implemented recategorize

README.md

@@ -113,6 +113,20 @@ Each is also exposed as a Celery task (`api.<name>`) and triggerable from the ad
 - `python manage.py seed_banners`: Updates procedurally generated banner configurations from `backend/banners/*.json`.
 - `python manage.py backup_db` (`--full` for media too): Creates a consistent SQLite snapshot in `backend/backups/`.
 
+#### Categories & `recategorize_flavors`
+
+`sync_flavors` derives a flavor's category from the Shopify `product_type` (e.g. `42 - Syrup Bottle` → **Syrup**), stripping the numeric prefix and the form-factor word. Any new drink line Holy ships **auto-creates** its category on the next sync — no code change needed. Packs, shakers, merch, stickers and other non-drink types funnel to **Packs and other**.
+
+One catch: `sync_flavors` only assigns a category when it **creates** a row. Existing flavors keep their original category, so a categorization change won't retro-apply to rows already in the DB. Use the one-off command below to fix mis-shelved rows. It is run manually (not exposed as a Celery task / Jobs-tab entry) and is **dry-run by default** — nothing is written without `--apply`.
+
+```bash
+python manage.py recategorize_flavors                      # dry-run: report what would move
+python manage.py recategorize_flavors --apply              # persist the moves
+python manage.py recategorize_flavors --apply --merge-collisions  # also fold name-clashes
+```
+
+It re-fetches the catalog, matches rows by `external_id`, and re-runs the exact `sync_flavors` categorization logic. Rows not in the catalog (discontinued) are never touched. When a move would collide with an existing flavor of the same name in the target category (`UniqueConstraint(name, category)`), it is skipped and reported unless `--merge-collisions` is passed, which folds the rows via `merge_flavors` (preserving ratings and replies).
+
 ---
 
