載完畢後直接彈出瀏覽器,以此可以分辨出哪些資料是ajax請求 version #scrapy version 檢視scrapy的版本,scrapy version -v檢視scrapy依賴庫的版本 Project-only commands: crawl #執行爬蟲,必須建立專案才行,確保配置檔案中ROBOTSTXT_OBEY = False check #檢測專案中有無語法錯誤 list #列出專案中所包含的爬蟲名 edit #編輯器,一般不用 parse #scrapy parse url地址 --callback 回撥函式 #以此可以驗證我們的回撥函式是否正確 bench #scrapy bentch壓力測試 #3 官網連結 https://docs.scrapy.org/en/latest/topics/commands.html
#1、執行全域性命令:請確保不在某個專案的目錄下,排除受該專案配置的影響 scrapy startproject MyProject cd MyProject scrapy genspider baidu www.baidu.com scrapy settings --get XXX #如果切換到專案目錄下,看到的則是該專案的配置 scrapy runspider baidu.py scrapy shell https://www.baidu.com response response.status response.body view(response) scrapy view https://www.taobao.com #如果頁面顯示內容不全,不全的內容則是ajax請求實現的,以此快速定位問題 scrapy fetch --nolog --headers https://www.taobao.com scrapy version #scrapy的版本 scrapy version -v #依賴庫的版本 #2、執行專案命令:切到專案目錄下 scrapy crawl baidu scrapy check scrapy list scrapy parse http://quotes.toscrape.com/ --callback parse scrapy bench示範用法
四 專案結構以及爬蟲應用簡介
project_name/ scrapy.cfg project_name/ __init__.py items.py pipelines.py settings.py spiders/ __init__.py 爬蟲1.py 爬蟲2.py 爬蟲3.py
檔案說明:
- scrapy.cfg 專案的主配置資訊,用來部署scrapy時使用,爬蟲相關的配置資訊在settings.py檔案中。
- items.py 設定資料儲存模板,用於結構化資料,如:Django的Model
- pipelines 資料處理行為,如:一般結構化的資料持久化
- settings.py 配置檔案,如:遞迴的層數、併發數,延遲下載等。強調:配置檔案的選項必須大寫否則視為無效,正確寫法USER_AGENT='xxxx'
- spiders 爬蟲目錄,如:建立檔案,編寫爬蟲規則
注意:一般建立爬蟲檔案時,以網站域名命名
#在專案目錄下新建:entrypoint.py from scrapy.cmdline import execute execute(['scrapy', 'crawl', 'xiaohua'])預設只能在cmd中執行爬蟲,如果想在pycharm中執行需要做
import sys,os sys.stdout=io.TextIOWrapper(sys.stdout.buffer,encoding='gb18030')關於windows編碼
五 Spiders
1、介紹
#1、Spiders是由一系列類(定義了一個網址或一組網址將被爬取)組成,具體包括如何執行爬取任務並且如何從頁面中提取結構化的資料。 #2、換句話說,Spiders是你為了一個特定的網址或一組網址自定義爬取和解析頁面行為的地方
2、Spiders會迴圈做如下事情
#1、生成初始的Requests來爬取第一個URLS,並且標識一個回撥函式 第一個請求定義在start_requests()方法內預設從start_urls列表中獲得url地址來生成Request請求,預設的回撥函式是parse方法。回撥函式在下載完成返回response時自動觸發 #2、在回撥函式中,解析response並且返回值 返回值可以4種: 包含解析資料的字典 Item物件 新的Request物件(新的Requests也需要指定一個回撥函式) 或者是可迭代物件(包含Items或Request) #3、在回撥函式中解析頁面內容 通常使用Scrapy自帶的Selectors,但很明顯你也可以使用Beutifulsoup,lxml或其他你愛用啥用啥。 #4、最後,針對返回的Items物件將會被持久化到資料庫 通過Item Pipeline元件存到資料庫:https://docs.scrapy.org/en/latest/topics/item-pipeline.html#topics-item-pipeline) 或者匯出到不同的檔案(通過Feed exports:https://docs.scrapy.org/en/latest/topics/feed-exports.html#topics-feed-exports)
3、Spiders總共提供了五種類:
#1、scrapy.spiders.Spider #scrapy.Spider等同於scrapy.spiders.Spider #2、scrapy.spiders.CrawlSpider #3、scrapy.spiders.XMLFeedSpider #4、scrapy.spiders.CSVFeedSpider #5、scrapy.spiders.SitemapSpider
4、匯入使用
# -*- coding: utf-8 -*- import scrapy from scrapy.spiders import Spider,CrawlSpider,XMLFeedSpider,CSVFeedSpider,SitemapSpider class AmazonSpider(scrapy.Spider): #自定義類,繼承Spiders提供的基類 name = 'amazon' allowed_domains = ['www.amazon.cn'] start_urls = ['http://www.amazon.cn/'] def parse(self, response): pass
5、class scrapy.spiders.Spider
這是最簡單的spider類,任何其他的spider類都需要繼承它(包含你自己定義的)。
該類不提供任何特殊的功能,它僅提供了一個預設的start_requests方法預設從start_urls中讀取url地址傳送requests請求,並且預設parse作為回撥函式
class AmazonSpider(scrapy.Spider): name = 'amazon' allowed_domains = ['www.amazon.cn'] start_urls = ['http://www.amazon.cn/'] custom_settings = { 'BOT_NAME' : 'Egon_Spider_Amazon', 'REQUEST_HEADERS' : { 'Accept': 'text/html,application/xhtml+xml,application/xml;q=0.9,*/*;q=0.8', 'Accept-Language': 'en', } } def parse(self, response): pass
