diff options
Diffstat (limited to 'crawler/nevrax/spiders/scrape.py')
| -rw-r--r-- | crawler/nevrax/spiders/scrape.py | 4 |
1 files changed, 3 insertions, 1 deletions
diff --git a/crawler/nevrax/spiders/scrape.py b/crawler/nevrax/spiders/scrape.py index 785ec3f..d27aecf 100644 --- a/crawler/nevrax/spiders/scrape.py +++ b/crawler/nevrax/spiders/scrape.py @@ -5,6 +5,7 @@ from scrapy.linkextractors import LinkExtractor from scrapy import Selector import config +import datetime class NevraxSpider(CrawlSpider): name = "nevrax" @@ -42,5 +43,6 @@ class NevraxSpider(CrawlSpider): 'url': response.url, 'title': response.css('title::text').extract_first(), 'content': ''.join(sel.select("//body//text()").extract()).strip(), - 'content_length': len(response.body) + 'content_length': len(response.body), + 'date_updated': datetime.datetime.now() } |