remove .old_id entirely and make ABID generation only happen once on initial save

archivebox/abid_utils/admin.py

@@ -69,7 +69,6 @@ def get_abid_info(self, obj, request=None):
             &nbsp; &nbsp; SUBTYPE: &nbsp; &nbsp; &nbsp; <code style="font-size: 10px;"><b style="user-select: all">{}</b> &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; {}</code> &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; <code style="font-size: 10px;"><b>{}</b></code> {}: <code style="user-select: all">{}</code><br/>
             &nbsp; &nbsp; RAND: &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; <code style="font-size: 10px;"><b style="user-select: all">{}</b> &nbsp; &nbsp; &nbsp; {}</code> &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; <code style="font-size: 10px;"><b>{}</b></code> {}: <code style="user-select: all">{}</code></code>
             <br/><hr/>
-            &nbsp; &nbsp; <small style="opacity: 0.5">.old_id: &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp;<code style="font-size: 10px; user-select: all">{}</code></small><br/>
             </div>
             ''',
             obj.api_url + (f'?api_key={get_or_create_api_token(request.user)}' if request and request.user else ''), obj.api_url, obj.api_docs_url,
@@ -82,7 +81,6 @@ def get_abid_info(self, obj, request=None):
             highlight_diff(obj.ABID.uri, derived_uri), highlight_diff(str(obj.ABID.uuid)[14:26], str(fresh_abid.uuid)[14:26]), mark_safe(uri_diff), obj.abid_uri_src, str(obj.ABID_FRESH_VALUES['uri']),
             highlight_diff(obj.ABID.subtype, derived_subtype), highlight_diff(str(obj.ABID.uuid)[26:28], str(fresh_abid.uuid)[26:28]), mark_safe(subtype_diff), obj.abid_subtype_src, str(obj.ABID_FRESH_VALUES['subtype']),
             highlight_diff(obj.ABID.rand, derived_rand), highlight_diff(str(obj.ABID.uuid)[28:36], str(fresh_abid.uuid)[28:36]), mark_safe(rand_diff), obj.abid_rand_src, str(obj.ABID_FRESH_VALUES['rand'])[-7:],
-            highlight_diff(getattr(obj, 'old_id', ''), obj.pk),
         )
     except Exception as e:
         return str(e)

archivebox/abid_utils/models.py

@@ -73,27 +73,30 @@ class ABIDModel(models.Model):
     """
     Abstract Base Model for other models to depend on. Provides ArchiveBox ID (ABID) interface.
     """
-    abid_prefix: str = DEFAULT_ABID_PREFIX   # e.g. 'tag_'
+    abid_prefix: str = DEFAULT_ABID_PREFIX            # e.g. 'tag_'
-    abid_ts_src = 'None'                    # e.g. 'self.created'
+    abid_ts_src = 'self.created'                     # e.g. 'self.created'
-    abid_uri_src = 'None'                   # e.g. 'self.uri'
+    abid_uri_src = 'None'                            # e.g. 'self.uri'
-    abid_subtype_src = 'None'               # e.g. 'self.extractor'
+    abid_subtype_src = 'self.__class__.__name__'     # e.g. 'self.extractor'
-    abid_rand_src = 'None'                  # e.g. 'self.uuid' or 'self.id'
+    abid_rand_src = 'self.id'                        # e.g. 'self.uuid' or 'self.id'
     abid_salt: str = DEFAULT_ABID_URI_SALT
 
-    # id = models.UUIDField(primary_key=True, default=uuid4, editable=True)
+    # id = models.UUIDField(primary_key=True, default=None, null=False, editable=False, unique=True, verbose_name='ID')
-    # uuid = models.UUIDField(blank=True, null=True, editable=True, unique=True)
+    # abid = ABIDField(prefix=abid_prefix)
-    abid = ABIDField(prefix=abid_prefix)
 
-    created_by = models.ForeignKey(settings.AUTH_USER_MODEL, on_delete=models.CASCADE, default=get_or_create_system_user_pk)
+    # created_by = models.ForeignKey(settings.AUTH_USER_MODEL, on_delete=models.CASCADE, default=None, null=False)
-    created = AutoDateTimeField(default=None, null=False, db_index=True)
+    # created = AutoDateTimeField(default=None, null=False, db_index=True)
-    modified = models.DateTimeField(auto_now=True)
+    # modified = models.DateTimeField(auto_now=True)
 
     class Meta(TypedModelMeta):
         abstract = True
 
     def save(self, *args: Any, **kwargs: Any) -> None:
         if self._state.adding:
-            self.issue_new_abid()
+            self.pk = self.id = self.id or uuid4()
+            self.created = ts_from_abid(abid_part_from_ts(timezone.now()))  # cut off precision to match precision of TS component
+            self.modified = self.created
+            self.created_by = self.created_by or get_or_create_system_user_pk()
+            self.abid = str(self.issue_new_abid())
         return super().save(*args, **kwargs)
 
