mirrors/ArchiveBox
1
0
Fork 0
mirror of https://github.com/ArchiveBox/ArchiveBox.git synced 2025-05-17 08:34:26 -04:00
Code Issues Projects Releases Packages Wiki Activity Actions

improve latest title logic to take longest title

Browse source
This commit is contained in:
Nick Sweeting 2021-02-18 02:33:28 -05:00
parent ad4bd49163
commit f1823381d0
1 changed files with 21 additions and 8 deletions
Download patch file Download diff file Expand all files Collapse all files

29
archivebox/core/models.py
View file

@ -170,18 +170,31 @@ class Snapshot(models.Model):
@cached_property
def latest_title(self):
if self.title:
return self.title
return self.title # whoopdedoo that was easy
try:
return self.archiveresult_set.filter(extractor='title', status='succeeded')[0].output
except ArchiveResult.DoesNotExist:
# take longest successful title from ArchiveResult db history
return sorted(
self.archiveresult_set\
.filter(extractor='title', status='succeeded', output__isnull=False)\
.values_list('output', flat=True),
key=lambda r: len(r),
)[-1]
except IndexError:
pass
if ('title' in self.history
and self.history['title']
and (self.history['title'][-1].status == 'succeeded')
and self.history['title'][-1].output.strip()):
return self.history['title'][-1].output.strip()
try:
# take longest successful title from Link json index file history
return sorted(
(
result.output.strip()
for result in self.history['title']
if result.status == 'succeeded' and result.output.strip()
),
key=lambda r: len(r),
)[-1]
except (KeyError, IndexError):
pass
return None