CrawlSpiders是Spider的派生類蝇闭，Spider類的設(shè)計(jì)原則是只爬取start_url列表中的網(wǎng)頁蜘拉，而CrawlSpider類定義了一些規(guī)則(rule)來提供跟進(jìn)link的方便的機(jī)制爆惧，從爬取的網(wǎng)頁中獲取link并繼續(xù)爬取的工作更適合。
一、我們先來分析一下CrawlSpiders源碼

源碼解析

class CrawlSpider(Spider):
    rules = ()
    def __init__(self, *a, **kw):
        super(CrawlSpider, self).__init__(*a, **kw)
        self._compile_rules()

    # 首先調(diào)用parse()來處理start_urls中返回的response對象
    # parse()則將這些response對象傳遞給了_parse_response()函數(shù)處理，并設(shè)置回調(diào)函數(shù)為parse_start_url()
    # 設(shè)置了跟進(jìn)標(biāo)志位True
    # parse將返回item和跟進(jìn)了的Request對象    
    def parse(self, response):
        return self._parse_response(response, self.parse_start_url, cb_kwargs={}, follow=True)

    # 處理start_url中返回的response至非，需要重寫
    def parse_start_url(self, response):
        return []

    def process_results(self, response, results):
        return results

    # 從response中抽取符合任一用戶定義'規(guī)則'的鏈接，并構(gòu)造成Resquest對象返回
    def _requests_to_follow(self, response):
        if not isinstance(response, HtmlResponse):
            return
        seen = set()
        # 抽取之內(nèi)的所有鏈接糠聪，只要通過任意一個(gè)'規(guī)則'荒椭，即表示合法
        for n, rule in enumerate(self._rules):
            links = [l for l in rule.link_extractor.extract_links(response) if l not in seen]
            # 使用用戶指定的process_links處理每個(gè)連接
            if links and rule.process_links:
                links = rule.process_links(links)
            # 將鏈接加入seen集合，為每個(gè)鏈接生成Request對象舰蟆，并設(shè)置回調(diào)函數(shù)為_repsonse_downloaded()
            for link in links:
                seen.add(link)
                # 構(gòu)造Request對象趣惠，并將Rule規(guī)則中定義的回調(diào)函數(shù)作為這個(gè)Request對象的回調(diào)函數(shù)
                r = Request(url=link.url, callback=self._response_downloaded)
                r.meta.update(rule=n, link_text=link.text)
                # 對每個(gè)Request調(diào)用process_request()函數(shù)。該函數(shù)默認(rèn)為indentify身害，即不做任何處理味悄，直接返回該Request.
                yield rule.process_request(r)

    # 處理通過rule提取出的連接，并返回item以及request
    def _response_downloaded(self, response):
        rule = self._rules[response.meta['rule']]
        return self._parse_response(response, rule.callback, rule.cb_kwargs, rule.follow)

    # 解析response對象塌鸯，會(huì)用callback解析處理他侍瑟，并返回request或Item對象
    def _parse_response(self, response, callback, cb_kwargs, follow=True):
        # 首先判斷是否設(shè)置了回調(diào)函數(shù)。（該回調(diào)函數(shù)可能是rule中的解析函數(shù)，也可能是 parse_start_url函數(shù)）
        # 如果設(shè)置了回調(diào)函數(shù)（parse_start_url()）涨颜，那么首先用parse_start_url()處理response對象费韭，
        # 然后再交給process_results處理。返回cb_res的一個(gè)列表
        if callback:
            #如果是parse調(diào)用的庭瑰，則會(huì)解析成Request對象
            #如果是rule callback星持，則會(huì)解析成Item
            cb_res = callback(response, **cb_kwargs) or ()
            cb_res = self.process_results(response, cb_res)
            for requests_or_item in iterate_spider_output(cb_res):
                yield requests_or_item