#1、name = 'amazon' 定義爬蟲名,scrapy會根據該值定位爬蟲程式 所以它必須要有且必須唯一(In Python 2 this must be ASCII only.) #2、allowed_domains = ['www.amazon.cn'] 定義允許爬取的域名,如果OffsiteMiddleware啟動(預設就啟動), 那麼不屬於該列表的域名及其子域名都不允許爬取 如果爬取的網址為:https://www.example.com/1.html,那就新增'example.com'到列表. #3、start_urls = ['http://www.amazon.cn/'] 如果沒有指定url,就從該列表中讀取url來生成第一個請求 #4、custom_settings 值為一個字典,定義一些配置資訊,在執行爬蟲程式時,這些配置會覆蓋專案級別的配置 所以custom_settings必須被定義成一個類屬性,由於settings會在類例項化前被載入 #5、settings 通過self.settings['配置項的名字']可以訪問settings.py中的配置,如果自己定義了custom_settings還是以自己的為準 #6、logger 日誌名預設為spider的名字 self.logger.debug('=============>%s' %self.settings['BOT_NAME']) #5、crawler:瞭解 該屬性必須被定義到類方法from_crawler中 #6、from_crawler(crawler, *args, **kwargs):瞭解 You probably won’t need to override this directly because the default implementation acts as a proxy to the __init__() method, calling it with the given arguments args and named arguments kwargs. #7、start_requests() 該方法用來發起第一個Requests請求,且必須返回一個可迭代的物件。它在爬蟲程式開啟時就被Scrapy呼叫,Scrapy只調用它一次。 預設從start_urls裡取出每個url來生成Request(url, dont_filter=True) #針對引數dont_filter,請看自定義去重規則 如果你想要改變起始爬取的Requests,你就需要覆蓋這個方法,例如你想要起始傳送一個POST請求,如下 class MySpider(scrapy.Spider): name = 'myspider' def start_requests(self): return [scrapy.FormRequest("http://www.example.com/login", formdata={'user': 'john', 'pass': 'secret'}, callback=self.logged_in)] def logged_in(self, response): # here you would extract links to follow and return Requests for # each of them, with another callback pass #8、parse(response) 這是預設的回撥函式,所有的回撥函式必須返回an iterable of Request and/or dicts or Item objects. #9、log(message[, level, component]):瞭解 Wrapper that sends a log message through the Spider’s logger, kept for backwards compatibility. For more information see Logging from Spiders. #10、closed(reason) 爬蟲程式結束時自動觸發定製scrapy.spider屬性與方法詳解
去重規則應該多個爬蟲共享的,但凡一個爬蟲爬取了,其他都不要爬了,實現方式如下 #方法一: 1、新增類屬性 visited=set() #類屬性 2、回撥函式parse方法內: def parse(self, response): if response.url in self.visited: return None ....... self.visited.add(response.url) #方法一改進:針對url可能過長,所以我們存放url的hash值 def parse(self, response): url=md5(response.request.url) if url in self.visited: return None ....... self.visited.add(url) #方法二:Scrapy自帶去重功能 配置檔案: DUPEFILTER_CLASS = 'scrapy.dupefilter.RFPDupeFilter' #預設的去重規則幫我們去重,去重規則在記憶體中 DUPEFILTER_DEBUG = False JOBDIR = "儲存範文記錄的日誌路徑,如:/root/" # 最終路徑為 /root/requests.seen,去重規則放檔案中 scrapy自帶去重規則預設為RFPDupeFilter,只需要我們指定 Request(...,dont_filter=False) ,如果dont_filter=True則告訴Scrapy這個URL不參與去重。 #方法三: 我們也可以仿照RFPDupeFilter自定義去重規則, from scrapy.dupefilter import RFPDupeFilter,看原始碼,仿照BaseDupeFilter #步驟一:在專案目錄下自定義去重檔案dup.py class UrlFilter(object): def __init__(self): self.visited = set() #或者放到資料庫 @classmethod def from_settings(cls, settings): return cls() def request_seen(self, request): if request.url in self.visited: return True self.visited.add(request.url) def open(self): # can return deferred pass def close(self, reason): # can return a deferred pass def log(self, request, spider): # log that a request has been filtered pass #步驟二:配置檔案settings.py: DUPEFILTER_CLASS = '專案名.dup.UrlFilter' # 原始碼分析: from scrapy.core.scheduler import Scheduler 見Scheduler下的enqueue_request方法:self.df.request_seen(request)去重規則:去除重複的url
#例一: import scrapy class 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): self.logger.info('A response from %s just arrived!', response.url) #例二:一個回撥函式返回多個Requests和Items import scrapy class 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): for h3 in response.xpath('//h3').extract(): yield {"title": h3} for url in response.xpath('//a/@href').extract(): yield scrapy.Request(url, callback=self.parse) #例三:在start_requests()內直接指定起始爬取的urls,start_urls就沒有用了, import scrapy from myproject.items import MyItem class MySpider(scrapy.Spider): name = 'example.com' allowed_domains = ['example.com'] def start_requests(self): yield scrapy.Request('http://www.example.com/1.html', self.parse) yield scrapy.Request('http://www.example.com/2.html', self.parse) yield scrapy.Request('http://www.example.com/3.html', self.parse) def parse(self, 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)例子
我們可能需要在命令列為爬蟲程式傳遞引數,比如傳遞初始的url,像這樣 #命令列執行 scrapy crawl myspider -a category=electronics #在__init__方法中可以接收外部傳進來的引數 import scrapy class MySpider(scrapy.Spider): name = 'myspider' def __init__(self, category=None, *args, **kwargs): super(MySpider, self).__init__(*args, **kwargs) self.start_urls = ['http://www.example.com/categories/%s' % category] #... #注意接收的引數全都是字串,如果想要結構化的資料,你需要用類似json.loads的方法引數傳遞
6、其他通用Spiders:https://docs.scrapy.org/en/latest/topics/spiders.html#generic-spiders
六 Selectors
#1 //與/ #2 text #3、extract與extract_first:從selector物件中解出內容 #4、屬性:xpath的屬性加字首@ #4、巢狀查詢 #5、設定預設值 #4、按照屬性查詢 #5、按照屬性模糊查詢 #6、正則表示式 #7、xpath相對路徑 #8、帶變數的xpath
response.selector.css() response.selector.xpath() 可簡寫為 response.css() response.xpath() #1 //與/ response.xpath('//body/a/')# response.css('div a::text') >>> response.xpath('//body/a') #開頭的//代表從整篇文件中尋找,body之後的/代表body的兒子 [] >>> response.xpath('//body//a') #開頭的//代表從整篇文件中尋找,body之後的//代表body的子子孫孫 [<Selector xpath='//body//a' data='<a href="image1.html">Name: My image 1 <'>, <Selector xpath='//body//a' data='<a href="image2.html">Name: My image 2 <'>, <Selector xpath='//body//a' data='<a href=" image3.html">Name: My image 3 <'>, <Selector xpath='//body//a' data='<a href="image4.html">Name: My image 4 <'>, <Selector xpath='//body//a' data='<a href="image5.html">Name: My image 5 <'>] #2 text >>> response.xpath('//body//a/text()') >>> response.css('body a::text') #3、extract與extract_first:從selector物件中解出內容 >>> response.xpath('//div/a/text()').extract() ['Name: My image 1 ', 'Name: My image 2 ', 'Name: My image 3 ', 'Name: My image 4 ', 'Name: My image 5 '] >>> response.css('div a::text').extract() ['Name: My image 1 ', 'Name: My image 2 ', 'Name: My image 3 ', 'Name: My image 4 ', 'Name: My image 5 '] >>> response.xpath('//div/a/text()').extract_first() 'Name: My image 1 ' >>> response.css('div a::text').extract_first() 'Name: My image 1 ' #4、屬性:xpath的屬性加字首@ >>> response.xpath('//div/a/@href').extract_first() 'image1.html' >>> response.css('div a::attr(href)').extract_first() 'image1.html' #4、巢狀查詢 >>> response.xpath('//div').css('a').xpath('@href').extract_first() 'image1.html' #5、設定預設值 >>> response.xpath('//div[@id="xxx"]').extract_first(default="not found") 'not found' #4、按照屬性查詢 response.xpath('//div[@id="images"]/a[@href="image3.html"]/text()').extract() response.css('#images a[@href="image3.html"]/text()').extract() #5、按照屬性模糊查詢 response.xpath('//a[contains(@href,"image")]/@href').extract() response.css('a[href*="image"]::attr(href)').extract() response.xpath('//a[contains(@href,"image")]/img/@src').extract() response.css('a[href*="imag"] img::attr(src)').extract() response.xpath('//*[@href="image1.html"]') response.css('*[href="image1.html"]') #6、正則表示式 response.xpath('//a/text()').re(r'Name: (.*)') response.xpath('//a/text()').re_first(r'Name: (.*)') #7、xpath相對路徑 >>> res=response.xpath('//a[contains(@href,"3")]')[0] >>> res.xpath('img') [<Selector xpath='img' data='<img src="image3_thumb.jpg">'>] >>> res.xpath('./img') [<Selector xpath='./img' data='<img src="image3_thumb.jpg">'>] >>> res.xpath('.//img') [<Selector xpath='.//img' data='<img src="image3_thumb.jpg">'>] >>> res.xpath('//img') #這就是從頭開始掃描 [<Selector xpath='//img' data='<img src="image1_thumb.jpg">'>, <Selector xpath='//img' data='<img src="image2_thumb.jpg">'>, <Selector xpath='//img' data='<img src="image3_thumb.jpg">'>, <Selector xpa th='//img' data='<img src="image4_thumb.jpg">'>, <Selector xpath='//img' data='<img src="image5_thumb.jpg">'>] #8、帶變數的xpath >>> response.xpath('//div[@id=$xxx]/a/text()',xxx='images').extract_first() 'Name: My image 1 ' >>> response.xpath('//div[count(a)=$yyy]/@id',yyy=5).extract_first() #求有5個a標籤的div的id 'images'View Code