         # assert str(self.id) == str(self.ABID.uuid), f'self.id {self.id} does not match self.ABID {self.ABID.uuid}'
@@ -119,48 +122,28 @@ class ABIDModel(models.Model):
     def ABID_FRESH_HASHES(self) -> Dict[str, str]:
         return abid_hashes_from_values(**self.ABID_FRESH_VALUES)
 
+    def issue_new_abid(self):
+        assert self.abid is None, f'Can only issue new ABID for new objects that dont already have one {self.abid}'
+        assert self._state.adding, 'Can only issue new ABID when model._state.adding is True'
+        assert eval(self.abid_uri_src), f'Can only issue new ABID if self.abid_uri_src is defined ({self.abid_uri_src}={eval(self.abid_uri_src)})'
 
-    @property
+        self.abid = None
-    def ABID_FRESH(self) -> ABID:
+        self.pk = self.id = self.id or uuid4()
-        """
+        self.created = ts_from_abid(abid_part_from_ts(timezone.now()))  # cut off precision to match precision of TS component
-        Return a pure freshly derived ABID (assembled from attrs defined in ABIDModel.abid_*_src).
-        """
 
         abid_fresh_values = self.ABID_FRESH_VALUES
         assert all(abid_fresh_values.values()), f'All ABID_FRESH_VALUES must be set {abid_fresh_values}'
         abid_fresh_hashes = self.ABID_FRESH_HASHES
         assert all(abid_fresh_hashes.values()), f'All ABID_FRESH_HASHES must be able to be generated {abid_fresh_hashes}'
         
-        abid = ABID(**abid_fresh_hashes)
+        new_abid = ABID(**abid_fresh_hashes)
         
-        assert abid.ulid and abid.uuid and abid.typeid, f'Failed to calculate {abid_fresh_values["prefix"]}_ABID for {self.__class__.__name__}'
+        assert new_abid.ulid and new_abid.uuid and new_abid.typeid, f'Failed to calculate {abid_fresh_values["prefix"]}_ABID for {self.__class__.__name__}'
-        return abid
-
-
-    def issue_new_abid(self):
-        assert self.abid is None, f'Can only issue new ABID for new objects that dont already have one {self.abid}'
-        assert self._state.adding, 'Can only issue new ABID when model._state.adding is True'
-        assert eval(self.abid_uri_src), f'Can only issue new ABID if self.abid_uri_src is defined ({self.abid_uri_src}={eval(self.abid_uri_src)})'
-
-        self.old_id = getattr(self, 'old_id', None) or self.id or uuid4()
-        self.abid = None
-        self.created = ts_from_abid(abid_part_from_ts(getattr(self, 'bookmarked', None) or timezone.now()))  # cut off precision to match precision of TS component
-        self.added = getattr(self, 'added', None) or self.created
-        self.modified = self.created
-
-        assert all(self.ABID_FRESH_VALUES.values()), f'Can only issue new ABID if all self.ABID_FRESH_VALUES are defined {self.ABID_FRESH_VALUES}'
-
-        new_abid = self.ABID_FRESH
 
         # store stable ABID on local fields, overwrite them because we are adding a new entry and existing defaults havent touched db yet
         self.abid = str(new_abid)
-        self.id = new_abid.uuid
+        assert str(self.ABID.uuid) == str(new_abid.uuid)
-        self.pk = new_abid.uuid
+        return new_abid
-
-        assert self.ABID == new_abid
-        assert str(self.ABID.uuid) == str(self.id) == str(self.pk) == str(ABID.parse(self.abid).uuid)
-        
-        self._ready_to_save_as_new = True
 
 
     @property
@@ -169,27 +152,12 @@ class ABIDModel(models.Model):
         aka get_or_generate_abid -> ULIDParts(timestamp='01HX9FPYTR', url='E4A5CCD9', subtype='00', randomness='ZYEBQE')
         """
 
-        # otherwise DB is single source of truth, load ABID from existing db pk
+        if not self.abid:
-        abid: ABID | None = None
+            pre_save_abid = self.issue_new_abid()
-        try:
+            self.abid = str(pre_save_abid)
-            abid = abid or ABID.parse(cast(str, self.abid))
+            return pre_save_abid
-        except Exception:
-            pass
-
-        try:
-            abid = abid or ABID.parse(cast(str, self.id))
-        except Exception:
-            pass
-
-        try:
-            abid = abid or ABID.parse(cast(str, self.pk))
-        except Exception:
-            pass
-
-        abid = abid or self.ABID_FRESH
-
-        return abid
 
