Features: 1) Add bulk_prefetch_filterable_attributes for optimized attribute fetching; 2) Introduce FilterableAttribute typing for standardizing attribute data across serializers and models; 3) Enhance CategoryDetailSerializer with custom list_serializer_class for bulk operations.

Fixes: 1) Replace redundant `get_filterable_attributes` logic with `filterable_attributes` property to avoid duplications; 2) Remove unnecessary imports and redundant comments in various modules.

Extra: Refactor admin queryset methods for optimization; remove unused `categories` field in `Attribute` model; improve clarity and maintainability of Graphene resolvers and related logic.

core/admin.py

@@ -145,6 +145,13 @@ class AttributeValueInline(TabularInline):  # type: ignore [type-arg]
     verbose_name_plural = _("attribute values")
     icon = "fa-solid fa-list-ul"
 
+    def get_queryset(self, request):
+        return (
+            super()
+            .get_queryset(request)
+            .select_related("attribute", "product")
+        )
+
 
 class ProductImageInline(TabularInline):  # type: ignore [type-arg]
     model = ProductImage
@@ -154,6 +161,9 @@ class ProductImageInline(TabularInline):  # type: ignore [type-arg]
     verbose_name_plural = _("images")
     icon = "fa-regular fa-images"
 
+    def get_queryset(self, request):
+        return super().get_queryset(request).select_related("product")
+
 
 class StockInline(TabularInline):  # type: ignore [type-arg]
     model = Stock
@@ -163,6 +173,13 @@ class StockInline(TabularInline):  # type: ignore [type-arg]
     verbose_name_plural = _("stocks")
     icon = "fa-solid fa-boxes-stacked"
 
+    def get_queryset(self, request):
+        return (
+            super()
+            .get_queryset(request)
+            .select_related("vendor", "product")
+        )
+
 
 class OrderProductInline(TabularInline):  # type: ignore [type-arg]
     model = OrderProduct
@@ -439,6 +456,20 @@ class ProductAdmin(FieldsetsMixin, ActivationActionsMixin, ModelAdmin):  # type:
         "tags",
     ]
 
+    def get_queryset(self, request):
+        # Optimize product change page to avoid N+1 queries
+        return (
+            super()
+            .get_queryset(request)
+            .select_related("category", "brand")
+            .prefetch_related(
+                "tags",
+                "images",
+                "stocks__vendor",
+                "attributes__attribute",
+            )
+        )
+
 
 @register(ProductTag)
 class ProductTagAdmin(FieldsetsMixin, ActivationActionsMixin, ModelAdmin):  # type: ignore [misc, type-arg]

core/graphene/object_types.py

@@ -1,10 +1,10 @@
 import logging
+from contextlib import suppress
 from typing import Any
 
 from constance import config
 from django.core.cache import cache
 from django.db.models import Max, Min, QuerySet
-from django.db.models.functions import Length
 from django.utils.translation import gettext_lazy as _
 from graphene import (
     UUID,
@@ -212,7 +212,6 @@ class CategoryType(DjangoObjectType):  # type: ignore [misc]
         fields = (
             "uuid",
             "markup_percent",
-            "attributes",
             "children",
             "name",
             "slug",
@@ -223,11 +222,16 @@ class CategoryType(DjangoObjectType):  # type: ignore [misc]
         filter_fields = ["uuid"]
         description = _("categories")
 