https://docs.scrapy.org/en/latest/topics/selectors.html
七 Items
https://docs.scrapy.org/en/latest/topics/items.html
八 Item Pipeline
#一:可以寫多個Pipeline類 #1、如果優先順序高的Pipeline的process_item返回一個值或者None,會自動傳給下一個pipline的process_item, #2、如果只想讓第一個Pipeline執行,那得讓第一個pipline的process_item丟擲異常raise DropItem() #3、可以用spider.name == '爬蟲名' 來控制哪些爬蟲用哪些pipeline 二:示範 from scrapy.exceptions import DropItem class CustomPipeline(object): def __init__(self,v): self.value = v @classmethod def from_crawler(cls, crawler): """ Scrapy會先通過getattr判斷我們是否自定義了from_crawler,有則調它來完 成例項化 """ val = crawler.settings.getint('MMMM') return cls(val) def open_spider(self,spider): """ 爬蟲剛啟動時執行一次 """ print('000000') def close_spider(self,spider): """ 爬蟲關閉時執行一次 """ print('111111') def process_item(self, item, spider): # 操作並進行持久化 # return表示會被後續的pipeline繼續處理 return item # 表示將item丟棄,不會被後續pipeline處理 # raise DropItem()自定義pipeline
#1、settings.py HOST="127.0.0.1" PORT=27017 USER="root" PWD="123" DB="amazon" TABLE="goods" ITEM_PIPELINES = { 'Amazon.pipelines.CustomPipeline': 200, } #2、pipelines.py class CustomPipeline(object): def __init__(self,host,port,user,pwd,db,table): self.host=host self.port=port self.user=user self.pwd=pwd self.db=db self.table=table @classmethod def from_crawler(cls, crawler): """ Scrapy會先通過getattr判斷我們是否自定義了from_crawler,有則調它來完 成例項化 """ HOST = crawler.settings.get('HOST') PORT = crawler.settings.get('PORT') USER = crawler.settings.get('USER') PWD = crawler.settings.get('PWD') DB = crawler.settings.get('DB') TABLE = crawler.settings.get('TABLE') return cls(HOST,PORT,USER,PWD,DB,TABLE) def open_spider(self,spider): """ 爬蟲剛啟動時執行一次 """ self.client = MongoClient('mongodb://%s:%s@%s:%s' %(self.user,self.pwd,self.host,self.port)) def close_spider(self,spider): """ 爬蟲關閉時執行一次 """ self.client.close() def process_item(self, item, spider): # 操作並進行持久化 self.client[self.db][self.table].save(dict(item))示範
https://docs.scrapy.org/en/latest/topics/item-pipeline.html
九 Dowloader Middeware
下載中介軟體的用途 1、在process——request內,自定義下載,不用scrapy的下載 2、對請求進行二次加工,比如 設定請求頭 設定cookie 新增代理 scrapy自帶的代理元件: from scrapy.downloadermiddlewares.httpproxy import HttpProxyMiddleware from urllib.request import getproxies
class DownMiddleware1(object): def process_request(self, request, spider): """ 請求需要被下載時,經過所有下載器中介軟體的process_request呼叫 :param request: :param spider: :return: None,繼續後續中介軟體去下載; Response物件,停止process_request的執行,開始執行process_response Request物件,停止中介軟體的執行,將Request重新排程器 raise IgnoreRequest異常,停止process_request的執行,開始執行process_exception """ pass def process_response(self, request, response, spider): """ spider處理完成,返回時呼叫 :param response: :param result: :param spider: :return: Response 物件:轉交給其他中介軟體process_response Request 物件:停止中介軟體,request會被重新排程下載 raise IgnoreRequest 異常:呼叫Request.errback """ print('response1') return response def process_exception(self, request, exception, spider): """ 當下載處理器(download handler)或 process_request() (下載中介軟體)丟擲異常 :param response: :param exception: :param spider: :return: None:繼續交給後續中介軟體處理異常; Response物件:停止後續process_exception方法 Request物件:停止中介軟體,request將會被重新呼叫下載 """ return None下載器中介軟體
#1、與middlewares.py同級目錄下新建proxy_handle.py import requests def get_proxy(): return requests.get("http://127.0.0.1:5010/get/").text def delete_proxy(proxy): requests.get("http://127.0.0.1:5010/delete/?proxy={}".format(proxy)) #2、middlewares.py from Amazon.proxy_handle import get_proxy,delete_proxy class DownMiddleware1(object): def process_request(self, request, spider): """ 請求需要被下載時,經過所有下載器中介軟體的process_request呼叫 :param request: :param spider: :return: None,繼續後續中介軟體去下載; Response物件,停止process_request的執行,開始執行process_response Request物件,停止中介軟體的執行,將Request重新排程器 raise