Coverage Summary for Class: CategorySpendHistoryKt (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.DateTimeUnit
 import kotlinx.datetime.YearMonth
 import kotlinx.datetime.minus
 import kotlinx.datetime.plus
 
 /** One column of the spend trend. [total] is zero for a month the category booked nothing in. */
 data class CategorySpendMonth(
     val month: YearMonth,
     val total: Money,
 )
 
 /** One row of the subcategory breakdown — a direct child, with its own subtree rolled in. */
 data class CategorySpendBreakdown(
     val categoryId: CategoryId,
     val total: Money,
     val transactionCount: Int,
 )
 
 /**
  * Everything the category detail screen needs in numbers, for one category over a fixed window.
  *
  * Totals here are *rolled up*: a parent's figure includes its whole subtree, matching the
  * promise the manage screen makes ("subcategories roll up into their parent"). [subcategories]
  * is empty for a leaf, which is what drives the leaf empty state — a leaf is not an error case,
  * it is the common shape.
  */
 data class CategorySpendHistory(
     val months: List<CategorySpendMonth>,
     val currentMonth: Money,
     val transactionCount: Int,
     val subcategories: List<CategorySpendBreakdown>,
 ) {
     /** Mean over the whole window, including months with nothing in them. */
     val monthlyAverage: Money
         get() = if (months.isEmpty()) {
             Money.zero()
         } else {
             months.fold(Money.zero()) { acc, m ->
                 acc + m.total
             } / months.size
         }
 
     /** Mean per transaction this month, or null when nothing was booked. */
     val perTransaction: Money?
         get() = if (transactionCount <= 0) null else currentMonth / transactionCount
 
     companion object {
         /** Columns the trend draws. Six fits the mockup's bar row without crowding on a phone. */
         const val TREND_MONTHS: Int = 6
 
         val Empty = CategorySpendHistory(
             months = emptyList(),
             currentMonth = Money.zero(),
             transactionCount = 0,
             subcategories = emptyList(),
         )
     }
 }
 
 /** [count] months ending at [anchor] inclusive, oldest first. */
 fun trailingMonths(anchor: YearMonth, count: Int): List<YearMonth> =
     (count - 1 downTo 0).map { anchor.minus(it, DateTimeUnit.MONTH) }
 
 /**
  * Folds raw per-category-per-month rows into the shape the detail screen renders.
  *
  * Pure, and deliberately tolerant of a sparse [totals]: the aggregate query only returns months
  * a category actually booked something in, so every month in [months] is emitted here whether or
  * not a row backs it. Without that the trend would silently shrink to however many months had
  * spend, and a quiet category would draw a two-bar chart labelled "last 6 months".
  *
  * Rows for categories outside [rootId]'s subtree are ignored rather than trusted, so a caller
  * passing a wider query result cannot inflate the total.
  */
 fun buildCategorySpendHistory(
     categories: List<Category>,
     rootId: CategoryId,
     totals: List<CategoryMonthlyTotal>,
     months: List<YearMonth>,
 ): CategorySpendHistory {
     val subtree = CategoryTree.descendantsOf(categories, rootId) + rootId
     val currentMonth = months.lastOrNull()
 
     val byCategory = totals.groupBy { it.categoryId }
     fun rollUp(ids: Set<CategoryId>, month: YearMonth?): Pair<Money, Int> {
         var total = Money.zero()
         var count = 0
         ids.forEach { id ->
             byCategory[id].orEmpty().forEach { row ->
                 if (month == null || row.month == month) {
                     total += row.total
                     count += row.transactionCount
                 }
             }
         }
         return total to count
     }
 
     val trend = months.map { month ->
         CategorySpendMonth(month = month, total = rollUp(subtree, month).first)
     }
     val (thisMonth, thisMonthCount) = rollUp(subtree, currentMonth)
 
     val breakdown = CategoryTree.childrenOf(categories, rootId).map { child ->
         // A child may itself have children — the mockup shows one flat row per direct child, so
         // a grandchild's spend belongs on its parent's row rather than vanishing from the split.
         val childSubtree = CategoryTree.descendantsOf(categories, child.id) + child.id
         val (total, count) = rollUp(childSubtree, currentMonth)
         CategorySpendBreakdown(categoryId = child.id, total = total, transactionCount = count)
     }
 
     return CategorySpendHistory(
         months = trend,
         currentMonth = thisMonth,
         transactionCount = thisMonthCount,
         subcategories = breakdown,
     )
 }
 
 /** The month after this one — the exclusive upper bound of a window ending at it. */
 fun YearMonth.next(): YearMonth = plus(1, DateTimeUnit.MONTH)