diff options
| author | Étienne Loks <etienne.loks@iggdrasil.net> | 2019-08-20 11:52:02 +0200 |
|---|---|---|
| committer | Étienne Loks <etienne.loks@iggdrasil.net> | 2019-08-20 11:52:02 +0200 |
| commit | a465ac1fa4271e26bdc659aff32bb9ebeaae7922 (patch) | |
| tree | ef091e8bae6c6c69677e0b869e907c552824c40d /commcrawler/scrapy.py | |
| parent | 01d72fa7c94359049e2a7beb068167cb7f047805 (diff) | |
| download | Comm-on-net-a465ac1fa4271e26bdc659aff32bb9ebeaae7922.tar.bz2 Comm-on-net-a465ac1fa4271e26bdc659aff32bb9ebeaae7922.zip | |
Slice crawls - add a crontab example
Diffstat (limited to 'commcrawler/scrapy.py')
| -rw-r--r-- | commcrawler/scrapy.py | 5 |
1 files changed, 5 insertions, 0 deletions
diff --git a/commcrawler/scrapy.py b/commcrawler/scrapy.py index 490142c..6af1c0b 100644 --- a/commcrawler/scrapy.py +++ b/commcrawler/scrapy.py @@ -435,6 +435,11 @@ def launch_crawl(crawl_item, excluded_domains=None): process.start(stop_after_crawl=ONLY_FIRST_PAGE) page += 1 crawl_item.crawl_ended = timezone.now() + if ONLY_FIRST_PAGE and page <= page_number: + crawl_item.status = "A" + crawl_item.save() + return + crawl_item.status = "M" crawl_item.save() launch_match(crawl_item) |