+        return ABID.parse(cast(str, self.abid))
 
     @property
     def ULID(self) -> ULID:

archivebox/api/migrations/0008_alter_apitoken_created_alter_apitoken_created_by_and_more.py

@@ -0,0 +1,47 @@
+# Generated by Django 5.1 on 2024-09-04 23:32
+
+import abid_utils.models
+import django.db.models.deletion
+from django.conf import settings
+from django.db import migrations, models
+
+
+class Migration(migrations.Migration):
+
+    dependencies = [
+        ('api', '0007_alter_apitoken_created_by'),
+        migrations.swappable_dependency(settings.AUTH_USER_MODEL),
+    ]
+
+    operations = [
+        migrations.AlterField(
+            model_name='apitoken',
+            name='created',
+            field=abid_utils.models.AutoDateTimeField(db_index=True, default=None),
+        ),
+        migrations.AlterField(
+            model_name='apitoken',
+            name='created_by',
+            field=models.ForeignKey(default=None, on_delete=django.db.models.deletion.CASCADE, to=settings.AUTH_USER_MODEL),
+        ),
+        migrations.AlterField(
+            model_name='apitoken',
+            name='id',
+            field=models.UUIDField(default=None, editable=False, primary_key=True, serialize=False, unique=True, verbose_name='ID'),
+        ),
+        migrations.AlterField(
+            model_name='outboundwebhook',
+            name='created',
+            field=abid_utils.models.AutoDateTimeField(db_index=True, default=None),
+        ),
+        migrations.AlterField(
+            model_name='outboundwebhook',
+            name='created_by',
+            field=models.ForeignKey(default=None, on_delete=django.db.models.deletion.CASCADE, to=settings.AUTH_USER_MODEL),
+        ),
+        migrations.AlterField(
+            model_name='outboundwebhook',
+            name='id',
+            field=models.UUIDField(default=None, editable=False, primary_key=True, serialize=False, unique=True, verbose_name='ID'),
+        ),
+    ]

archivebox/api/models.py

@@ -12,7 +12,7 @@ from signal_webhooks.models import WebhookBase
 
 from django_stubs_ext.db.models import TypedModelMeta
 
-from abid_utils.models import ABIDModel, ABIDField, get_or_create_system_user_pk
+from abid_utils.models import ABIDModel, ABIDField, AutoDateTimeField
 
 
 
@@ -32,11 +32,12 @@ class APIToken(ABIDModel):
     abid_subtype_src = 'self.created_by_id'
     abid_rand_src = 'self.id'
 
-    id = models.UUIDField(primary_key=True, default=uuid.uuid4, editable=False)
+    id = models.UUIDField(primary_key=True, default=None, null=False, editable=False, unique=True, verbose_name='ID')
     abid = ABIDField(prefix=abid_prefix)
 
-    created_by = models.ForeignKey(settings.AUTH_USER_MODEL, on_delete=models.CASCADE, default=get_or_create_system_user_pk)
+    created_by = models.ForeignKey(settings.AUTH_USER_MODEL, on_delete=models.CASCADE, default=None, null=False)
-    created = models.DateTimeField(auto_now_add=True)
+    created = AutoDateTimeField(default=None, null=False, db_index=True)
+    modified = models.DateTimeField(auto_now=True)
 
     token = models.CharField(max_length=32, default=generate_secret_token, unique=True)
     expires = models.DateTimeField(null=True, blank=True)
@@ -99,11 +100,11 @@ class OutboundWebhook(ABIDModel, WebhookBase):
     abid_subtype_src = 'self.ref'
     abid_rand_src = 'self.id'
 
-    id = models.UUIDField(primary_key=True, default=uuid.uuid4, editable=False)
+    id = models.UUIDField(primary_key=True, default=None, null=False, editable=False, unique=True, verbose_name='ID')
     abid = ABIDField(prefix=abid_prefix)
 
-    created_by = models.ForeignKey(settings.AUTH_USER_MODEL, on_delete=models.CASCADE, default=get_or_create_system_user_pk)
+    created_by = models.ForeignKey(settings.AUTH_USER_MODEL, on_delete=models.CASCADE, default=None, null=False)
-    created = models.DateTimeField(auto_now_add=True)
+    created = AutoDateTimeField(default=None, null=False, db_index=True)
     modified = models.DateTimeField(auto_now=True)
 
