How to limit number of followed pages per site in Python Scrapy

Question

I am trying to build a spider that could efficiently scrape text information from many websites. Since I am a Python user I was referred to Scrapy. However, in order to avoid scraping huge websites, I want to limit the spider to scrape no more than 20 pages of a certain "depth" per website. Here is my spider:

class DownloadSpider(CrawlSpider):
    name = 'downloader'
    download_path = '/home/MyProjects/crawler'
    rules = (Rule(SgmlLinkExtractor(), callback='parse_item', follow=True),)

    def __init__(self, *args, **kwargs):
        super(DownloadSpider, self).__init__(*args, **kwargs)
        self.urls_file_path = [kwargs.get('urls_file')]
        data = open(self.urls_file_path[0], 'r').readlines()
        self.allowed_domains = [urlparse(i).hostname.strip() for i in data] 
        self.start_urls = ['http://' + domain for domain in self.allowed_domains]

    def parse_start_url(self, response):
        return self.parse_item(response)

    def parse_item(self, response):
        self.fname = self.download_path + urlparse(response.url).hostname.strip()
        open(str(self.fname)+ '.txt', 'a').write(response.url)
        open(str(self.fname)+ '.txt', 'a').write('\n')

urls_file is a path to a text file with urls. I have also set the max depth in the settings file. Here is my problem: if I set the CLOSESPIDER_PAGECOUNT exception it closes the spider when the total number of scraped pages (regardless for which site) reaches the exception value. However, I need to stop scraping when I have scraped say 20 pages from each url. I also tried keeping count with a variable like self.parsed_number += 1, but this didn't work either -- it seems that scrapy doesn't go url by url but mixes them up. Any advice is much appreciated !

Answer 1

To do this you can create your own link extractor class based on SgmlLinkExtractor. It should look something like this:

from scrapy.selector import Selector
from scrapy.utils.response import get_base_url

from scrapy.contrib.linkextractors.sgml import SgmlLinkExtractor

class LimitedLinkExtractor(SgmlLinkExtractor):
    def __init__(self, allow=(), deny=(), allow_domains=(), deny_domains=(), restrict_xpaths=(),
                 tags=('a', 'area'), attrs=('href'), canonicalize=True, unique=True, process_value=None,
                 deny_extensions=None, max_pages=20):
        self.max_pages=max_pages

        SgmlLinkExtractor.__init__(self, allow=allow, deny=deny, allow_domains=allow_domains, deny_domains=deny_domains, restrict_xpaths=restrict_xpaths,
                 tags=tags, attrs=attrs, canonicalize=canonicalize, unique=unique, process_value=process_value,
                 deny_extensions=deny_extensions)

    def extract_links(self, response):
        base_url = None
        if self.restrict_xpaths:
            sel = Selector(response)
            base_url = get_base_url(response)
            body = u''.join(f
                            for x in self.restrict_xpaths
                            for f in sel.xpath(x).extract()
                            ).encode(response.encoding, errors='xmlcharrefreplace')
        else:
            body = response.body

        links = self._extract_links(body, response.url, response.encoding, base_url)
        links = self._process_links(links)
        links = links[0:self.max_pages]
        return links

The code of this subclass completely based on the code of the class SgmlLinkExtractor. I've just added variable self.max_pages to the class constructor and line which cut the list of links in the end of extract_links method. But you can cut this list in more intelligent way.

Answer 2

I'd make per-class variable, initialize it with stats = defaultdict(int) and increment self.stats[response.url] (or may be the key could be a tuple like (website, depth) in your case) in parse_item.

This is how I imagine this - should work in theory. Let me know if you need an example.

FYI, you can extract base url and calculate depth with the help of urlparse.urlparse (see docs).