Coverage Summary for Class: MonthlyNet (com.georgeci.moneysurfer.domain.model)

 package com.georgeci.moneysurfer.domain.model
 
 import com.georgeci.moneysurfer.domain.primitives.CategoryId
 import com.georgeci.moneysurfer.domain.primitives.Money
 import kotlinx.datetime.LocalDate
 import kotlinx.datetime.YearMonth
 
 /*
  * Return shapes of SpendAnalyticsRepository.
  *
  * All of them are magnitudes in the workspace base currency: the queries behind them sum
  * `ABS(amount)`, so an expense reads as a positive number rather than a negative one. Nothing
  * here is persisted or pushed — a change to the maths never needs a migration.
  */
 
 /**
  * What one category booked inside a spend window.
  *
  * [categoryId] is `null` for the uncategorized bucket. That is a real slice, not a missing one:
  * `Transaction.categoryId` is nullable, and a workspace with half its rows uncategorized would
  * otherwise show a donut that does not add up to what was spent.
  */
 data class CategorySpendSlice(
     val categoryId: CategoryId?,
     val total: Money,
     val transactionCount: Int,
 )
 
 /**
  * One month of the income-vs-expense trend. Both figures are magnitudes, so [net] is the
  * subtraction rather than a sum of signed rows.
  *
  * A month with income but no spend still produces a row — the query groups by month and type,
  * and the missing side folds to zero here rather than dropping the column from the chart.
  *
  * [month] names the calendar month the rows fell in; it does *not* promise the whole month was
  * counted. A query window that starts or ends mid-month yields a part-month row indistinguishable
  * from a full one — see
  * [com.georgeci.moneysurfer.domain.repositories.SpendAnalyticsRepository.netByMonth].
  */
 data class MonthlyNet(
     val month: YearMonth,
     val income: Money,
     val expense: Money,
 ) {
     /** Negative once the month spent more than it earned. */
     val net: Money get() = income - expense
 }
 
 /**
  * One day of the spend series. Days with nothing booked are absent — the heatmap and the burn
  * rate both need to tell "no spend" from "no data", and only the caller knows which window it
  * asked for.
  */
 data class DailySpendPoint(
     val date: LocalDate,
     val total: Money,
 )
 
 /**
  * What one merchant took inside a spend window. [merchant] is never blank — see
  * [com.georgeci.moneysurfer.domain.repositories.SpendAnalyticsRepository.topMerchants].
  */
 data class MerchantSpend(
     val merchant: String,
     val total: Money,
     val transactionCount: Int,
 )