     # More fields here: WebhookBase...

archivebox/api/v1_core.py

@@ -64,7 +64,6 @@ class ArchiveResultSchema(Schema):
     TYPE: str = 'core.models.ArchiveResult'
 
     id: UUID
-    old_id: int
     abid: str
 
     modified: datetime
@@ -127,9 +126,9 @@ class ArchiveResultSchema(Schema):
 
 
 class ArchiveResultFilterSchema(FilterSchema):
-    id: Optional[str] = Field(None, q=['id__startswith', 'abid__icontains', 'old_id__startswith', 'snapshot__id__startswith', 'snapshot__abid__icontains', 'snapshot__timestamp__startswith'])
+    id: Optional[str] = Field(None, q=['id__startswith', 'abid__icontains', 'snapshot__id__startswith', 'snapshot__abid__icontains', 'snapshot__timestamp__startswith'])
 
-    search: Optional[str] = Field(None, q=['snapshot__url__icontains', 'snapshot__title__icontains', 'snapshot__tags__name__icontains', 'extractor', 'output__icontains', 'id__startswith', 'abid__icontains', 'old_id__startswith', 'snapshot__id__startswith', 'snapshot__abid__icontains', 'snapshot__timestamp__startswith'])
+    search: Optional[str] = Field(None, q=['snapshot__url__icontains', 'snapshot__title__icontains', 'snapshot__tags__name__icontains', 'extractor', 'output__icontains', 'id__startswith', 'abid__icontains', 'snapshot__id__startswith', 'snapshot__abid__icontains', 'snapshot__timestamp__startswith'])
     snapshot_id: Optional[str] = Field(None, q=['snapshot__id__startswith', 'snapshot__abid__icontains', 'snapshot__timestamp__startswith'])
     snapshot_url: Optional[str] = Field(None, q='snapshot__url__icontains')
     snapshot_tag: Optional[str] = Field(None, q='snapshot__tags__name__icontains')
@@ -157,8 +156,8 @@ def get_archiveresults(request, filters: ArchiveResultFilterSchema = Query(...))
 
 @router.get("/archiveresult/{archiveresult_id}", response=ArchiveResultSchema, url_name="get_archiveresult")
 def get_archiveresult(request, archiveresult_id: str):
-    """Get a specific ArchiveResult by pk, abid, or old_id."""
+    """Get a specific ArchiveResult by id or abid."""
-    return ArchiveResult.objects.get(Q(id__icontains=archiveresult_id) | Q(abid__icontains=archiveresult_id) | Q(old_id__icontains=archiveresult_id))
+    return ArchiveResult.objects.get(Q(id__icontains=archiveresult_id) | Q(abid__icontains=archiveresult_id))
 
 
 # @router.post("/archiveresult", response=ArchiveResultSchema)
@@ -193,7 +192,6 @@ class SnapshotSchema(Schema):
     TYPE: str = 'core.models.Snapshot'
 
     id: UUID
-    old_id: UUID
     abid: str
 
     modified: datetime
@@ -251,9 +249,7 @@ class SnapshotSchema(Schema):
 
 
 class SnapshotFilterSchema(FilterSchema):
-    id: Optional[str] = Field(None, q=['id__icontains', 'abid__icontains', 'old_id__icontains', 'timestamp__startswith'])
+    id: Optional[str] = Field(None, q=['id__icontains', 'abid__icontains', 'timestamp__startswith'])
-
-    old_id: Optional[str] = Field(None, q='old_id__icontains')
     abid: Optional[str] = Field(None, q='abid__icontains')
 
     created_by_id: str = Field(None, q='created_by_id')
@@ -266,7 +262,7 @@ class SnapshotFilterSchema(FilterSchema):
     modified__gte: datetime = Field(None, q='modified__gte')
     modified__lt: datetime = Field(None, q='modified__lt')
 
-    search: Optional[str] = Field(None, q=['url__icontains', 'title__icontains', 'tags__name__icontains', 'id__icontains', 'abid__icontains', 'old_id__icontains', 'timestamp__startswith'])
+    search: Optional[str] = Field(None, q=['url__icontains', 'title__icontains', 'tags__name__icontains', 'id__icontains', 'abid__icontains', 'timestamp__startswith'])
     url: Optional[str] = Field(None, q='url')
     tag: Optional[str] = Field(None, q='tags__name')
     title: Optional[str] = Field(None, q='title__icontains')
@@ -293,12 +289,12 @@ def get_snapshot(request, snapshot_id: str, with_archiveresults: bool=True):
     request.with_archiveresults = with_archiveresults
     snapshot = None
     try:
-        snapshot = Snapshot.objects.get(Q(abid__startswith=snapshot_id) | Q(id__startswith=snapshot_id) | Q(old_id__startswith=snapshot_id) | Q(timestamp__startswith=snapshot_id))
+        snapshot = Snapshot.objects.get(Q(abid__startswith=snapshot_id) | Q(id__startswith=snapshot_id) | Q(timestamp__startswith=snapshot_id))
     except Snapshot.DoesNotExist:
         pass
 
     try:
-        snapshot = snapshot or Snapshot.objects.get(Q(abid__icontains=snapshot_id) | Q(id__icontains=snapshot_id) | Q(old_id__icontains=snapshot_id))
+        snapshot = snapshot or Snapshot.objects.get(Q(abid__icontains=snapshot_id) | Q(id__icontains=snapshot_id))
     except Snapshot.DoesNotExist:
         pass
 
