mirrors/bookwyrm
1
0
Fork 1
mirror of https://github.com/bookwyrm-social/bookwyrm.git synced 2025-01-16 04:05:38 +00:00
Code Issues Projects Releases Wiki Activity

move some code around for readability

Browse source
This commit is contained in:
Mouse Reeve 2020-12-31 16:30:04 -08:00
parent c5a3e57bb4
commit 1246042c12
1 changed files with 45 additions and 30 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

75
bookwyrm/management/commands/deduplicate_book_data.py
View file

@ -5,6 +5,44 @@ from django.db.models import Count
from bookwyrm import models
def update_related(canonical, obj):
''' update all the models with fk to the object being removed '''
# move related models to canonical
related_models = [
(r.remote_field.name, r.related_model) for r in \
canonical._meta.related_objects]
for (related_field, related_model) in related_models:
related_objs = related_model.objects.filter(
**{related_field: obj})
for related_obj in related_objs:
print(
'replacing in',
related_model.__name__,
related_field,
related_obj.id
)
try:
setattr(related_obj, related_field, canonical)
related_obj.save()
except TypeError:
getattr(related_obj, related_field).add(canonical)
getattr(related_obj, related_field).remove(obj)
def copy_data(canonical, obj):
''' try to get the most data possible '''
for data_field in obj._meta.get_fields():
if not hasattr(data_field, 'activitypub_field'):
continue
data_value = getattr(obj, data_field.name)
if not data_value:
continue
if not getattr(canonical, data_field.name):
print('setting data field', data_field.name, data_value)
setattr(canonical, data_field.name, data_value)
canonical.save()
def dedupe_model(model):
''' combine duplicate editions and update related models '''
fields = model._meta.get_fields()
@ -25,42 +63,19 @@ def dedupe_model(model):
**{field.name: value}
).order_by('id')
canonical = objs.first()
print('keeping', canonical.remote_id, canonical.id)
print('keeping', canonical.remote_id)
for obj in objs[1:]:
print(obj.remote_id, obj.id)
# try to get the most data possible
for data_field in obj._meta.get_fields():
if not hasattr(data_field, 'activitypub_field'):
continue
data_value = getattr(obj, data_field.name)
if not data_value:
continue
if not getattr(canonical, data_field.name):
print('setting data field', data_field.name, data_value)
setattr(canonical, data_field.name, data_value)
canonical.save()
# move related models to canonical
related_models = [
(r.remote_field.name, r.related_model) for r in \
canonical._meta.related_objects]
for (related_field, related_model) in related_models:
related_objs = related_model.objects.filter(
**{related_field: obj})
for related_obj in related_objs:
print(
'replacing in', related_model.__name__, related_obj)
try:
setattr(related_obj, related_field, canonical)
related_obj.save()
except TypeError:
getattr(related_obj, related_field).add(canonical)
getattr(related_obj, related_field).remove(obj)
print(obj.remote_id)
copy_data(canonical, obj)
update_related(canonical, obj)
# remove the outdated entry
obj.delete()
class Command(BaseCommand):
''' dedplucate allllll the book data models '''
help = 'merges duplicate book data'
# pylint: disable=no-self-use,unused-argument
def handle(self, *args, **options):
''' run deudplications '''
dedupe_model(models.Edition)