IgnoreRequest異常,停止process_request的執行,開始執行process_exception """ proxy="http://" + get_proxy() request.meta['download_timeout']=20 request.meta["proxy"] = proxy print('為%s 新增代理%s ' % (request.url, proxy),end='') print('元資料為',request.meta) def process_response(self, request, response, spider): """ spider處理完成,返回時呼叫 :param response: :param result: :param spider: :return: Response 物件:轉交給其他中介軟體process_response Request 物件:停止中介軟體,request會被重新排程下載 raise IgnoreRequest 異常:呼叫Request.errback """ print('返回狀態嗎',response.status) return response def process_exception(self, request, exception, spider): """ 當下載處理器(download handler)或 process_request() (下載中介軟體)丟擲異常 :param response: :param exception: :param spider: :return: None:繼續交給後續中介軟體處理異常; Response物件:停止後續process_exception方法 Request物件:停止中介軟體,request將會被重新呼叫下載 """ print('代理%s,訪問%s出現異常:%s' %(request.meta['proxy'],request.url,exception)) import time time.sleep(5) delete_proxy(request.meta['proxy'].split("//")[-1]) request.meta['proxy']='http://'+get_proxy() return request配置代理
十 Spider Middleware
1、爬蟲中介軟體方法介紹
from scrapy import signals class SpiderMiddleware(object): # Not all methods need to be defined. If a method is not defined, # scrapy acts as if the spider middleware does not modify the # passed objects. @classmethod def from_crawler(cls, crawler): # This method is used by Scrapy to create your spiders. s = cls() crawler.signals.connect(s.spider_opened, signal=signals.spider_opened) #當前爬蟲執行時觸發spider_opened return s def spider_opened(self, spider): # spider.logger.info('我是egon派來的爬蟲1: %s' % spider.name) print('我是egon派來的爬蟲1: %s' % spider.name) def process_start_requests(self, start_requests, spider): # Called with the start requests of the spider, and works # similarly to the process_spider_output() method, except # that it doesn’t have a response associated. # Must return only requests (not items). print('start_requests1') for r in start_requests: yield r def process_spider_input(self, response, spider): # Called for each response that goes through the spider # middleware and into the spider. # 每個response經過爬蟲中介軟體進入spider時呼叫 # 返回值:Should return None or raise an exception. #1、None: 繼續執行其他中介軟體的process_spider_input #2、丟擲異常: # 一旦丟擲異常則不再執行其他中介軟體的process_spider_input # 並且觸發request繫結的errback # errback的返回值倒著傳給中介軟體的process_spider_output # 如果未找到errback,則倒著執行中介軟體的process_spider_exception print("input1") return None def process_spider_output(self, response, result, spider): # Called with the results returned from the Spider, after # it has processed the response. # Must return an iterable of Request, dict or Item objects. print('output1') # 用yield返回多次,與return返回一次是一個道理 # 如果生成器掌握不好(函式內有yield執行函式得到的是生成器而並不會立刻執行),生成器的形式會容易誤導你對中介軟體執行順序的理解 # for i in result: # yield i return result def process_spider_exception(self, response, exception, spider): # Called when a spider or process_spider_input() method # (from other spider middleware) raises an exception. # Should return either None or an iterable of Response, dict # or Item objects. print('exception1')爬蟲中介軟體
2、當前爬蟲啟動時以及初始請求產生時
#步驟一: ''' 開啟註釋: SPIDER_MIDDLEWARES = { 'Baidu.middlewares.SpiderMiddleware1': 200, 'Baidu.middlewares.SpiderMiddleware2': 300, 'Baidu.middlewares.SpiderMiddleware3': 400, } ''' #步驟二:middlewares.py from scrapy import signals class SpiderMiddleware1(object): @classmethod def from_crawler(cls, crawler): s = cls() crawler.signals.connect(s.spider_opened, signal=signals.spider_opened) #當前爬蟲執行時觸發spider_opened return s def spider_opened(self, spider): print('我是egon派來的爬蟲1: %s' % spider.name) def process_start_requests(self, start_requests, spider): # Must return only requests (not items). print('start_requests1') for r in start_requests: yield r class SpiderMiddleware2(object): @classmethod def from_crawler(cls, crawler): s = cls() crawler.signals.connect(s.spider_opened, signal=signals.spider_opened) # 當前爬蟲執行時觸發spider_opened return s def spider_opened(self, spider): print('我是egon派來的爬蟲2: %s' % spider.name) def process_start_requests(self, start_requests, spider): print('start_requests2') for r in start_requests: yield r class SpiderMiddleware3(object): @classmethod def from_crawler(cls, crawler): s = cls() crawler.signals.connect(s.spider_opened, signal=signals.spider_opened) # 當前爬蟲執行時觸發spider_opened return s def spider_opened(self, spider): print('我是egon派來的爬蟲3: %s' % spider.name) def process_start_requests(self, start_requests, spider): print('start_requests3') for r in start_requests: yield r #步驟三:分析執行結果 #1、啟動爬蟲時則立刻執行: 我是egon派來的爬蟲1: baidu 我是egon派來的爬蟲2: baidu 我是egon派來的爬蟲3: baidu #2、然後產生一個初始的request請求,依次經過爬蟲中介軟體1,2,3: start_requests1 start_requests2 start_requests3View Code
3、process_spider_input返回None時
#步驟一:開啟註釋: SPIDER_MIDDLEWARES = { 'Baidu.middlewares.SpiderMiddleware1': 200, 'Baidu.middlewares.SpiderMiddleware2': 300, 'Baidu.middlewares.SpiderMiddleware3': 400, } ''' #步驟二:middlewares.py from scrapy import signals class SpiderMiddleware1(object): def process_spider_input(self, response, spider): print("input1") def process_spider_output(self, response, result, spider): print('output1') return result def process_spider_exception(self, response, exception, spider): print('exception1') class SpiderMiddleware2(object): def process_spider_input(self, response, spider): print("input2") return None def process_spider_output(self, response, result, spider): print('output2') return result def process_spider_exception(self, response, exception, spider): print('exception2') class SpiderMiddleware3(object): def process_spider_input(self, response, spider): print("input3") return None def process_spider_output(self, response, result, spider): print('output3') return result def process_spider_exception(self, response, exception, spider): print('exception3') #步驟三:執行結果分析 #1、返回response時,依次經過爬蟲中介軟體1,2,3 input1 input2 input3 #2、spider處理完畢後,依次經過爬蟲中介軟體3,2,1 output3 output2 output1View Code
4、process_spider_input丟擲異常時
#步驟一: ''' 開啟註釋: SPIDER_MIDDLEWARES = { 'Baidu.middlewares.SpiderMiddleware1': 200, 'Baidu.middlewares.SpiderMiddleware2': 300, 'Baidu.middlewares.SpiderMiddleware3': 400, } ''' #步驟二:middlewares.py from scrapy import signals class SpiderMiddleware1(object): def process_spider_input(self, response, spider): print("input1") def process_spider_output(self, response, result, spider): print('output1') return result def process_spider_exception(self, response, exception, spider): print('exception1') class SpiderMiddleware2(object): def process_spider_input(self, response, spider): print("input2") raise Type def process_spider_output(self, response, result, spider): print('output2') return result def process_spider_exception(self, response, exception, spider): print('exception2') class SpiderMiddleware3(object): def process_spider_input(self, response, spider): print("input3") return None def process_spider_output(self, response, result, spider): print('output3') return result def process_spider_exception(self, response, exception, spider): print('exception3') #執行結果 input1 input2 exception3 exception2 exception1 #分析: #1、當response經過中介軟體1的 process_spider_input返回None,繼續交給中介軟體2的process_spider_input #2、中介軟體2的process_spider_input丟擲異常,則直接跳過後續的process_spider_input,將異常資訊傳遞給Spiders裡該請求的errback #3、沒有找到errback,則該response既沒有被Spiders正常的callback執行,也沒有被errback執行,即Spiders啥事也沒有幹,那麼開始倒著執行process_spider_exception #4、如果process_spider_exception返回None,代表該方法推卸掉責任,並沒處理異常,而是直接交給下一個process_spider_exception,全都返回None,則異常最終交給Engine丟擲View Code
5、指定errback