-    def resolve_children(self, info) -> TreeQuerySet:
+    def resolve_children(self, info) -> TreeQuerySet | list[Category]:
         categories = Category.objects.filter(parent=self)
-        if info.context.user.has_perm("core.view_category"):
+        if not info.context.user.has_perm("core.view_category"):
-            return categories
+            categories = categories.filter(is_active=True)
-        return categories.filter(is_active=True)
+        result = categories
+        with suppress(Exception):
+            items = list(categories)
+            Category.bulk_prefetch_filterable_attributes(items)
+            result = items
+        return result
 
     def resolve_image(self: Category, info) -> str:
         return info.context.build_absolute_uri(self.image.url) if self.image else ""
@@ -238,38 +242,7 @@ class CategoryType(DjangoObjectType):  # type: ignore [misc]
         return 0.0
 
     def resolve_filterable_attributes(self: Category, info):
-        filterable_results = cache.get(f"{self.uuid}_filterable_results", [])
+        return self.filterable_attributes
-
-        if len(filterable_results) > 0:
-            return filterable_results
-
-        for attr in (
-            self.attributes.all()
-            if info.context.user.has_perm("view_attribute")
-            else self.attributes.filter(is_active=True)
-        ):
-            distinct_vals = (
-                AttributeValue.objects.annotate(value_length=Length("value"))
-                .filter(attribute=attr, attribute__is_filterable=True, product__category=self, value_length__lte=30)
-                .values_list("value", flat=True)
-                .distinct()
-            )
-
-            distinct_vals_list = list(distinct_vals)
-
-            if len(distinct_vals_list) <= 128:
-                filterable_results.append(
-                    {
-                        "attribute_name": attr.name,
-                        "possible_values": distinct_vals_list,
-                    }
-                )
-            else:
-                pass
-
-        cache.set(f"{self.uuid}_filterable_results", filterable_results, 86400)
-
-        return filterable_results
 
     def resolve_min_max_prices(self: Category, _info):
         min_max_prices = cache.get(key=f"{self.name}_min_max_prices", default={})

core/models.py

@@ -3,7 +3,7 @@ import json
 import logging
 import traceback
 from contextlib import suppress
-from typing import Any, Optional, Self
+from typing import Any, Optional, Self, Iterable
 
 from constance import config
 from django.contrib.gis.db.models import PointField
@@ -37,6 +37,7 @@ from django.db.models import (
     URLField,
 )
 from django.db.models.indexes import Index
+from django.db.models.functions import Length
 from django.http import Http404
 from django.utils import timezone
 from django.utils.encoding import force_bytes
@@ -52,6 +53,7 @@ from core.abstract import NiceModel
 from core.choices import ORDER_PRODUCT_STATUS_CHOICES, ORDER_STATUS_CHOICES
 from core.errors import DisabledCommerceError, NotEnoughMoneyError
 from core.managers import AddressManager, ProductManager
+from core.typing import FilterableAttribute
 from core.utils import (
     generate_human_readable_id,
     generate_human_readable_token,
@@ -340,6 +342,81 @@ class Category(ExportModelOperationsMixin("category"), NiceModel, MPTTModel):  #
             return 0
         return self.get_descendants().aggregate(max_depth=Max("level"))["max_depth"] - self.get_level()
 
