feat: split a transaction across multiple categories

Один чек часто смешивает несколько категорий трат (гипермаркет: часть — продукты,
часть — бытовая химия). Раньше такую транзакцию приходилось огрублять до одной
категории или дробить вручную на несколько записей. Теперь у транзакции есть
второй режим: 2+ доли (TransactionSplit) с суммами по разным категориям, которые
обязаны в точности складываться в сумму транзакции.

- Новая модель TransactionSplit + миграция.
- split_rule_violation() проверяет инварианты на создании и редактировании.
- category_rollup.py — общий helper, которым теперь пользуются Dashboard, отчёт
  по категориям и бюджеты, учитывая доли сплитов наравне с обычными транзакциями.
- Фильтр по категории в списке транзакций находит и сплит-транзакции.
- Сплиты попадают в полный бэкап/восстановление БД.
- 13 новых backend-тестов.
- Фронтенд: переключатель «Разбить на несколько категорий» в форме транзакции
  с повторяемыми строками и индикатором остатка суммы; строка транзакции в
  списке показывает все категории сплита.
This commit is contained in:
ZProger
2026-08-30 18:42:13 +03:00
parent b045d378f0
commit 0134cdb6bf
20 changed files with 1008 additions and 144 deletions

View File

@@ -70,6 +70,16 @@ class TransactionBackup(BaseModel):
tag_ids: list[int] = Field(default_factory=list)
class TransactionSplitBackup(BaseModel):
model_config = ConfigDict(from_attributes=True)
id: int
transaction_id: int
category_id: int | None
amount: Decimal
note: str | None
class AssetBackup(BaseModel):
model_config = ConfigDict(from_attributes=True)
@@ -202,6 +212,9 @@ class BackupPayload(BaseModel):
# cleanly under the same format version.
tags: list[TagBackup] = Field(default_factory=list)
transactions: list[TransactionBackup]
# Defaulted so a backup exported before transaction splitting existed
# still imports cleanly under the same format version.
transaction_splits: list[TransactionSplitBackup] = Field(default_factory=list)
assets: list[AssetBackup]
asset_valuations: list[AssetValuationBackup]
# Defaulted so a backup exported before crypto holdings existed still

View File

@@ -39,6 +39,37 @@ def transfer_rule_violation(
return None
def split_rule_violation(
*,
type: TransactionType,
amount: Decimal,
category_id: int | None,
split_count: int,
split_total: Decimal | None,
) -> str | None:
"""The split invariants in one place, same shape as transfer_rule_violation
above — checked on create, and against the row *as it would look after
a patch* on update (see routes/transactions.py), so an edit can't leave
a transaction whose splits no longer add up to its own amount.
Takes counts/totals rather than the split objects themselves: the
update path's "splits weren't touched by this patch" case has to check
the existing ORM rows, whose category_id can be None (the category was
since deleted) — nothing here needs a live category to check the sum.
"""
if split_count == 0:
return None
if type == TransactionType.TRANSFER:
return "splits are not valid for transfer transactions"
if category_id is not None:
return "category_id must be omitted when splitting a transaction across categories"
if split_count < 2:
return "splitting a transaction needs at least 2 categories"
if split_total != amount:
return f"split amounts ({split_total}) must add up to the transaction amount ({amount})"
return None
class TransactionFields(BaseModel):
"""A transaction's shape, without the rules that only make sense while
writing one. Reading goes through this: rows already in the database can
@@ -83,8 +114,34 @@ class TransactionBase(TransactionFields):
return self
class TransactionSplitInput(BaseModel):
category_id: int
amount: Decimal = Field(gt=0, max_digits=14, decimal_places=2)
note: str | None = Field(default=None, max_length=200)
class TransactionCreate(TransactionBase):
tag_ids: list[int] = Field(default_factory=list)
# None/omitted -> a normal single-category transaction, unchanged from
# before. 2+ entries -> the amount is divided across categories instead
# of using category_id (which must then be omitted — see
# split_rule_violation). A single entry isn't accepted: that's just
# category_id with extra steps.
splits: list[TransactionSplitInput] | None = None
@model_validator(mode="after")
def _validate_splits(self) -> "TransactionCreate":
splits = self.splits or []
violation = split_rule_violation(
type=self.type,
amount=self.amount,
category_id=self.category_id,
split_count=len(splits),
split_total=sum((s.amount for s in splits), Decimal("0")) if splits else None,
)
if violation:
raise ValueError(violation)
return self
class TransactionUpdate(BaseModel):
@@ -99,6 +156,10 @@ class TransactionUpdate(BaseModel):
date: date_ | None = None
# Omitted -> tags untouched; sent (even as []) -> replaces the full tag set.
tag_ids: list[int] | None = None
# Omitted -> splits untouched; sent (even as []) -> replaces the full
# split set (send [] together with a category_id to turn a split
# transaction back into a normal single-category one).
splits: list[TransactionSplitInput] | None = None
@field_validator("description")
@classmethod
@@ -106,6 +167,16 @@ class TransactionUpdate(BaseModel):
return capitalize_first_letter(value) if value is not None else None
class TransactionSplitRead(BaseModel):
model_config = ConfigDict(from_attributes=True)
id: int
category_id: int | None
category: CategoryRead | None = None
amount: Decimal
note: str | None
class TransactionRead(TransactionFields):
model_config = ConfigDict(from_attributes=True)
@@ -113,6 +184,7 @@ class TransactionRead(TransactionFields):
account: AccountRead
category: CategoryRead | None = None
tags: list[TagRead] = Field(default_factory=list)
splits: list[TransactionSplitRead] = Field(default_factory=list)
class TransactionPage(BaseModel):