Skip to content

Commit 36d9894

Browse files
authored
Merge pull request #471 from WikipediaLibrary/kgraessle/T404879
Backfill missing Wall Street Journal aggregate data
2 parents e53ff4c + 606cb60 commit 36d9894

File tree

5 files changed

+3440
-3
lines changed

5 files changed

+3440
-3
lines changed

extlinks/aggregates/management/commands/fill_link_aggregates.py

Lines changed: 3 additions & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -8,7 +8,7 @@
88
from django.db.models.fields import DateField
99

1010
from ...models import LinkAggregate
11-
from extlinks.links.models import LinkEvent
11+
from extlinks.links.models import LinkEvent, URLPattern
1212
from extlinks.organisations.models import Collection
1313

1414

@@ -117,6 +117,8 @@ def _process_single_collection(self, link_event_filter, collection):
117117
None
118118
"""
119119
url_patterns = collection.get_url_patterns()
120+
if len(url_patterns) == 0:
121+
url_patterns = URLPattern.objects.filter(collection=collection).all()
120122
for url_pattern in url_patterns:
121123
link_events_with_annotated_timestamp = url_pattern.link_events.annotate(
122124
timestamp_date=Cast("timestamp", DateField())

extlinks/aggregates/management/commands/fill_pageproject_aggregates.py

Lines changed: 3 additions & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -9,7 +9,7 @@
99
from django.db.models.fields import DateField
1010

1111
from ...models import PageProjectAggregate
12-
from extlinks.links.models import LinkEvent
12+
from extlinks.links.models import LinkEvent, URLPattern
1313
from extlinks.organisations.models import Collection
1414

1515
logger = logging.getLogger("django")
@@ -118,6 +118,8 @@ def _process_single_collection(self, link_event_filter, collection):
118118
None
119119
"""
120120
url_patterns = collection.get_url_patterns()
121+
if len(url_patterns) == 0:
122+
url_patterns = URLPattern.objects.filter(collection=collection).all()
121123
for url_pattern in url_patterns:
122124
link_events_with_annotated_timestamp = url_pattern.link_events.annotate(
123125
timestamp_date=Cast("timestamp", DateField())

extlinks/aggregates/management/commands/fill_user_aggregates.py

Lines changed: 3 additions & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -8,7 +8,7 @@
88
from django.db.models.fields import DateField
99

1010
from ...models import UserAggregate
11-
from extlinks.links.models import LinkEvent
11+
from extlinks.links.models import LinkEvent, URLPattern
1212
from extlinks.organisations.models import Collection
1313

1414

@@ -117,6 +117,8 @@ def _process_single_collection(self, link_event_filter, collection):
117117
None
118118
"""
119119
url_patterns = collection.get_url_patterns()
120+
if len(url_patterns) == 0:
121+
url_patterns = URLPattern.objects.filter(collection=collection).all()
120122
for url_pattern in url_patterns:
121123
link_events_with_annotated_timestamp = url_pattern.link_events.annotate(
122124
timestamp_date=Cast("timestamp", DateField())

0 commit comments

Comments
 (0)