        # 如果需要跟進(jìn)，那么使用定義的Rule規(guī)則提取并返回這些Request對象
        if follow and self._follow_links:
            #返回每個(gè)Request對象
            for request_or_item in self._requests_to_follow(response):
                yield request_or_item

    def _compile_rules(self):
        def get_method(method):
            if callable(method):
                return method
            elif isinstance(method, basestring):
                return getattr(self, method, None)

        self._rules = [copy.copy(r) for r in self.rules]
        for rule in self._rules:
            rule.callback = get_method(rule.callback)
            rule.process_links = get_method(rule.process_links)
            rule.process_request = get_method(rule.process_request)

    def set_crawler(self, crawler):
        super(CrawlSpider, self).set_crawler(crawler)
        self._follow_links = crawler.settings.getbool('CRAWLSPIDER_FOLLOW_LINKS', True)

二弹灭、 CrawlSpider爬蟲文件字段的介紹
1督暂、 CrawlSpider繼承于Spider類，除了繼承過來的屬性外（name穷吮、allow_domains）逻翁，還提供了新的屬性和方法:class scrapy.linkextractors.LinkExtractorLink Extractors 的目的很簡單: 提取鏈接?每個(gè)LinkExtractor有唯一的公共方法是 extract_links()，它接收一個(gè) Response 對象酒来，并返回一個(gè) scrapy.link.Link 對象卢未。
Link Extractors要實(shí)例化一次，并且 extract_links 方法會(huì)根據(jù)不同的 response 調(diào)用多次提取鏈接?

class scrapy.linkextractors.LinkExtractor(
    allow = (),
    deny = (),
    allow_domains = (),
    deny_domains = (),
    deny_extensions = None,
    restrict_xpaths = (),
    tags = ('a','area'),
    attrs = ('href'),
    canonicalize = True,
    unique = True,
    process_value = None
)

主要參數(shù)：

① allow：滿足括號中“正則表達(dá)式”的值會(huì)被提取堰汉，如果為空辽社，則全部匹配。
② deny：與這個(gè)正則表達(dá)式(或正則表達(dá)式列表)不匹配的URL一定不提取翘鸭。
③ allow_domains：會(huì)被提取的鏈接的domains滴铅。
④ deny_domains：一定不會(huì)被提取鏈接的domains。
⑤ restrict_xpaths：使用xpath表達(dá)式就乓，和allow共同作用過濾鏈接汉匙。

2、 在rules中包含一個(gè)或多個(gè)Rule對象生蚁，每個(gè)Rule對爬取網(wǎng)站的動(dòng)作定義了特定操作噩翠。如果多個(gè)rule匹配了相同的鏈接，則根據(jù)規(guī)則在本集合中被定義的順序邦投，第一個(gè)會(huì)被使用伤锚。

class scrapy.spiders.Rule(
        link_extractor, 
        callback = None, 
        cb_kwargs = None, 
        follow = None, 
        process_links = None, 
        process_request = None
)

① link_extractor：是一個(gè)Link Extractor對象，用于定義需要提取的鏈接志衣。
② callback： 從link_extractor中每獲取到鏈接時(shí)屯援，參數(shù)所指定的值作為回調(diào)函數(shù)，該回調(diào)函數(shù)接受一個(gè)response作為其第一個(gè)參數(shù)念脯。
注意：當(dāng)編寫爬蟲規(guī)則時(shí)狞洋，避免使用parse作為回調(diào)函數(shù)。由于CrawlSpider使用parse方法來實(shí)現(xiàn)其邏輯绿店，如果覆蓋了 parse方法吉懊，crawl spider將會(huì)運(yùn)行失敗。
③ follow：是一個(gè)布爾(boolean)值，指定了根據(jù)該規(guī)則從response提取的鏈接是否需要跟進(jìn)借嗽。 如果callback為None怕午，follow 默認(rèn)設(shè)置為True ，否則默認(rèn)為False淹魄。
④ process_links：指定該spider中哪個(gè)的函數(shù)將會(huì)被調(diào)用，從link_extractor中獲取到鏈接列表時(shí)將會(huì)調(diào)用該函數(shù)堡距。該方法主要用來過濾甲锡。
⑤ process_request：指定該spider中哪個(gè)的函數(shù)將會(huì)被調(diào)用， 該規(guī)則提取到每個(gè)request時(shí)都會(huì)調(diào)用該函數(shù)羽戒。 (用來過濾request)