 ## 💾 Backups & Restoration

backend/api/management/commands/recategorize_flavors.py

@@ -0,0 +1,143 @@
+import requests
+from django.core.management.base import BaseCommand
+from django.db import transaction
+
+from api.management.commands.sync_flavors import Command as SyncCommand
+from api.models import Category, Flavor
+from api.services.flavor import merge_flavors
+
+_CATALOG_URL = "https://weareholy.com/products.json?limit=250"
+_PACKS_SLUG = "packs-and-other"
+_PACKS_NAME = "Packs and other"
+
+
+class Command(BaseCommand):
+    help = (
+        "Re-evaluate the category of every already-synced flavor against the current "
+        "sync_flavors logic and move mis-shelved rows. Dry-run unless --apply is given. "
+        "Existing flavors keep their creation-time category during normal sync, so this "
+        "is the way to retro-apply categorization changes to rows that already exist."
+    )
+
+    def add_arguments(self, parser):
+        parser.add_argument(
+            "--apply",
+            action="store_true",
+            help="Persist changes. Without this flag the command only reports what would change.",
+        )
+        parser.add_argument(
+            "--merge-collisions",
+            action="store_true",
+            help=(
+                "If the target category already holds a flavor with the same name, merge "
+                "this row into it (preserves ratings) instead of skipping. Implies a write; "
+                "only acts under --apply."
+            ),
+        )
+
+    def fetch_products(self) -> list[dict]:
+        response = requests.get(_CATALOG_URL, headers={"User-Agent": "Mozilla/5.0"}, timeout=15)
+        response.raise_for_status()
+        return response.json().get("products", [])
+
+    def handle(self, *args, **options):
+        apply = options["apply"]
+        merge_collisions = options["merge_collisions"]
+
+        try:
+            products = self.fetch_products()
+        except Exception as e:
+            self.stdout.write(self.style.ERROR(f"Failed to fetch catalog: {e}"))
+            return
+
+        by_id = {p["id"]: p for p in products if p.get("id") is not None}
+        self.stdout.write(f"Fetched {len(by_id)} catalog products.")
+
+        sync = SyncCommand()
+        moved = merged = collisions = unresolved = not_in_catalog = unchanged = 0
+
+        # Everything mutating runs in one transaction. resolve_category auto-creates
+        # categories as a side effect, so dry-run rolls the whole thing back to stay pure.
+        with transaction.atomic():
+            packs_category = Category.objects.get_or_create(
+                slug=_PACKS_SLUG, defaults={"name": _PACKS_NAME}
+            )[0]
+            moved, merged, collisions, unresolved, not_in_catalog, unchanged = self._process(
+                by_id, sync, packs_category, apply, merge_collisions
+            )
+            if not apply:
+                transaction.set_rollback(True)
+
+        verb = "Applied" if apply else "Would change (dry-run)"
+        self.stdout.write(
+            self.style.SUCCESS(
+                f"{verb}: moved={moved}, merged={merged}, collisions={collisions}, "
+                f"unresolved={unresolved}, unchanged={unchanged}, not_in_catalog={not_in_catalog}"
+            )
+        )
+        if not apply and (moved or collisions):
+            self.stdout.write("Re-run with --apply to persist.")
+
+    def _process(self, by_id, sync, packs_category, apply, merge_collisions):
+        moved = merged = collisions = unresolved = not_in_catalog = unchanged = 0
+
+        for flavor in Flavor.objects.filter(external_id__isnull=False).select_related("category"):
+            product = by_id.get(flavor.external_id)
+            if product is None:
+                not_in_catalog += 1
+                continue
+
+            title = product.get("title", "").strip()
+            tags = product.get("tags", [])
+            if isinstance(tags, str):
+                tags = [t.strip() for t in tags.split(",")]
+            tags_lower = [t.lower() for t in tags]
+
+            target = sync.resolve_category(product, title.lower(), tags_lower, packs_category)
+
+            if target is None:
+                unresolved += 1
+                self.stdout.write(
+                    self.style.WARNING(
+                        f"  ? {flavor.name!r} [{flavor.category.slug}]: resolver yields no "
+                        f"category (product_type={product.get('product_type')!r}) — left as-is"
+                    )
+                )
+                continue
+
+            if target.id == flavor.category_id:
+                unchanged += 1
+                continue
+
+            conflict = (
+                Flavor.objects.filter(name=flavor.name, category=target)
+                .exclude(pk=flavor.pk)
+                .first()
+            )
+            if conflict is not None:
+                collisions += 1
+                if merge_collisions:
+                    self.stdout.write(
+                        f"  MERGE {flavor.name!r}: {flavor.category.slug} -> {target.slug} "
+                        f"(into existing row #{conflict.pk})"
+                    )
+                    if apply:
+                        merge_flavors(conflict, flavor)
+                        merged += 1
+                else:
+                    self.stdout.write(
+                        self.style.WARNING(
+                            f"  SKIP {flavor.name!r}: target {target.slug} already has a flavor "
+                            f"named {flavor.name!r} (#{conflict.pk}). Re-run with "
+                            f"--merge-collisions to fold them."
+                        )
+                    )
+                continue
+
+            self.stdout.write(f"  MOVE {flavor.name!r}: {flavor.category.slug} -> {target.slug}")
+            if apply:
+                flavor.category = target
+                flavor.save(update_fields=["category"])
+            moved += 1
+
+        return moved, merged, collisions, unresolved, not_in_catalog, unchanged

backend/api/management/commands/sync_flavors.py

@@ -7,11 +7,93 @@ from urllib.parse import urlparse
 import requests
 from django.conf import settings
 from django.core.management.base import BaseCommand
+from django.utils.text import slugify
 