+    @classmethod
+    def bulk_prefetch_filterable_attributes(cls, categories: Iterable["Category"]) -> None:
+        cat_list = [c for c in categories]
+        if not cat_list:
+            return
+        cat_ids = [c.id for c in cat_list if c.id]
+        if not cat_ids:
+            return
+
+        rows = (
+            AttributeValue.objects.annotate(value_length=Length("value"))
+            .filter(
+                product__category_id__in=cat_ids,
+                attribute__is_filterable=True,
+                value_length__lte=30,
+            )
+            .values_list(
+                "product__category_id",
+                "attribute_id",
+                "attribute__name",
+                "attribute__value_type",
+                "value",
+            )
+            .distinct()
+        )
+
+        per_cat: dict[int, dict[int, dict]] = {}
+        for cat_id, attr_id, attr_name, value_type, value in rows:
+            cat_bucket = per_cat.get(cat_id)
+            if cat_bucket is None:
+                cat_bucket = {}
+                per_cat[cat_id] = cat_bucket
+            bucket = cat_bucket.get(attr_id)
+            if bucket is None:
+                bucket = {
+                    "attribute_name": attr_name,
+                    "possible_values": [],
+                    "value_type": value_type,
+                }
+                cat_bucket[attr_id] = bucket
+            if len(bucket["possible_values"]) < 128 and value not in bucket["possible_values"]:
+                bucket["possible_values"].append(value)
+
+        for c in cat_list:
+            data = list(per_cat.get(c.id, {}).values())
+            c.__dict__["filterable_attributes"] = data
+
+    @cached_property
+    def filterable_attributes(self) -> list[FilterableAttribute]:
+        rows = (
+            AttributeValue.objects.annotate(value_length=Length("value"))
+            .filter(
+                product__category=self,
+                attribute__is_filterable=True,
+                value_length__lte=30,
+            )
+            .values_list("attribute_id", "attribute__name", "attribute__value_type", "value")
+            .distinct()
+        )
+
+        by_attr: dict[int, dict] = {}
+        for attr_id, attr_name, value_type, value in rows:
+            bucket = by_attr.get(attr_id)
+            if bucket is None:
+                bucket = {
+                    "attribute_name": attr_name,
+                    "possible_values": [],
+                    "value_type": value_type,
+                }
+                by_attr[attr_id] = bucket
+            if len(bucket["possible_values"]) < 128 and value not in bucket["possible_values"]:
+                bucket["possible_values"].append(value)
+
+        return list(by_attr.values())  # type: ignore [arg-type]
+
     class Meta:
         verbose_name = _("category")
         verbose_name_plural = _("categories")
@@ -627,13 +704,6 @@ class Attribute(ExportModelOperationsMixin("attribute"), NiceModel):  # type: ig
     )
     is_publicly_visible = True
 
-    categories = ManyToManyField(
-        "core.Category",
-        related_name="attributes",
-        help_text=_("category of this attribute"),
-        verbose_name=_("categories"),
-    )
-
     group = ForeignKey(
         "core.AttributeGroup",
         on_delete=CASCADE,

core/serializers/detail.py

@@ -1,12 +1,9 @@
 import logging
-from collections import defaultdict
 from contextlib import suppress
 from typing import Collection, Any
 
-from django.core.cache import cache
-from django.db.models.functions import Length
 from rest_framework.fields import JSONField, SerializerMethodField
-from rest_framework.serializers import ModelSerializer
+from rest_framework.serializers import ModelSerializer, ListSerializer
 from rest_framework_recursive.fields import RecursiveField
 
 from core.models import (
@@ -28,6 +25,7 @@ from core.models import (
     Wishlist,
 )
 from core.serializers.simple import CategorySimpleSerializer, ProductSimpleSerializer
+from core.typing import FilterableAttribute
 from core.serializers.utility import AddressSerializer
 
 logger = logging.getLogger("django")
@@ -47,6 +45,15 @@ class AttributeGroupDetailSerializer(ModelSerializer):
         ]
 
 
+class CategoryDetailListSerializer(ListSerializer):
+
+    def to_representation(self, data):  # type: ignore[override]
+        items = list(data)
+        with suppress(Exception):
+            Category.bulk_prefetch_filterable_attributes(items)
+        return super().to_representation(items)
+
+
 class CategoryDetailSerializer(ModelSerializer):
     children = SerializerMethodField()
     image = SerializerMethodField()
@@ -54,6 +61,7 @@ class CategoryDetailSerializer(ModelSerializer):
 
     class Meta:
         model = Category
+        list_serializer_class = CategoryDetailListSerializer
         fields = [
             "uuid",
             "name",
@@ -72,46 +80,8 @@ class CategoryDetailSerializer(ModelSerializer):
             return obj.image.url
         return None
 