3缤沦、Scrapy提供了log功能，可以通過 logging 模塊使用易稠「追希可以修改配置文件settings.py，任意位置添加下面兩行驶社，效果會(huì)清爽很多企量。

LOG_FILE = "TencentSpider.log"
LOG_LEVEL = "INFO"

Scrapy提供5層logging級別:
① CRITICAL - 嚴(yán)重錯(cuò)誤(critical)
② ERROR - 一般錯(cuò)誤(regular errors)
③ WARNING - 警告信息(warning messages)
④ INFO - 一般信息(informational messages)
⑤ DEBUG - 調(diào)試信息(debugging messages)

通過在setting.py中進(jìn)行以下設(shè)置可以被用來配置logging:

① LOG_ENABLED 默認(rèn): True，啟用logging
② LOG_ENCODING 默認(rèn): 'utf-8'亡电，logging使用的編碼
③ LOG_FILE 默認(rèn): None届巩，在當(dāng)前目錄里創(chuàng)建logging輸出文件的文件名
④ LOG_LEVEL 默認(rèn): 'DEBUG'，log的最低級別
⑤ LOG_STDOUT 默認(rèn): False 如果為 True份乒，進(jìn)程所有的標(biāo)準(zhǔn)輸出(及錯(cuò)誤)將會(huì)被重定向到log中恕汇。例如，執(zhí)行 print "hello" 或辖，其將會(huì)在Scrapy log中顯示瘾英。

三、 CrawlSpider爬蟲案例分析

1颂暇、創(chuàng)建項(xiàng)目：scrapy startproject CrawlYouYuan

2缺谴、創(chuàng)建爬蟲文件：scrapy genspider -t crawl youyuan youyuan.com

3、項(xiàng)目文件分析
items.py

模型類
import scrapy
class CrawlyouyuanItem(scrapy.Item):
    # 用戶名
    username = scrapy.Field()
    # 年齡
    age = scrapy.Field()
    # 頭像圖片的鏈接
    header_url = scrapy.Field()
    # 相冊圖片的鏈接
    images_url = scrapy.Field()
    # 內(nèi)心獨(dú)白
    content = scrapy.Field()
    # 籍貫
    place_from = scrapy.Field()
    # 學(xué)歷
    education = scrapy.Field()
    # 興趣愛好
    hobby = scrapy.Field()
    # 個(gè)人主頁
    source_url = scrapy.Field()
    # 數(shù)據(jù)來源網(wǎng)站
    sourec = scrapy.Field()
    # utc 時(shí)間
    time = scrapy.Field()
    # 爬蟲名
    spidername = scrapy.Field()

youyuan.py

爬蟲文件
# -*- coding: utf-8 -*-
import scrapy
from scrapy.linkextractors import LinkExtractor
from scrapy.spiders import CrawlSpider, Rule
from CrawlYouYuan.items import CrawlyouyuanItem
import re
class YouyuanSpider(CrawlSpider):
    name = 'youyuan'
    allowed_domains = ['youyuan.com']
    start_urls = ['http://www.youyuan.com/find/beijing/mm18-25/advance-0-0-0-0-0-0-0/p1/']
    # 自動(dòng)生成的文件不需要改東西蟀架，只需要添加rules文件里面Rule角色就可以
    # 每一頁匹配規(guī)則
    page_links = LinkExtractor(allow=(r"youyuan.com/find/beijing/mm18-25/advance-0-0-0-0-0-0-0/p\d+/"))
    # 每個(gè)人個(gè)人主頁匹配規(guī)則
    profile_links = LinkExtractor(allow=(r"youyuan.com/\d+-profile/"))
    rules = (
        # 沒有回調(diào)函數(shù)瓣赂，說明follow是True
        Rule(page_links),
        # 有回調(diào)函數(shù)，說明follow是False
        Rule(profile_links, callback='parse_item', follow=True),
    )