@@ -338,7 +334,6 @@ class TagSchema(Schema):
     TYPE: str = 'core.models.Tag'
 
     id: UUID
-    old_id: str
     abid: str
 
     modified: datetime
@@ -351,10 +346,6 @@ class TagSchema(Schema):
     num_snapshots: int
     snapshots: List[SnapshotSchema]
 
-    @staticmethod
-    def resolve_old_id(obj):
-        return str(obj.old_id)
-
     @staticmethod
     def resolve_created_by_id(obj):
         return str(obj.created_by_id)
@@ -386,11 +377,6 @@ def get_tag(request, tag_id: str, with_snapshots: bool=True):
     request.with_snapshots = with_snapshots
     request.with_archiveresults = False
     tag = None
-    try:
-        tag = tag or Tag.objects.get(old_id__icontains=tag_id)
-    except (Tag.DoesNotExist, ValidationError, ValueError):
-        pass
-
     try:
         tag = Tag.objects.get(abid__icontains=tag_id)
     except (Tag.DoesNotExist, ValidationError):

archivebox/core/admin.py

@@ -240,8 +240,8 @@ class ArchiveResultInline(admin.TabularInline):
     # fk_name = 'snapshot'
     extra = 0
     sort_fields = ('end_ts', 'extractor', 'output', 'status', 'cmd_version')
-    readonly_fields = ('result_id', 'completed', 'extractor', 'command', 'version')
+    readonly_fields = ('id', 'result_id', 'completed', 'command', 'version')
-    fields = ('id', 'start_ts', 'end_ts', *readonly_fields, 'cmd', 'cmd_version', 'pwd', 'created_by', 'status', 'output')
+    fields = ('start_ts', 'end_ts', *readonly_fields, 'extractor', 'cmd', 'cmd_version', 'pwd', 'created_by', 'status', 'output')
     # exclude = ('id',)
     ordering = ('end_ts',)
     show_change_link = True
@@ -273,7 +273,7 @@ class ArchiveResultInline(admin.TabularInline):
         snapshot = self.get_parent_object_from_request(request)
 
         # import ipdb; ipdb.set_trace()
-        formset.form.base_fields['id'].widget = formset.form.base_fields['id'].hidden_widget()
+        # formset.form.base_fields['id'].widget = formset.form.base_fields['id'].hidden_widget()
         
         # default values for new entries
         formset.form.base_fields['status'].initial = 'succeeded'
@@ -351,13 +351,13 @@ class SnapshotActionForm(ActionForm):
 
 @admin.register(Snapshot, site=archivebox_admin)
 class SnapshotAdmin(SearchResultsAdminMixin, ABIDModelAdmin):
-    list_display = ('added', 'title_str', 'files', 'size', 'url_str')
+    list_display = ('created', 'title_str', 'files', 'size', 'url_str')
-    sort_fields = ('title_str', 'url_str', 'added')
+    sort_fields = ('title_str', 'url_str', 'created')
-    readonly_fields = ('tags_str', 'timestamp', 'admin_actions', 'status_info', 'bookmarked', 'created', 'added', 'updated', 'modified', 'API', 'link_dir')
+    readonly_fields = ('tags_str', 'timestamp', 'admin_actions', 'status_info', 'bookmarked', 'created', 'created', 'updated', 'modified', 'API', 'link_dir')
-    search_fields = ('id', 'url', 'abid', 'old_id', 'timestamp', 'title', 'tags__name')
+    search_fields = ('id', 'url', 'abid', 'timestamp', 'title', 'tags__name')
-    list_filter = ('added', 'updated', 'archiveresult__status', 'created_by', 'tags__name')
+    list_filter = ('created', 'updated', 'archiveresult__status', 'created_by', 'tags__name')
     fields = ('url', 'created_by', 'title',*readonly_fields)
-    ordering = ['-added']
+    ordering = ['-created']
     actions = ['add_tags', 'remove_tags', 'update_titles', 'update_snapshots', 'resnapshot_snapshot', 'overwrite_snapshots', 'delete_snapshots']
     inlines = [TagInline, ArchiveResultInline]
     list_per_page = min(max(5, CONFIG.SNAPSHOTS_PER_PAGE), 5000)
@@ -391,12 +391,6 @@ class SnapshotAdmin(SearchResultsAdminMixin, ABIDModelAdmin):
         except (Snapshot.DoesNotExist, ValidationError):
             pass
 
