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

chg: [importer url_extract] fix item id

pull/497/head
Terrtia 2020-04-30 09:57:26 +02:00
parent 3d4131a193
commit eddeb0b3b0
No known key found for this signature in database
GPG Key ID: 1E1B1F50D84613D0
1 changed files with 4 additions and 1 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

5
bin/import/ail_json_importer/Ail_feeder_url_extract.py
View File

@ -32,7 +32,10 @@ class Ail_feeder_twitter(Default_json):
# use twitter timestamp ? # use twitter timestamp ?
item_date = datetime.date.today().strftime("%Y/%m/%d") item_date = datetime.date.today().strftime("%Y/%m/%d")
item_id = str(self.json_item['meta']['twitter:url-extracted']) item_id = str(self.json_item['meta']['twitter:url-extracted'])
item_id = '{}{}.gz'.format(file_name_paste[:215], str(uuid.uuid4())) if len(item_id) > 215:
item_id = '{}{}.gz'.format(item_id[:215], str(uuid.uuid4()))
else:
item_id = '{}{}.gz'.format(item_id, str(uuid.uuid4()))
return os.path.join('urlextract', item_date, item_id) + '.gz' return os.path.join('urlextract', item_date, item_id) + '.gz'
# # TODO: # # TODO: