前言
在之前的文章中我们都是对目标站点进行全量爬取，只要爬虫run起来就会对所有的链接都爬取一遍，这其实是很傻的做法，因为很多情况下我们并不需要爬取已经爬过的链接，除非你需要定期更新这个链接对应页面上的数据。好了，回归正题，本文介绍scrapy使用scrapy-deltafetch这个插件来实现增量爬取，这里以爬取【美食杰】上的菜谱信息为例。
正文
安装scrapy-deltafetch
$ pip install scrapy-deltafetch
如安装过程报错Can't find a local Berkeley DB installation.请参考：http://jinbitou.net/2018/01/27/2579.html
新建项目和爬虫
$ scrapy startproject meishijie PycharmProjects/meishijie
$ cd PycharmProjects/meishijie
$ scrapy genspider meishi meishij.net
settings.py
用Pycharm打开生成的项目，编辑settings.py，添加如下内容：
SPIDER_MIDDLEWARES = {
‘scrapy_deltafetch.DeltaFetch’: 100
}
DELTAFETCH_ENABLED = True
meishi.py
编辑爬虫文件meishi.py，因为是测试我就不写太多逻辑了，大家知道就好。
# -*- coding: utf-8 -*-import scrapyclass MeishiSpider(scrapy.Spider): name = 'meishi' allowed_domains = ['meishij.net'] start_urls = ['http://www.meishij.net/yaoshanshiliao/jibingtiaoli/weiyan/'] def parse(self, response): for ms in response.xpath("//div[contains(@class,'i_w')]"): item = {} title = ms.xpath("div/div/strong/text()").extract_first() hot = ms.xpath("div/div/span/text()").extract_first() item["title"] = title
item["hot"] = hot
yield item
next_page = response.xpath("//a[@class='next']/@href").extract_first() print("下一页：", next_page) if next_page: yield scrapy.Request(response.urljoin(next_page), callback=self.parse)
如上，items.py，pipelines.py，middlewares.py不做任何改动，保持默认就好。
运行
运行如下命令就可以看到scrapy已经愉快地跑起来了
$ scrapy crawl meishi
很快，所有链接已经爬取完毕，查看运行日志
2018-01-27 14:11:18 [scrapy.statscollectors] INFO: Dumping Scrapy stats:{'deltafetch/stored': 1500,'downloader/request_bytes': 25530,'downloader/request_count': 76,'downloader/request_method_count/GET': 76,'downloader/response_bytes': 1280237,'downloader/response_count': 76,'downloader/response_status_count/200': 76,'finish_reason': 'finished','finish_time': datetime.datetime(2018, 1, 27, 6, 11, 18, 772936),'item_scraped_count': 1500,'log_count/DEBUG': 1577,'log_count/INFO': 7,'memusage/max': 52551680,'memusage/startup': 52551680,'request_depth_max': 74,'response_received_count': 76,'scheduler/dequeued': 75,'scheduler/dequeued/memory': 75,'scheduler/enqueued': 75,'scheduler/enqueued/memory': 75,'start_time': datetime.datetime(2018, 1, 27, 6, 10, 55, 141746)}2018-01-27 14:11:18 [scrapy.core.engine] INFO: Spider closed (finished)
可以知道，scrapy共发起了包括入口在内的76次请求，爬取了1500个item，deltafetch存储了这1500个item对应请求的指纹信息。
测试增量爬取
再次运行scrapy crawl meishi命令
2018-01-27 14:27:39 [scrapy_deltafetch.middleware] INFO: Ignoring already visited: 2018-01-27 14:27:39 [scrapy.core.engine] INFO: Closing spider (finished)2018-01-27 14:27:39 [scrapy.statscollectors] INFO: Dumping Scrapy stats:{'deltafetch/skipped': 1,'deltafetch/stored': 20,'downloader/request_bytes': 471,'downloader/request_count': 2,'downloader/request_method_count/GET': 2,'downloader/response_bytes': 17879,'downloader/response_count': 2,'downloader/response_status_count/200': 2,'finish_reason': 'finished','finish_time': datetime.datetime(2018, 1, 27, 6, 27, 39, 940365),'item_scraped_count': 20,'log_count/DEBUG': 23,'log_count/INFO': 8,'memusage/max': 52420608,'memusage/startup': 52420608,'request_depth_max': 1,'response_received_count': 2,'scheduler/dequeued': 1,'scheduler/dequeued/memory': 1,'scheduler/enqueued': 1,'scheduler/enqueued/memory': 1,'start_time': datetime.datetime(2018, 1, 27, 6, 27, 39, 547330)}2018-01-27 14:27:39 [scrapy.core.engine] INFO: Spider closed (finished)
可以看到scrapy除了入口的请求外，之前请求过的链接都已经跳过，Done！
补充
如果想重新爬取之前已经爬取过的链接，可以通过重置DeltaFetch的缓存来实现，具体做法是给你的爬虫传一个参数deltafetch_reset，例如：
$ scrapy crawl meishi -a deltafetch_reset=1
