Corp/backend/commerce/services/metrics.py

from collections import defaultdict
from datetime import datetime
from decimal import Decimal

from django.db.models import (
    Case,
    CharField,
    Count,
    DecimalField,
    ExpressionWrapper,
    F,
    OuterRef,
    Q,
    Subquery,
    Sum,
    Value,
    When,
)
from django.db.models.functions import Coalesce

from commerce.models import Cost, Product, SaleEventType, SaleLine, StockSnapshot

MONEY_FIELD = DecimalField(max_digits=20, decimal_places=2)
ZERO_MONEY = Decimal("0.00")


def _date_filter(prefix: str, date_from: datetime | None, date_to: datetime | None):
    query = Q()
    if date_from is not None:
        query &= Q(**{f"{prefix}__gte": date_from})
    if date_to is not None:
        query &= Q(**{f"{prefix}__lte": date_to})
    return query


def _sale_metric_rows(
    product: Product,
    date_from: datetime | None,
    date_to: datetime | None,
    *group_fields: str,
):
    signed_quantity = ExpressionWrapper(
        F("quantity") * models_case_return_multiplier(),
        output_field=DecimalField(max_digits=20, decimal_places=4),
    )
    signed_gross = ExpressionWrapper(
        F("gross_amount") * models_case_return_multiplier(), output_field=MONEY_FIELD
    )
    signed_discount = ExpressionWrapper(
        F("discount_amount") * models_case_return_multiplier(),
        output_field=MONEY_FIELD,
    )
    signed_cost = ExpressionWrapper(
        F("cost_amount") * models_case_return_multiplier(), output_field=MONEY_FIELD
    )

    rows = (
        SaleLine.objects.filter(product=product, sale__counts_as_sale=True)
        .filter(_date_filter("sale__occurred_at", date_from, date_to))
        .values(*group_fields)
        .annotate(
            quantity=Coalesce(Sum(signed_quantity), Decimal("0")),
            gross=Coalesce(Sum(signed_gross), ZERO_MONEY),
            discounts=Coalesce(Sum(signed_discount), ZERO_MONEY),
            cost_of_goods=Coalesce(Sum(signed_cost), ZERO_MONEY),
            missing_cost_lines=Count(
                "id",
                filter=Q(
                    metadata__cost_source="unavailable",
                ),
            ),
        )
        .order_by(*group_fields)
    )
    return list(rows)


def _sale_metrics(
    product: Product,
    date_from: datetime | None,
    date_to: datetime | None,
):
    rows = _sale_metric_rows(product, date_from, date_to, "sale__currency")
    return {row["sale__currency"]: row for row in rows}


def models_case_return_multiplier():
    from django.db.models import Case, IntegerField, Value, When

    return Case(
        When(sale__event_type=SaleEventType.RETURN, then=Value(-1)),
        default=Value(1),
        output_field=IntegerField(),
    )


def _cost_metrics(
    product: Product,
    date_from: datetime | None,
    date_to: datetime | None,
):
    rows = (
        Cost.objects.filter(Q(product=product) | Q(sale_line__product=product))
        .filter(_date_filter("occurred_at", date_from, date_to))
        .values("currency", "cost_type")
        .annotate(amount=Coalesce(Sum("amount"), ZERO_MONEY))
        .order_by("currency", "cost_type")
    )
    result: dict[str, dict[str, Decimal]] = defaultdict(dict)
    for row in rows:
        result[row["currency"]][row["cost_type"]] = row["amount"]
    return result


def _channel_metrics(
    product: Product,
    date_from: datetime | None,
    date_to: datetime | None,
):
    sale_rows = _sale_metric_rows(
        product,
        date_from,
        date_to,
        "sale__channel",
        "sale__currency",
    )
    sales = {(row["sale__channel"], row["sale__currency"]): row for row in sale_rows}