#步驟一:spider.py import scrapy class BaiduSpider(scrapy.Spider): name = 'baidu' allowed_domains = ['www.baidu.com'] start_urls = ['http://www.baidu.com/'] def start_requests(self): yield scrapy.Request(url='http://www.baidu.com/', callback=self.parse, errback=self.parse_err, ) def parse(self, response): pass def parse_err(self,res): #res 為異常資訊,異常已經被該函式處理了,因此不會再拋給因此,於是開始走process_spider_output return [1,2,3,4,5] #提取異常資訊中有用的資料以可迭代物件的形式存放於管道中,等待被process_spider_output取走 #步驟二: ''' 開啟註釋: SPIDER_MIDDLEWARES = { 'Baidu.middlewares.SpiderMiddleware1': 200, 'Baidu.middlewares.SpiderMiddleware2': 300, 'Baidu.middlewares.SpiderMiddleware3': 400, } ''' #步驟三:middlewares.py from scrapy import signals class SpiderMiddleware1(object): def process_spider_input(self, response, spider): print("input1") def process_spider_output(self, response, result, spider): print('output1',list(result)) return result def process_spider_exception(self, response, exception, spider): print('exception1') class SpiderMiddleware2(object): def process_spider_input(self, response, spider): print("input2") raise TypeError('input2 丟擲異常') def process_spider_output(self, response, result, spider): print('output2',list(result)) return result def process_spider_exception(self, response, exception, spider): print('exception2') class SpiderMiddleware3(object): def process_spider_input(self, response, spider): print("input3") return None def process_spider_output(self, response, result, spider): print('output3',list(result)) return result def process_spider_exception(self, response, exception, spider): print('exception3') #步驟四:執行結果分析 input1 input2 output3 [1, 2, 3, 4, 5] #parse_err的返回值放入管道中,只能被取走一次,在output3的方法內可以根據異常資訊封裝一個新的request請求 output2 [] output1 []View Code
十一 自定義擴充套件
自定義擴充套件(與django的訊號類似) 1、django的訊號是django是預留的擴充套件,訊號一旦被觸發,相應的功能就會執行 2、scrapy自定義擴充套件的好處是可以在任意我們想要的位置新增功能,而其他元件中提供的功能只能在規定的位置執行
#1、在與settings同級目錄下新建一個檔案,檔名可以為extentions.py,內容如下 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') obj = cls(val) crawler.signals.connect(obj.spider_opened, signal=signals.spider_opened) crawler.signals.connect(obj.spider_closed, signal=signals.spider_closed) return obj def spider_opened(self, spider): print('=============>open') def spider_closed(self, spider): print('=============>close') #2、配置生效 EXTENSIONS = { "Amazon.extentions.MyExtension":200 }View Code
十二 settings.py
#==>第一部分:基本配置<=== #1、專案名稱,預設的USER_AGENT由它來構成,也作為日誌記錄的日誌名 BOT_NAME = 'Amazon' #2、爬蟲應用路徑 SPIDER_MODULES = ['Amazon.spiders'] NEWSPIDER_MODULE = 'Amazon.spiders' #3、客戶端User-Agent請求頭 #USER_AGENT = 'Amazon (+http://www.yourdomain.com)' #4、是否遵循爬蟲協議 # Obey robots.txt rules ROBOTSTXT_OBEY = False #5、是否支援cookie,cookiejar進行操作cookie,預設開啟 #COOKIES_ENABLED = False #6、Telnet用於檢視當前爬蟲的資訊,操作爬蟲等...使用telnet ip port ,然後通過命令操作 #TELNETCONSOLE_ENABLED = False #TELNETCONSOLE_HOST = '127.0.0.1' #TELNETCONSOLE_PORT = [6023,] #7、Scrapy傳送HTTP請求預設使用的請求頭 #DEFAULT_REQUEST_HEADERS = { # 'Accept': 'text/html,application/xhtml+xml,application/xml;q=0.9,*/*;q=0.8', # 'Accept-Language': 'en', #} #===>第二部分:併發與延遲<=== #1、下載器總共最大處理的併發請求數,預設值16 #CONCURRENT_REQUESTS = 32 #2、每個域名能夠被執行的最大併發請求數目,預設值8 #CONCURRENT_REQUESTS_PER_DOMAIN = 16 #3、能夠被單個IP處理的併發請求數,預設值0,代表無限制,需要注意兩點 #I、如果不為零,那CONCURRENT_REQUESTS_PER_DOMAIN將被忽略,即併發數的限制是按照每個IP來計算,而不是每個域名 #II、該設定也影響DOWNLOAD_DELAY,如果該值不為零,那麼DOWNLOAD_DELAY下載延遲是限制每個IP而不是每個域 #CONCURRENT_REQUESTS_PER_IP = 16 #4、如果沒有開啟智慧限速,這個值就代表一個規定死的值,代表對同一網址延遲請求的秒數 #DOWNLOAD_DELAY = 3 #===>第三部分:智慧限速/自動節流:AutoThrottle extension<=== #一:介紹 from scrapy.contrib.throttle import AutoThrottle #http://scrapy.readthedocs.io/en/latest/topics/autothrottle.html#topics-autothrottle 設定目標: 1、比使用預設的下載延遲對站點更好 2、自動調整scrapy到最佳的爬取速度,所以使用者無需自己調整下載延遲到最佳狀態。使用者只需要定義允許最大併發的請求,剩下的事情由該擴充套件元件自動完成 #二:如何實現? 