CIRCL
/
AIL-framework
mirror of https://github.com/CIRCL/AIL-framework
2
1
Fork 0
Code Issues Releases Wiki Activity

fix: [crawler] error catcher

pull/487/head
Terrtia 2020-04-01 14:58:27 +02:00
parent eea2e1714d
commit 179fba2ecc
No known key found for this signature in database
GPG Key ID: 1E1B1F50D84613D0
1 changed files with 4 additions and 4 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

8
bin/torcrawler/TorSplashCrawler.py
View File

@ -156,7 +156,7 @@ class TorSplashCrawler():
self.parse,
errback=self.errback_catcher,
endpoint='execute',
meta={'father': self.original_item},
meta={'father': self.original_item, 'current_url': self.start_urls},
args=l_cookies
)
@ -217,7 +217,7 @@ class TorSplashCrawler():
self.parse,
errback=self.errback_catcher,
endpoint='execute',
meta={'father': item_id},
meta={'father': item_id, 'current_url': link.url},
args=l_cookies
)
@ -227,7 +227,7 @@ class TorSplashCrawler():
if failure.check(ResponseNeverReceived):
request = failure.request
url= response.data['last_url']
url= request.meta['current_url']
father = request.meta['father']
self.logger.error('Splash, ResponseNeverReceived for %s, retry in 10s ...', url)
@ -242,7 +242,7 @@ class TorSplashCrawler():
errback=self.errback_catcher,
endpoint='execute',
cache_args=['lua_source'],
meta={'father': father},
meta={'father': father, 'current_url': url},
args=self.build_request_arg(response.cookiejar)
)