Org Web Adapter

Editing hungryroot/brand/cookbook-dietary-tag-backfill.org

:PROPERTIES:
:ID:       6239515d-8f03-48df-9835-8e7e04839335
:END:
#+title: Cookbook Dietary Tag Backfill

One-off script to populate =Cookbook.dietary_tags= (M2M to =app.Tag=) by unioning,
across every pairing and product referenced from the cookbook's sections:
- the =HAS_X= allergen tags (IDs 66-75), computed from the =has_*= boolean fields, and
- the protein tags (tag groups =MEAT_PROTEIN=, =SEAFOOD_PROTEIN=, =PLANT_PROTEIN=),
  sourced directly from each item's existing =tags= M2M.

Tag IDs and exclusion semantics follow the Confluence doc
[[https://hungryroot.atlassian.net/wiki/spaces/WD/pages/1597276176/Discover+Cookbooks+Filter+to+API+Mapping][Discover Cookbooks - Filter to API Mapping]]:
- =/api/v3/cookbooks/= filters protein via =tag_id= against this set
  and filters dietary restrictions via =exclude_tag_id= against the =HAS_*= subset
  (e.g. Vegan -> =exclude_tag_id=66,67,74,69,72,75=).
- VEGAN/GLUTEN_FREE badge tags are still NOT written here.

* Usage

1. SSH to server, =python manage.py shell=.
2. Paste the script below.
3. With =DRY_RUN=True= (default) it prints =existing= / =computed= / =adding= per cookbook
   without writing.
4. Optionally set =COOKBOOK_ID= to a single cookbook id to scope the run.
5. When the diff looks right, flip =DRY_RUN=False= and paste again.

Existing tags are preserved; the script only adds missing =HAS_*= tag IDs (merge, not replace).

* Script

#SRC_BEGIN python
from django.db import transaction
from cookbooks.models import Cookbook, CookbookDietaryTag, CookbookItem, CookbookItemType
from app.models.tag import Tag
from app.models.tag_group import TagGroup

# Set to a single Cookbook ID to test on one; leave None to process all.
COOKBOOK_ID = None
DRY_RUN = True  # Flip to False to actually write.

# has_* field on DietaryMixin -> Tag.ID for the HAS_X allergen tag.
# Mirrors exclude_tag_id sets in the Confluence "Discover Cookbooks - Filter to API Mapping" doc.
HAS_FIELD_TO_TAG_ID = {
    "has_meat":      Tag.ID.HAS_MEAT,       # 66
    "has_fish":      Tag.ID.HAS_FISH,       # 67
    "has_gluten":    Tag.ID.HAS_GLUTEN,     # 68
    "has_dairy":     Tag.ID.HAS_DAIRY,      # 69
    "has_soy":       Tag.ID.HAS_SOY,        # 70
    "has_nut":       Tag.ID.HAS_NUT,        # 71
    "has_egg":       Tag.ID.HAS_EGG,        # 72
    "has_peanut":    Tag.ID.HAS_PEANUT,     # 73
    "has_byproduct": Tag.ID.HAS_BYPRODUCT,  # 74
    "has_shellfish": Tag.ID.HAS_SHELLFISH,  # 75
}

PROTEIN_TAG_GROUP_IDS = {
    TagGroup.ID.MEAT_PROTEIN,     # 1
    TagGroup.ID.SEAFOOD_PROTEIN,  # 2
    TagGroup.ID.PLANT_PROTEIN,    # 3
}

def allergen_tag_ids(obj) -> set[int]:
    """Return HAS_X tag IDs for any allergen the obj contains."""
    return {tag_id for field, tag_id in HAS_FIELD_TO_TAG_ID.items() if getattr(obj, field)}

def protein_tag_ids(obj) -> set[int]:
    """Return protein Tag IDs already assigned to obj via its tags M2M."""
    return set(
        obj.tags.filter(tag_group_id__in=PROTEIN_TAG_GROUP_IDS).values_list("id", flat=True)
    )

def compute_cookbook_tag_ids(cookbook: Cookbook) -> set[int]:
    items = (
        CookbookItem.objects
        .filter(section__cookbook=cookbook)
        .select_related("pairing", "product")
        .prefetch_related("pairing__tags", "product__tags")
    )
    tag_ids: set[int] = set()
    for item in items:
        obj = item.pairing if item.item_type == CookbookItemType.PAIRING else item.product
        if obj is None:
            continue
        tag_ids |= allergen_tag_ids(obj)
        tag_ids |= protein_tag_ids(obj)
    return tag_ids

def update_tags(cookbook_id=None, dry_run=True):
    qs = Cookbook.objects.all()
    if cookbook_id is not None:
        qs = qs.filter(id=cookbook_id)

for cookbook in qs:
        computed = compute_cookbook_tag_ids(cookbook)
        existing = set(cookbook.dietary_tags.values_list("id", flat=True))
        to_add = computed - existing
        print(
            f"Cookbook {cookbook.id} {cookbook.slug!r}: existing={sorted(existing)} "
            f"computed={sorted(computed)} adding={sorted(to_add)}"
        )

if dry_run:
            print("Dry run, no changes made")
            continue
        if not to_add:
            print("No new tags to add, no changes made")
            continue

with transaction.atomic():
            CookbookDietaryTag.objects.bulk_create(
                [CookbookDietaryTag(cookbook=cookbook, tag_id=tid) for tid in to_add],
                ignore_conflicts=True,
            )

print("done")
#SRC_END