 from api.models import Category, Flavor
 
 _DOWNLOAD_WORKERS = 8
 
+# Shopify product_type looks like "42 - Syrup Bottle" / "01 - Energy Bundle".
+# Strip the numeric prefix, then the trailing form-factor word, to get the
+# drink family ("Syrup", "Energy", "Iced Tea") that becomes a Category.
+_PRODUCT_TYPE_PREFIX = re.compile(r"^\s*\d+\s*[-–—]\s*")
+
+# Form-factor words are Holy's internal SKU taxonomy, NOT a pack signal:
+# "Energy Bundle" is a single can, "Caipirinha Crab" is its title. Pack
+# detection stays title-driven (see _is_pack). These only get stripped to
+# reveal the family.
+_FORM_WORDS = frozenset(
+    {"bundle", "sachet", "sachetbox", "bottle", "box", "pack", "set", "sachets"}
+)
+
+# When the product_type family itself is one of these, it's not a drink line —
+# funnel to "Packs and other" instead of minting a category for it.
+_NON_FLAVOR_TOKENS = frozenset(
+    {
+        "value",
+        "shaker",
+        "merch",
+        "sticker",
+        "packaging",
+        "material",
+        "other",
+        "items",
+        "mixed",
+        "collector",
+        "collector's",
+        "free",
+        "products",
+        "naturalrabatt",
+        "sample",
+        "samplebox",
+    }
+)
+
+_PACKS_SLUG = "packs-and-other"
+_PACKS_NAME = "Packs and other"
+
+_PACK_TITLE_KEYWORDS = (
+    "bundle",
+    "set",
+    "box",
+    "probe",
+    "sample",
+    "taster",
+    "starter",
+    "collection",
+    "probier",
+)
+
+# Fallback only — product_type covers these, but some rows lack one.
+_TAG_FAMILY = (
+    (("holy energy", "energy"), "Energy"),
+    (("holy hydration", "hydration"), "Hydration"),
+    (("holy iced tea", "iced tea"), "Iced Tea"),
+    (("milkshake",), "Milkshake"),
+)
+
+
+def _family_from_product_type(product_type: str) -> str | None:
+    """Return the drink family from a Shopify product_type, or None.
+
+    "42 - Syrup Bottle" -> "Syrup"; "Value Pack" / "Shaker" -> None (non-flavor).
+    """
+    if not product_type:
+        return None
+    stripped = _PRODUCT_TYPE_PREFIX.sub("", product_type).strip()
+    if not stripped:
+        return None
+
+    tokens = stripped.split()
+    if any(tok.lower() in _NON_FLAVOR_TOKENS for tok in tokens):
+        return None
+
+    while tokens and tokens[-1].lower() in _FORM_WORDS:
+        tokens.pop()
+    if not tokens:
+        return None
+    return " ".join(tokens)
+
 
 class Command(BaseCommand):
     help = "Syncs flavors from Holy Energy Shopify API"
@@ -80,6 +162,40 @@ class Command(BaseCommand):
                     results[i] = saved
         return [results[i] for i, _, _ in targets if i in results]
 
+    @staticmethod
+    def _category_for_family(family_name: str) -> Category:
+        """get_or_create a Category from a family name. slugify('Iced Tea') == 'iced-tea',
+        which matches the pre-existing slugs, so known families are reused not duplicated."""
+        slug = slugify(family_name)
+        return Category.objects.get_or_create(slug=slug, defaults={"name": family_name})[0]
+
+    @staticmethod
+    def _is_pack(title_lower: str, tags_lower: list[str]) -> bool:
+        if any(k in title_lower for k in _PACK_TITLE_KEYWORDS):
+            return True
+        return "shaker" in title_lower or "merch" in tags_lower
+
+    def resolve_category(self, product, title_lower, tags_lower, packs_category) -> Category | None:
+        """Pick a Category for a Shopify product. None means skip (no usable signal).
+
+        Order: pack detection (title/tags) -> product_type family (dynamic, auto-creates)
+        -> tag-based family fallback -> shaker tag -> skip.
+        """
+        if self._is_pack(title_lower, tags_lower):
+            return packs_category
+
+        family = _family_from_product_type(product.get("product_type", ""))
+        if family:
+            return self._category_for_family(family)
+
+        for keywords, name in _TAG_FAMILY:
+            if any(k in tags_lower for k in keywords):
+                return self._category_for_family(name)
+
+        if "shaker" in tags_lower:
+            return packs_category
+        return None
+
     def handle(self, *args, **options):
         url = "https://weareholy.com/products.json?limit=250"
         self.stdout.write(f"Fetching from {url}...")
@@ -95,21 +211,9 @@ class Command(BaseCommand):
         products = data.get("products", [])
         self.stdout.write(f"Found {len(products)} products.")
 