-
-        try:
-            snapshot = snapshot or Snapshot.objects.get(old_id=object_id)
-        except (Snapshot.DoesNotExist, Snapshot.MultipleObjectsReturned, ValidationError):
-            pass
-
         if snapshot:
             object_id = str(snapshot.id)
 
@@ -690,7 +684,7 @@ class ArchiveResultAdmin(ABIDModelAdmin):
     list_display = ('start_ts', 'snapshot_info', 'tags_str', 'extractor', 'cmd_str', 'status', 'output_str')
     sort_fields = ('start_ts', 'extractor', 'status')
     readonly_fields = ('cmd_str', 'snapshot_info', 'tags_str', 'created', 'modified', 'API', 'output_summary')
-    search_fields = ('id', 'old_id', 'abid', 'snapshot__url', 'extractor', 'output', 'cmd_version', 'cmd', 'snapshot__timestamp')
+    search_fields = ('id', 'abid', 'snapshot__url', 'extractor', 'output', 'cmd_version', 'cmd', 'snapshot__timestamp')
     fields = ('snapshot', 'extractor', 'status', 'output', 'pwd', 'start_ts', 'end_ts', 'created_by', 'cmd_version', 'cmd', *readonly_fields)
     autocomplete_fields = ['snapshot']
 

archivebox/core/migrations/0070_alter_archiveresult_created_by_alter_snapshot_added_and_more.py

@@ -0,0 +1,52 @@
+# Generated by Django 5.1 on 2024-09-04 09:00
+
+import abid_utils.models
+import django.db.models.deletion
+from django.conf import settings
+from django.db import migrations, models
+
+
+class Migration(migrations.Migration):
+
+    dependencies = [
+        ('core', '0069_alter_archiveresult_created_alter_snapshot_added_and_more'),
+        migrations.swappable_dependency(settings.AUTH_USER_MODEL),
+    ]
+
+    operations = [
+        migrations.AlterField(
+            model_name='archiveresult',
+            name='created_by',
+            field=models.ForeignKey(default=abid_utils.models.get_or_create_system_user_pk, on_delete=django.db.models.deletion.CASCADE, related_name='archiveresult_set', to=settings.AUTH_USER_MODEL),
+        ),
+        migrations.AlterField(
+            model_name='snapshot',
+            name='added',
+            field=abid_utils.models.AutoDateTimeField(db_index=True, default=None),
+        ),
+        migrations.AlterField(
+            model_name='snapshot',
+            name='created',
+            field=abid_utils.models.AutoDateTimeField(db_index=True, default=None),
+        ),
+        migrations.AlterField(
+            model_name='snapshot',
+            name='created_by',
+            field=models.ForeignKey(default=abid_utils.models.get_or_create_system_user_pk, on_delete=django.db.models.deletion.CASCADE, related_name='snapshot_set', to=settings.AUTH_USER_MODEL),
+        ),
+        migrations.AlterField(
+            model_name='snapshot',
+            name='id',
+            field=models.UUIDField(default=None, primary_key=True, serialize=False, unique=True),
+        ),
+        migrations.AlterField(
+            model_name='snapshot',
+            name='old_id',
+            field=models.UUIDField(default=None, editable=False, unique=True),
+        ),
+        migrations.AlterField(
+            model_name='tag',
+            name='created',
+            field=abid_utils.models.AutoDateTimeField(db_index=True, default=None),
+        ),
+    ]

archivebox/core/migrations/0071_remove_archiveresult_old_id_remove_snapshot_old_id_and_more.py