-    def get_filterable_attributes(self, obj: Category) -> list[dict]:
+    def get_filterable_attributes(self, obj: Category) -> list[FilterableAttribute]:
-        cache_key = f"{obj.uuid}_filterable_results"
+        return obj.filterable_attributes
-        filterable_results = cache.get(cache_key)
-        if filterable_results is not None:
-            return filterable_results
-
-        attrs_qs = obj.attributes.filter(is_active=True, is_filterable=True)
-        attributes = list(attrs_qs)
-
-        attr_ids = [a.id for a in attributes]
-        raw_vals = (
-            AttributeValue.objects.annotate(value_length=Length("value"))
-            .filter(
-                attribute_id__in=attr_ids,
-                product__category=obj,
-                value_length__lte=30,
-            )
-            .values_list("attribute_id", "value")
-            .distinct()
-        )
-
-        grouped = defaultdict(list)
-        for attr_id, val in raw_vals:
-            grouped[attr_id].append(val)
-
-        filterable_results = []
-        for attr in attributes:
-            vals = grouped.get(attr.id, [])  # type: ignore
-            slice_vals = vals[:128] if len(vals) > 128 else vals
-            filterable_results.append(
-                {
-                    "attribute_name": attr.name,
-                    "possible_values": slice_vals,
-                    "value_type": attr.value_type,
-                }
-            )
-
-        cache.set(cache_key, filterable_results, 3600)
-
-        return filterable_results
 
     def get_children(self, obj) -> Collection[Any]:
         request = self.context.get("request")

core/typing/__init__.py

@@ -0,0 +1,6 @@
+from .models import AttributeValueTypeLiteral, FilterableAttribute
+
+__all__ = [
+    "AttributeValueTypeLiteral",
+    "FilterableAttribute",
+]

core/typing/models.py

@@ -0,0 +1,16 @@
+from typing import Literal, TypedDict
+
+AttributeValueTypeLiteral = Literal[
+    "string",
+    "integer",
+    "float",
+    "boolean",
+    "array",
+    "object",
+]
+
+
+class FilterableAttribute(TypedDict):
+    attribute_name: str
+    possible_values: list[str]
+    value_type: AttributeValueTypeLiteral

vibes_auth/messaging/services.py

@@ -54,7 +54,6 @@ def create_anon_thread(email: str) -> ChatThread:
 def send_message(thread: ChatThread, *, sender_user: Optional[User], sender_type: SenderType, text: str) -> ChatMessage:
     if not text or len(text) > 1028:
         raise ValidationError({"text": _("Message must be 1..1028 characters.")})
-    # Permission rules: non-staff may only write to their own thread or anon thread they initiated
     if sender_user and not sender_user.is_staff:
         if thread.user_id != sender_user.pk:
             raise PermissionDenied
@@ -88,7 +87,6 @@ def send_message(thread: ChatThread, *, sender_user: Optional[User], sender_type
 
 
 def auto_reply(thread: ChatThread) -> None:
-    # Localizable text, do not translate here
     text = _("We're searching for the operator to answer you already, hold by!")
     msg = ChatMessage.objects.create(
         thread=thread,
@@ -119,7 +117,6 @@ def claim_thread(thread: ChatThread, staff_user: User) -> ChatThread:
     if not staff_user.is_staff:
         raise PermissionDenied
     if thread.assigned_to_id and not staff_user.is_superuser:
-        # already assigned, cannot reassign/unassign
         raise PermissionDenied
     thread.assigned_to = staff_user
     thread.save(update_fields=["assigned_to", "modified"])
@@ -144,9 +141,8 @@ def reassign_thread(thread: ChatThread, superuser: User, new_staff: User) -> Cha
 
 def close_thread(thread: ChatThread, actor: User | None) -> ChatThread:
     if actor and actor.is_staff:
-        pass  # allowed
+        pass
     elif actor and not actor.is_staff:
-        # non-staff allowed to close own thread? Keep simple: allowed only for staff for now
         raise PermissionDenied
     thread.status = ThreadStatus.CLOSED
     thread.save(update_fields=["status", "modified"])