    cost_rows = (
        Cost.objects.filter(Q(product=product) | Q(sale_line__product=product))
        .filter(_date_filter("occurred_at", date_from, date_to))
        .annotate(
            effective_channel=Case(
                When(
                    channel="",
                    then=Coalesce("sale_line__sale__channel", Value("")),
                ),
                default=F("channel"),
                output_field=CharField(),
            )
        )
        .values("effective_channel", "currency", "cost_type")
        .annotate(amount=Coalesce(Sum("amount"), ZERO_MONEY))
        .order_by("effective_channel", "currency", "cost_type")
    )
    costs: dict[tuple[str, str], dict[str, Decimal]] = defaultdict(dict)
    for row in cost_rows:
        costs[(row["effective_channel"], row["currency"])][row["cost_type"]] = row[
            "amount"
        ]

    result = []
    for channel, currency in sorted(set(sales) | set(costs)):
        sale = sales.get((channel, currency), {})
        gross = sale.get("gross", ZERO_MONEY)
        discounts = sale.get("discounts", ZERO_MONEY)
        cost_of_goods = sale.get("cost_of_goods", ZERO_MONEY)
        expense_breakdown = costs.get((channel, currency), {})
        additional_costs = sum(expense_breakdown.values(), ZERO_MONEY)
        net_revenue = gross - discounts
        result.append(
            {
                "channel": channel or "unattributed",
                "currency": currency,
                "quantity": sale.get("quantity", Decimal("0")),
                "net_revenue": net_revenue,
                "cost_of_goods": cost_of_goods,
                "additional_costs": additional_costs,
                "contribution_margin": net_revenue - cost_of_goods - additional_costs,
                "cost_breakdown": expense_breakdown,
                "missing_cost_lines": sale.get("missing_cost_lines", 0),
                "cost_complete": sale.get("missing_cost_lines", 0) == 0,
            }
        )
    return result


def _latest_stock(product: Product):
    latest_pk = (
        StockSnapshot.objects.filter(
            product=product,
            source_id=OuterRef("source_id"),
            warehouse_external_id=OuterRef("warehouse_external_id"),
        )
        .order_by("-captured_at", "-created_at")
        .values("pk")[:1]
    )
    snapshots = (
        StockSnapshot.objects.filter(product=product, pk=Subquery(latest_pk))
        .select_related("source")
        .order_by("source__code", "warehouse_external_id")
    )

    return [
        {
            "source": snapshot.source.code,
            "warehouse_external_id": snapshot.warehouse_external_id,
            "warehouse_name": snapshot.warehouse_name,
            "quantity": snapshot.quantity,
            "reserved": snapshot.reserved,
            "available": snapshot.available,
            "in_transit": snapshot.in_transit,
            "captured_at": snapshot.captured_at,
        }
        for snapshot in snapshots
    ]


def get_product_metrics(
    product: Product,
    *,
    date_from: datetime | None = None,
    date_to: datetime | None = None,
):
    sales = _sale_metrics(product, date_from, date_to)
    costs = _cost_metrics(product, date_from, date_to)
    currencies = sorted(set(sales) | set(costs))
    monetary = []
    for currency in currencies:
        sale = sales.get(currency, {})
        gross = sale.get("gross", ZERO_MONEY)
        discounts = sale.get("discounts", ZERO_MONEY)
        cost_of_goods = sale.get("cost_of_goods", ZERO_MONEY)
        expense_breakdown = costs.get(currency, {})
        additional_costs = sum(expense_breakdown.values(), ZERO_MONEY)
        net_revenue = gross - discounts
        monetary.append(
            {
                "currency": currency,
                "quantity": sale.get("quantity", Decimal("0")),
                "gross_revenue": gross,
                "discounts": discounts,
                "net_revenue": net_revenue,
                "cost_of_goods": cost_of_goods,
                "additional_costs": additional_costs,
                "contribution_margin": net_revenue - cost_of_goods - additional_costs,
                "cost_breakdown": expense_breakdown,
                "missing_cost_lines": sale.get("missing_cost_lines", 0),
                "cost_complete": sale.get("missing_cost_lines", 0) == 0,
            }
        )

    return {
        "product_id": product.id,
        "period": {"from": date_from, "to": date_to},
        "monetary": monetary,
        "by_channel": _channel_metrics(product, date_from, date_to),
        "stock": _latest_stock(product),
    }