@@ -0,0 +1,65 @@
+# Generated by Django 5.1 on 2024-09-04 23:23
+
+import abid_utils.models
+import django.db.models.deletion
+from django.conf import settings
+from django.db import migrations, models
+
+
+class Migration(migrations.Migration):
+
+    dependencies = [
+        ('core', '0070_alter_archiveresult_created_by_alter_snapshot_added_and_more'),
+        migrations.swappable_dependency(settings.AUTH_USER_MODEL),
+    ]
+
+    operations = [
+        migrations.RemoveField(
+            model_name='archiveresult',
+            name='old_id',
+        ),
+        migrations.RemoveField(
+            model_name='snapshot',
+            name='old_id',
+        ),
+        migrations.RemoveField(
+            model_name='tag',
+            name='old_id',
+        ),
+        
+        migrations.AlterField(
+            model_name='archiveresult',
+            name='created',
+            field=abid_utils.models.AutoDateTimeField(db_index=True, default=None),
+        ),
+        migrations.AlterField(
+            model_name='archiveresult',
+            name='created_by',
+            field=models.ForeignKey(default=None, on_delete=django.db.models.deletion.CASCADE, related_name='archiveresult_set', to=settings.AUTH_USER_MODEL),
+        ),
+        migrations.AlterField(
+            model_name='archiveresult',
+            name='id',
+            field=models.UUIDField(default=None, editable=False, primary_key=True, serialize=False, unique=True, verbose_name='ID'),
+        ),
+        migrations.AlterField(
+            model_name='snapshot',
+            name='created_by',
+            field=models.ForeignKey(default=None, on_delete=django.db.models.deletion.CASCADE, related_name='snapshot_set', to=settings.AUTH_USER_MODEL),
+        ),
+        migrations.AlterField(
+            model_name='snapshot',
+            name='id',
+            field=models.UUIDField(default=None, editable=False, primary_key=True, serialize=False, unique=True, verbose_name='ID'),
+        ),
+        migrations.AlterField(
+            model_name='tag',
+            name='created_by',
+            field=models.ForeignKey(default=None, on_delete=django.db.models.deletion.CASCADE, related_name='tag_set', to=settings.AUTH_USER_MODEL),
+        ),
+        migrations.AlterField(
+            model_name='tag',
+            name='id',
+            field=models.UUIDField(default=None, editable=False, primary_key=True, serialize=False, unique=True, verbose_name='ID'),
+        ),
+    ]

archivebox/core/models.py

@@ -54,13 +54,15 @@ class Tag(ABIDModel):
     abid_ts_src = 'self.created'
     abid_uri_src = 'self.slug'
     abid_subtype_src = '"03"'
-    abid_rand_src = 'self.old_id'
+    abid_rand_src = 'self.id'
 
-    old_id = models.BigIntegerField(unique=True, default=rand_int_id, serialize=False, verbose_name='Old ID')  # legacy PK
+    id = models.UUIDField(primary_key=True, default=None, null=False, editable=False, unique=True, verbose_name='ID')
-
-    id = models.UUIDField(primary_key=True, default=uuid.uuid4, editable=False, unique=True)
     abid = ABIDField(prefix=abid_prefix)
 
+    created_by = models.ForeignKey(settings.AUTH_USER_MODEL, on_delete=models.CASCADE, default=None, null=False, related_name='tag_set')
+    created = AutoDateTimeField(default=None, null=False, db_index=True)
+    modified = models.DateTimeField(auto_now=True)
+
     name = models.CharField(unique=True, blank=False, max_length=100)
     slug = models.SlugField(unique=True, blank=False, max_length=100, editable=False)
     # slug is autoset on save from name, never set it manually
@@ -74,10 +76,6 @@ class Tag(ABIDModel):
     def __str__(self):
         return self.name
 
-    # @property
-    # def old_id(self):
-    #     return self.id
-
     def slugify(self, tag, i=None):
         slug = slugify(tag)
         if i is not None:
@@ -133,16 +131,15 @@ class SnapshotManager(models.Manager):
 
 class Snapshot(ABIDModel):
     abid_prefix = 'snp_'
-    abid_ts_src = 'self.added'
+    abid_ts_src = 'self.created'
     abid_uri_src = 'self.url'
     abid_subtype_src = '"01"'
-    abid_rand_src = 'self.old_id'
+    abid_rand_src = 'self.id'
 
-    old_id = models.UUIDField(default=None, null=False, editable=False, unique=True)  # legacy pk
+    id = models.UUIDField(primary_key=True, default=None, null=False, editable=False, unique=True, verbose_name='ID')
-    id = models.UUIDField(default=None, null=False, primary_key=True, editable=True, unique=True)
     abid = ABIDField(prefix=abid_prefix)
 
-    created_by = models.ForeignKey(settings.AUTH_USER_MODEL, on_delete=models.CASCADE, default=get_or_create_system_user_pk, related_name='snapshot_set')
+    created_by = models.ForeignKey(settings.AUTH_USER_MODEL, on_delete=models.CASCADE, default=None, null=False, related_name='snapshot_set')
     created = AutoDateTimeField(default=None, null=False, db_index=True)
     modified = models.DateTimeField(auto_now=True)
 
@@ -161,7 +158,6 @@ class Snapshot(ABIDModel):
 
     objects = SnapshotManager()
 
-
     def __repr__(self) -> str:
         title = (self.title_stripped or '-')[:64]
         return f'[{self.timestamp}] {self.url[:64]} ({title})'
@@ -414,7 +410,7 @@ class ArchiveResult(ABIDModel):
     abid_ts_src = 'self.snapshot.added'
     abid_uri_src = 'self.snapshot.url'
     abid_subtype_src = 'self.extractor'
-    abid_rand_src = 'self.old_id'
+    abid_rand_src = 'self.id'
 