    def parse_item(self, response):
        item = CrawlyouyuanItem()

        item['username'] = self.get_username(response)
        # 年齡
        item['age'] = self.get_age(response)
        # 頭像圖片的鏈接
        item['header_url'] = self.get_header_url(response)
        # 相冊圖片的鏈接
        item['images_url'] = self.get_images_url(response)
        # 內(nèi)心獨(dú)白
        item['content'] = self.get_content(response)
        # 籍貫
        item['place_from'] = self.get_place_from(response)
        # 學(xué)歷
        item['education'] = self.get_education(response)
        # 興趣愛好
        item['hobby'] = self.get_hobby(response)
        # 個(gè)人主頁
        item['source_url'] = response.url
        # 數(shù)據(jù)來源網(wǎng)站
        item['sourec'] = "youyuan"

        yield item

    def get_username(self, response):
        username = response.xpath("http://dl[@class='personal_cen']//div[@class='main']/strong/text()").extract()
        if len(username):
            username = username[0]
        else:
            username = "NULL"
        return username.strip()

    def get_age(self, response):
        age = response.xpath("http://dl[@class='personal_cen']//dd/p/text()").extract()
        if len(age):
            age = re.findall(u"\d+歲", age[0])[0]
        else:
            age = "NULL"
        return age.strip()

    def get_header_url(self, response):
        header_url = response.xpath("http://dl[@class='personal_cen']/dt/img/@src").extract()
        if len(header_url):
            header_url = header_url[0]
        else:
            header_url = "NULL"
        return header_url.strip()

    def get_images_url(self, response):
        images_url = response.xpath("http://div[@class='ph_show']/ul/li/a/img/@src").extract()
        if len(images_url):
            images_url = ", ".join(images_url)
        else:
            images_url = "NULL"
        return images_url

    def get_content(self, response):
        content = response.xpath("http://div[@class='pre_data']/ul/li/p/text()").extract()
        if len(content):
            content = content[0]
        else:
            content = "NULL"
        return content.strip()

    def get_place_from(self, response):
        place_from = response.xpath("http://div[@class='pre_data']/ul/li[2]//ol[1]/li[1]/span/text()").extract()
        if len(place_from):
            place_from = place_from[0]
        else:
            place_from = "NULL"
        return place_from.strip()

    def get_education(self, response):
        education = response.xpath("http://div[@class='pre_data']/ul/li[3]//ol[2]/li[2]/span/text()").extract()
        if len(education):
            education = education[0]
        else:
            education = "NULL"
        return education.strip()

    def get_hobby(self, response):
        hobby = response.xpath("http://dl[@class='personal_cen']//ol/li/text()").extract()
        if len(hobby):
            hobby = ",".join(hobby).replace(" ", "")
        else:
            hobby = "NULL"
        return hobby.strip()

pipelines.py

管道文件
import json
import codecs
class CrawlyouyuanPipeline(object):

    def __init__(self):
        self.filename = codecs.open('content.json', 'w', encoding='utf-8')

    def process_item(self, item, spider):
        html = json.dumps(dict(item), ensure_ascii=False)
        self.filename.write(html + '\n')
        return item

    def spider_closed(self, spider):
        self.filename.close()

settings.py

BOT_NAME = 'CrawlYouYuan'
SPIDER_MODULES = ['CrawlYouYuan.spiders']
NEWSPIDER_MODULE = 'CrawlYouYuan.spiders'
# Crawl responsibly by identifying yourself (and your website) on the user-agent
USER_AGENT = 'Mozilla/5.0 (Macintosh; Intel Mac OS X 10.13; rv:56.0)'
# Obey robots.txt rules
ROBOTSTXT_OBEY = True
ITEM_PIPELINES = {
   'CrawlYouYuan.pipelines.CrawlyouyuanPipeline': 300,
}

begin.py

from scrapy import cmdline
cmdline.execute('scrapy crawl youyuan'.split())

在運(yùn)行程序之前需要使Scrapy版本和Twisted版本相吻合片拍，設(shè)置如下

scr?apy.png