我們在寫普通腳本的時候，從一個網站拿到一個文件的下載url，然後下載，直接將數據寫入文件或者保存下來，但是這個需要我們自己一點一點的寫出來，而且反復利用率並不高，為了不重復造輪子，scrapy提供很流暢的下載文件方式，只需要隨便寫寫便可用了。

mat.py文件

 1 # -*- coding: utf-8 -*-
 2 import scrapy
 3 from scrapy.linkextractor import LinkExtractor
 4 from weidashang.items import matplotlib
 5 
 6 class MatSpider(scrapy.Spider):
 
 7     name = "mat"
 8     allowed_domains = ["matplotlib.org"]
 9     start_urls = [‘https://matplotlib.org/examples‘]
10 
11     def parse(self, response):
　　　　　　　#抓取每個腳本文件的訪問頁面，拿到後下載
12         link = LinkExtractor(restrict_css=‘div.toctree-wrapper.compound li.toctree-l2‘)
13         for link in link.extract_links(response):
 
14             yield scrapy.Request(url=link.url,callback=self.example)
15 
16     def example(self,response):
　　　　　　#進入每個腳本的頁面，抓取源碼文件按鈕，並和base_url結合起來形成一個完整的url
17         href = response.css(‘a.reference.external::attr(href)‘).extract_first()
18         url = response.urljoin(href)
19         example = matplotlib()
 
20         example[‘file_urls‘] = [url]
21         return example

pipelines.py

1 class MyFilePlipeline(FilesPipeline):
2     def file_path(self, request, response=None, info=None):
3         path = urlparse(request.url).path
4         return join(basename(dirname(path)),basename(path))

settings.py

1 ITEM_PIPELINES = {
2    ‘weidashang.pipelines.MyFilePlipeline‘: 1,
3 }
4 FILES_STORE = ‘examples_src‘

items.py

class matplotlib(Item):
    file_urls = Field()
    files = Field()

run.py

1 from scrapy.cmdline import execute
2 execute([‘scrapy‘, ‘crawl‘, ‘mat‘,‘-o‘,‘example.json‘])

python爬蟲之scrapy文件下載