     EXTRACTOR_CHOICES = (
         ('htmltotext', 'htmltotext'),
@@ -438,13 +434,11 @@ class ArchiveResult(ABIDModel):
         ("skipped", "skipped")
     ]
 
-    old_id = models.BigIntegerField(default=rand_int_id, serialize=False, verbose_name='Old ID')
+    id = models.UUIDField(primary_key=True, default=None, null=False, editable=False, unique=True, verbose_name='ID')
-
-    id = models.UUIDField(primary_key=True, default=uuid.uuid4, editable=True, unique=True, verbose_name='ID')
     abid = ABIDField(prefix=abid_prefix)
 
-    created_by = models.ForeignKey(settings.AUTH_USER_MODEL, on_delete=models.CASCADE, default=get_or_create_system_user_pk, related_name='archiveresult_set')
+    created_by = models.ForeignKey(settings.AUTH_USER_MODEL, on_delete=models.CASCADE, default=None, null=False, related_name='archiveresult_set')
-    created = AutoDateTimeField(default=timezone.now, db_index=True)
+    created = AutoDateTimeField(default=None, null=False, db_index=True)
     modified = models.DateTimeField(auto_now=True)
 
     snapshot = models.ForeignKey(Snapshot, on_delete=models.CASCADE, to_field='id', db_column='snapshot_id')

archivebox/core/views.py

@@ -285,17 +285,17 @@ class SnapshotView(View):
         # ulid = slug.split('_', 1)[-1]
         # try:
         #     try:
-        #         snapshot = snapshot or Snapshot.objects.get(Q(abid=ulid) | Q(id=ulid) | Q(old_id=ulid))
+        #         snapshot = snapshot or Snapshot.objects.get(Q(abid=ulid) | Q(id=ulid))
         #     except Snapshot.DoesNotExist:
         #         pass
 
         #     try:
-        #         snapshot = Snapshot.objects.get(Q(abid__startswith=slug) | Q(abid__startswith=Snapshot.abid_prefix + slug) | Q(id__startswith=slug) | Q(old_id__startswith=slug))
+        #         snapshot = Snapshot.objects.get(Q(abid__startswith=slug) | Q(abid__startswith=Snapshot.abid_prefix + slug) | Q(id__startswith=slug))
         #     except (Snapshot.DoesNotExist, Snapshot.MultipleObjectsReturned):
         #         pass
 
         #     try:
-        #         snapshot = snapshot or Snapshot.objects.get(Q(abid__icontains=snapshot_id) | Q(id__icontains=snapshot_id) | Q(old_id__icontains=snapshot_id))
+        #         snapshot = snapshot or Snapshot.objects.get(Q(abid__icontains=snapshot_id) | Q(id__icontains=snapshot_id))
         #     except Snapshot.DoesNotExist:
         #         pass
         #     return redirect(f'/archive/{snapshot.timestamp}/index.html')
@@ -308,7 +308,7 @@ class SnapshotView(View):
                 # try exact match on full url / ABID first
                 snapshot = Snapshot.objects.get(
                     Q(url='http://' + path) | Q(url='https://' + path) | Q(id__startswith=path)
-                    | Q(abid__icontains=path) | Q(id__icontains=path) | Q(old_id__icontains=path)
+                    | Q(abid__icontains=path) | Q(id__icontains=path)
                 )
             except Snapshot.DoesNotExist:
                 # fall back to match on exact base_url
@@ -352,7 +352,7 @@ class SnapshotView(View):
                 )
                 for snap in Snapshot.objects.filter(
                     Q(url__startswith='http://' + base_url(path)) | Q(url__startswith='https://' + base_url(path))
-                    | Q(abid__icontains=path) | Q(id__icontains=path) | Q(old_id__icontains=path)
+                    | Q(abid__icontains=path) | Q(id__icontains=path)
                 ).only('url', 'timestamp', 'title', 'added').order_by('-added')
             )
             return HttpResponse(

archivebox/index/schema.py

@@ -192,7 +192,6 @@ class Link:
         if extended:
             info.update({
                 'snapshot_id': self.snapshot_id,
-                'snapshot_old_id': self.snapshot_old_id,
                 'snapshot_abid': self.snapshot_abid,
 
                 'link_dir': self.link_dir,
@@ -266,16 +265,12 @@ class Link:
     @cached_property
     def snapshot(self):
         from core.models import Snapshot
-        return Snapshot.objects.only('id', 'old_id', 'abid').get(url=self.url)
+        return Snapshot.objects.only('id', 'abid').get(url=self.url)
 
     @cached_property
     def snapshot_id(self):
         return str(self.snapshot.pk)
 
-    @cached_property
-    def snapshot_old_id(self):
-        return str(self.snapshot.old_id)
-
     @cached_property
     def snapshot_abid(self):
         return str(self.snapshot.ABID)