-        cat_map = {
+        packs_category = Category.objects.get_or_create(
-            "energy": Category.objects.get_or_create(name="Energy", defaults={"slug": "energy"})[0],
+            slug=_PACKS_SLUG, defaults={"name": _PACKS_NAME}
-            "hydration": Category.objects.get_or_create(
+        )[0]
-                name="Hydration", defaults={"slug": "hydration"}
-            )[0],
-            "iced-tea": Category.objects.get_or_create(
-                name="Iced Tea", defaults={"slug": "iced-tea"}
-            )[0],
-            "milkshake": Category.objects.get_or_create(
-                name="Milkshake", defaults={"slug": "milkshake"}
-            )[0],
-            "packs": Category.objects.get_or_create(
-                name="Packs and other", defaults={"slug": "packs-and-other"}
-            )[0],
-        }
 
         created_count = 0
         updated_count = 0
@@ -124,39 +228,8 @@ class Command(BaseCommand):
             tags_lower = [t.lower() for t in tags]
             title_lower = title.lower()
 
-            is_pack = False
+            category = self.resolve_category(p, title_lower, tags_lower, packs_category)
-            pack_keywords = [
-                "bundle",
-                "set",
-                "box",
-                "probe",
-                "sample",
-                "taster",
-                "starter",
-                "collection",
-                "probier",
-            ]
-            if any(k in title_lower for k in pack_keywords):
-                is_pack = True
-            if "shaker" in title_lower or "merch" in tags_lower:
-                is_pack = True
-
-            category = None
-            if is_pack:
-                category = cat_map["packs"]
-            elif "holy energy" in tags_lower or "energy" in tags_lower:
-                category = cat_map["energy"]
-            elif "holy hydration" in tags_lower or "hydration" in tags_lower:
-                category = cat_map["hydration"]
-            elif "holy iced tea" in tags_lower or "iced tea" in tags_lower:
-                category = cat_map["iced-tea"]
-            elif "milkshake" in tags_lower:
-                category = cat_map["milkshake"]
-
             if not category:
-                if "shaker" in tags_lower:
-                    category = cat_map["packs"]
-                else:
                 continue
 
             variants = p.get("variants", [])

backend/api/tests/test_recategorize.py

@@ -0,0 +1,69 @@
+from unittest.mock import patch
+
+from django.core.management import call_command
+from django.test import TestCase
+
+from api.models import Category, Flavor, Rating, User
+
+
+def _product(pid, title, product_type="", tags=None):
+    return {"id": pid, "title": title, "product_type": product_type, "tags": tags or []}
+
+
+class RecategorizeTests(TestCase):
+    def setUp(self) -> None:
+        self.packs = Category.objects.create(name="Packs and other", slug="packs-and-other")
+        self.energy = Category.objects.create(name="Energy", slug="energy")
+
+    def _run(self, products, **opts):
+        with patch(
+            "api.management.commands.recategorize_flavors.Command.fetch_products",
+            return_value=products,
+        ):
+            call_command("recategorize_flavors", **opts)
+
+    def test_dry_run_does_not_move(self) -> None:
+        # Mis-shelved: a real syrup product currently parked in packs.
+        flavor = Flavor.objects.create(name="Reinigungsbürste", category=self.packs, external_id=42)
+        self._run([_product(42, "Reinigungsbürste", "42 - Syrup Bottle")])
+        flavor.refresh_from_db()
+        self.assertEqual(flavor.category_id, self.packs.id)
+        self.assertFalse(Category.objects.filter(slug="syrup").exists())
+
+    def test_apply_moves_and_creates_category(self) -> None:
+        flavor = Flavor.objects.create(name="Reinigungsbürste", category=self.packs, external_id=42)
+        self._run([_product(42, "Reinigungsbürste", "42 - Syrup Bottle")], apply=True)
+        flavor.refresh_from_db()
+        self.assertEqual(flavor.category.slug, "syrup")
+
+    def test_unchanged_row_stays(self) -> None:
+        flavor = Flavor.objects.create(name="Mojito Macaw", category=self.energy, external_id=1)
+        self._run([_product(1, "Mojito Macaw", "01 - Energy Bundle")], apply=True)
+        flavor.refresh_from_db()
+        self.assertEqual(flavor.category_id, self.energy.id)
+
+    def test_row_not_in_catalog_is_skipped(self) -> None:
+        flavor = Flavor.objects.create(name="Gone", category=self.packs, external_id=999)
+        self._run([_product(1, "Mojito Macaw", "01 - Energy Bundle")], apply=True)
+        flavor.refresh_from_db()
+        self.assertEqual(flavor.category_id, self.packs.id)
+
+    def test_collision_skipped_without_flag(self) -> None:
+        syrup = Category.objects.create(name="Syrup", slug="syrup")
+        # Target already holds a row of the same name.
+        Flavor.objects.create(name="Dup", category=syrup, external_id=100)
+        moving = Flavor.objects.create(name="Dup", category=self.packs, external_id=42)
+        self._run([_product(42, "Dup", "42 - Syrup Bottle")], apply=True)
+        moving.refresh_from_db()
+        self.assertEqual(moving.category_id, self.packs.id)  # untouched
+        self.assertEqual(Flavor.objects.filter(name="Dup").count(), 2)
+
+    def test_collision_merges_with_flag(self) -> None:
+        user = User.objects.create_user(username="u", password="p")
+        syrup = Category.objects.create(name="Syrup", slug="syrup")
+        keep = Flavor.objects.create(name="Dup", category=syrup, external_id=100)
+        moving = Flavor.objects.create(name="Dup", category=self.packs, external_id=42)
+        Rating.objects.create(user=user, flavor=moving, score=5)
+        self._run([_product(42, "Dup", "42 - Syrup Bottle")], apply=True, merge_collisions=True)
+        self.assertFalse(Flavor.objects.filter(pk=moving.pk).exists())
+        self.assertTrue(Rating.objects.filter(flavor=keep, user=user).exists())

