BookDataModel: implement merge_into method

bookwyrm/management/commands/deduplicate_book_data.py

@@ -1,9 +1,9 @@
 """ PROCEED WITH CAUTION: uses deduplication fields to permanently
 merge book data objects """
+
 from django.core.management.base import BaseCommand
 from django.db.models import Count
 from bookwyrm import models
-from bookwyrm.management.merge import merge_objects
 
 
 def dedupe_model(model):
@@ -16,7 +16,8 @@ def dedupe_model(model):
         dupes = (
             model.objects.values(field.name)
             .annotate(Count(field.name))
-            .filter(**{"%s__count__gt" % field.name: 1})
+            .filter(merged_into__isnull=True, **{"%s__count__gt" % field.name: 1})
+            .exclude(**{field.name: "", "%s__isnull" % field.name: True})
         )
 
         for dupe in dupes:
@@ -30,13 +31,14 @@ def dedupe_model(model):
             print("keeping", canonical.remote_id)
             for obj in objs[1:]:
                 print(obj.remote_id)
-                merge_objects(canonical, obj)
+                obj.merge_into(canonical)
 
 
 class Command(BaseCommand):
     """deduplicate allllll the book data models"""
 
     help = "merges duplicate book data"
+
     # pylint: disable=no-self-use,unused-argument
     def handle(self, *args, **options):
         """run deduplications"""

bookwyrm/management/merge.py

@@ -1,50 +0,0 @@
-from django.db.models import ManyToManyField
-
-
-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:
-        # Skip the ManyToMany fields that aren’t auto-created. These
-        # should have a corresponding OneToMany field in the model for
-        # the linking table anyway. If we update it through that model
-        # instead then we won’t lose the extra fields in the linking
-        # table.
-        related_field_obj = related_model._meta.get_field(related_field)
-        if isinstance(related_field_obj, ManyToManyField):
-            through = related_field_obj.remote_field.through
-            if not through._meta.auto_created:
-                continue
-        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 merge_objects(canonical, obj):
-    copy_data(canonical, obj)
-    update_related(canonical, obj)
-    # remove the outdated entry
-    obj.delete()

bookwyrm/management/merge_command.py

@@ -1,4 +1,3 @@
-from bookwyrm.management.merge import merge_objects
 from django.core.management.base import BaseCommand
 
 
@@ -26,4 +25,4 @@ class MergeCommand(BaseCommand):
             print("other book doesn’t exist!")
             return
 
-        merge_objects(canonical, other)
+        other.merge_into(canonical)

bookwyrm/models/book.py

@@ -3,12 +3,13 @@
 from itertools import chain
 import re
 from typing import Any
+from typing_extensions import Self
 
 from django.contrib.postgres.search import SearchVectorField
 from django.contrib.postgres.indexes import GinIndex
 from django.core.cache import cache
 from django.db import models, transaction
-from django.db.models import Prefetch
+from django.db.models import Prefetch, ManyToManyField
 from django.dispatch import receiver
 from django.utils.translation import gettext_lazy as _
 from model_utils import FieldTracker
@@ -112,6 +113,58 @@ class BookDataModel(ObjectMixin, BookWyrmModel):
         """only send book data updates to other bookwyrm instances"""
         super().broadcast(activity, sender, software=software, **kwargs)
 
+    def merge_into(self, canonical: Self) -> None:
+        """merge this entity into another entity"""
+        if canonical.id == self.id:
+            raise ValueError(f"Cannot merge {self} into itself")
+        if canonical.merged_into:
+            raise ValueError(
+                f"Cannot merge {self} into {canonical} because "
+                f"{canonical} has itself been merged into {canonical.merged_into}"
+            )
+        canonical.adopt_data_from(self)
+        canonical.save()
+
+        # move related models to canonical
+        related_models = [
+            (r.remote_field.name, r.related_model) for r in self._meta.related_objects
+        ]
+        # pylint: disable=protected-access
+        for related_field, related_model in related_models:
+            # Skip the ManyToMany fields that aren’t auto-created. These
+            # should have a corresponding OneToMany field in the model for
+            # the linking table anyway. If we update it through that model
+            # instead then we won’t lose the extra fields in the linking
+            # table.
+            # pylint: disable=protected-access
+            related_field_obj = related_model._meta.get_field(related_field)
+            if isinstance(related_field_obj, ManyToManyField):
+                through = related_field_obj.remote_field.through
+                if not through._meta.auto_created:
+                    continue
+            related_objs = related_model.objects.filter(**{related_field: self})
+            for related_obj in related_objs:
+                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(self)
+
+        self.merged_into = canonical
+        self.save()
+
+    def adopt_data_from(self, other: Self) -> None:
+        """fill empty fields with values from another entity"""
+        for data_field in self._meta.get_fields():
+            if not hasattr(data_field, "activitypub_field"):
+                continue
+            data_value = getattr(other, data_field.name)
+            if not data_value:
+                continue
+            if not getattr(self, data_field.name):
+                setattr(self, data_field.name, data_value)
+
 
 class Book(BookDataModel):
     """a generic book, which can mean either an edition or a work"""