在Scrapy中,下載延遲是通過計算建立TCP連線到接收到HTTP包頭(header)之間的時間來測量的。 注意,由於Scrapy可能在忙著處理spider的回撥函式或者無法下載,因此在合作的多工環境下準確測量這些延遲是十分苦難的。 不過,這些延遲仍然是對Scrapy(甚至是伺服器)繁忙程度的合理測量,而這擴充套件就是以此為前提進行編寫的。 #三:限速演算法 自動限速演算法基於以下規則調整下載延遲 #1、spiders開始時的下載延遲是基於AUTOTHROTTLE_START_DELAY的值 #2、當收到一個response,對目標站點的下載延遲=收到響應的延遲時間/AUTOTHROTTLE_TARGET_CONCURRENCY #3、下一次請求的下載延遲就被設定成:對目標站點下載延遲時間和過去的下載延遲時間的平均值 #4、沒有達到200個response則不允許降低延遲 #5、下載延遲不能變的比DOWNLOAD_DELAY更低或者比AUTOTHROTTLE_MAX_DELAY更高 #四:配置使用 #開啟True,預設False AUTOTHROTTLE_ENABLED = True #起始的延遲 AUTOTHROTTLE_START_DELAY = 5 #最小延遲 DOWNLOAD_DELAY = 3 #最大延遲 AUTOTHROTTLE_MAX_DELAY = 10 #每秒併發請求數的平均值,不能高於 CONCURRENT_REQUESTS_PER_DOMAIN或CONCURRENT_REQUESTS_PER_IP,調高了則吞吐量增大強姦目標站點,調低了則對目標站點更加”禮貌“ #每個特定的時間點,scrapy併發請求的數目都可能高於或低於該值,這是爬蟲檢視達到的建議值而不是硬限制 AUTOTHROTTLE_TARGET_CONCURRENCY = 16.0 #除錯 AUTOTHROTTLE_DEBUG = True CONCURRENT_REQUESTS_PER_DOMAIN = 16 CONCURRENT_REQUESTS_PER_IP = 16 #===>第四部分:爬取深度與爬取方式<=== #1、爬蟲允許的最大深度,可以通過meta檢視當前深度;0表示無深度 # DEPTH_LIMIT = 3 #2、爬取時,0表示深度優先Lifo(預設);1表示廣度優先FiFo # 後進先出,深度優先 # DEPTH_PRIORITY = 0 # SCHEDULER_DISK_QUEUE = 'scrapy.squeue.PickleLifoDiskQueue' # SCHEDULER_MEMORY_QUEUE = 'scrapy.squeue.LifoMemoryQueue' # 先進先出,廣度優先 # DEPTH_PRIORITY = 1 # SCHEDULER_DISK_QUEUE = 'scrapy.squeue.PickleFifoDiskQueue' # SCHEDULER_MEMORY_QUEUE = 'scrapy.squeue.FifoMemoryQueue' #3、排程器佇列 # SCHEDULER = 'scrapy.core.scheduler.Scheduler' # from scrapy.core.scheduler import Scheduler #4、訪問URL去重 # DUPEFILTER_CLASS = 'step8_king.duplication.RepeatUrl' #===>第五部分:中介軟體、Pipelines、擴充套件<=== #1、Enable or disable spider middlewares # See http://scrapy.readthedocs.org/en/latest/topics/spider-middleware.html #SPIDER_MIDDLEWARES = { # 'Amazon.middlewares.AmazonSpiderMiddleware': 543, #} #2、Enable or disable downloader middlewares # See http://scrapy.readthedocs.org/en/latest/topics/downloader-middleware.html DOWNLOADER_MIDDLEWARES = { # 'Amazon.middlewares.DownMiddleware1': 543, } #3、Enable or disable extensions # See http://scrapy.readthedocs.org/en/latest/topics/extensions.html #EXTENSIONS = { # 'scrapy.extensions.telnet.TelnetConsole': None, #} #4、Configure item pipelines # See http://scrapy.readthedocs.org/en/latest/topics/item-pipeline.html ITEM_PIPELINES = { # 'Amazon.pipelines.CustomPipeline': 200, } #===>第六部分:快取<=== """ 1. 啟用快取 目的用於將已經發送的請求或相應快取下來,以便以後使用 from scrapy.downloadermiddlewares.httpcache import HttpCacheMiddleware from scrapy.extensions.httpcache import DummyPolicy from scrapy.extensions.httpcache import FilesystemCacheStorage """ # 是否啟用快取策略 # HTTPCACHE_ENABLED = True # 快取策略:所有請求均快取,下次在請求直接訪問原來的快取即可 # HTTPCACHE_POLICY = "scrapy.extensions.httpcache.DummyPolicy" # 快取策略:根據Http響應頭:Cache-Control、Last-Modified 等進行快取的策略 # HTTPCACHE_POLICY = "scrapy.extensions.httpcache.RFC2616Policy" # 快取超時時間 # HTTPCACHE_EXPIRATION_SECS = 0 # 快取儲存路徑 # HTTPCACHE_DIR = 'httpcache' # 快取忽略的Http狀態碼 # HTTPCACHE_IGNORE_HTTP_CODES = [] # 快取儲存的外掛 # HTTPCACHE_STORAGE = 'scrapy.extensions.httpcache.FilesystemCacheStorage' #===>第七部分:執行緒池<=== REACTOR_THREADPOOL_MAXSIZE = 10 #Default: 10 #scrapy基於twisted非同步IO框架,downloader是多執行緒的,執行緒數是Twisted執行緒池的預設大小(The maximum limit for Twisted Reactor thread pool size.) #關於twisted執行緒池: http://twistedmatrix.com/documents/10.1.0/core/howto/threading.html #執行緒池實現:twisted.python.threadpool.ThreadPool twisted調整執行緒池大小: from twisted.internet import reactor reactor.suggestThreadPoolSize(30) #scrapy相關原始碼: D:\python3.6\Lib\site-packages\scrapy\crawler.py #補充: windows下檢視程序內執行緒數的工具: https://docs.microsoft.com/zh-cn/sysinternals/downloads/pslist 或 https://pan.baidu.com/s/1jJ0pMaM 命令為: pslist |findstr python linux下:top -p 程序id #===>第八部分:其他預設配置參考<=== D:\python3.6\Lib\site-packages\scrapy\settings\default_settings.pysettings.py