backend/api/tests/test_sync_categories.py

@@ -0,0 +1,54 @@
+from django.test import TestCase
+
+from api.management.commands.sync_flavors import Command, _family_from_product_type
+from api.models import Category
+
+
+class FamilyFromProductTypeTests(TestCase):
+    def test_strips_prefix_and_form_word(self) -> None:
+        self.assertEqual(_family_from_product_type("42 - Syrup Bottle"), "Syrup")
+        self.assertEqual(_family_from_product_type("01 - Energy Bundle"), "Energy")
+        self.assertEqual(_family_from_product_type("04 - Energy Sachet"), "Energy")
+
+    def test_multiword_family(self) -> None:
+        self.assertEqual(_family_from_product_type("02 - Iced Tea Bundle"), "Iced Tea")
+
+    def test_non_flavor_families_return_none(self) -> None:
+        for pt in ("10 - Shaker", "11 - Merch", "12 - Value Pack", "16 - Packaging Material"):
+            self.assertIsNone(_family_from_product_type(pt), pt)
+
+    def test_empty_or_missing(self) -> None:
+        self.assertIsNone(_family_from_product_type(""))
+        self.assertIsNone(_family_from_product_type("99 - "))
+
+
+class ResolveCategoryTests(TestCase):
+    def setUp(self) -> None:
+        self.cmd = Command()
+        self.packs = Category.objects.create(name="Packs and other", slug="packs-and-other")
+
+    def _resolve(self, product, title="", tags=None):
+        tags = tags or []
+        return self.cmd.resolve_category(product, title.lower(), tags, self.packs)
+
+    def test_new_family_auto_creates_category(self) -> None:
+        cat = self._resolve({"product_type": "42 - Syrup Bottle"}, title="Reinigungsbürste")
+        self.assertIsNotNone(cat)
+        self.assertEqual(cat.slug, "syrup")
+        self.assertTrue(Category.objects.filter(slug="syrup").exists())
+
+    def test_known_family_reuses_existing_slug(self) -> None:
+        existing = Category.objects.create(name="Iced Tea", slug="iced-tea")
+        cat = self._resolve({"product_type": "02 - Iced Tea Bundle"}, title="Erdbeere")
+        self.assertEqual(cat.id, existing.id)
+
+    def test_pack_title_keyword_wins_over_family(self) -> None:
+        cat = self._resolve({"product_type": "43 - Syrup Bundle"}, title="3er Box Syrup")
+        self.assertEqual(cat.id, self.packs.id)
+
+    def test_tag_fallback_when_no_product_type(self) -> None:
+        cat = self._resolve({}, title="Mojito Macaw", tags=["holy energy"])
+        self.assertEqual(cat.slug, "energy")
+
+    def test_no_signal_returns_none(self) -> None:
+        self.assertIsNone(self._resolve({"product_type": "20 - Other Items"}, title="Mystery"))