Coverage Summary for Class: TransactionCsvFormat (com.georgeci.moneysurfer.domain.csv)

 package com.georgeci.moneysurfer.domain.csv
 
 /**
  * The column layout of a transactions CSV that this codec wrote at some point.
  *
  * The header row *is* the format version: a file's header names its columns in
  * order, so [forHeader] maps it back to the layout it was written with and the
  * decoder reads that layout — no version marker, no per-row `schema_version`
  * column, nothing outside the CSV grid a spreadsheet would mangle.
  *
  * Every layout stays listed here forever, and each one keeps its own exact
  * field count: a record is still rejected unless it carries every column of
  * *its* file's format, so a truncated row is rejected in a [V1] file and in a
  * [V2] file alike. Adding columns therefore means adding an entry here, never
  * relaxing the length check into "missing tail fields default to empty".
  */
 enum class TransactionCsvFormat(val columns: List<TransactionCsvColumn>) {
 
     /** Pre-#260 exports: ends at `transfer_id`, before merchant/tags/recurring rule. */
     V1(V1_COLUMNS),
 
     /** #260's fields appended: merchant, tags, recurring rule. */
     V2(V2_COLUMNS),
 
     /** #399's `split_id` appended — what [TransactionCsvCodec.encode] writes today. */
     V3(V3_COLUMNS),
     ;
 
     val header: List<String> = columns.map { it.header }
 
     private val indexByColumn: Map<TransactionCsvColumn, Int> =
         columns.withIndex().associate { (index, column) -> column to index }
 
     /** Field index of [column] in this layout, or null when the layout predates it. */
     fun indexOf(column: TransactionCsvColumn): Int? = indexByColumn[column]
 
     companion object {
         /**
          * The newest layout — what every export writes and what [decode] assumes
          * when a caller has no header row to resolve ([TransactionCsvCodec.decode]).
          * Newest is the last entry,
          * so a future format only has to be declared, not registered here.
          */
         val Latest: TransactionCsvFormat = entries.last()
 
         /** The layout [header] was written with, or null when it is not one of ours. */
         fun forHeader(header: List<String>): TransactionCsvFormat? =
             entries.find { it.header == header }
     }
 }
 
 /**
  * Each layout's columns, frozen. Every list here describes bytes already sitting
  * in some user's backup file, so none of them may be re-derived from
  * `TransactionCsvColumn.entries`: that would silently re-define an old format
  * the day a column is added, and files written in it would stop matching their
  * own header. A new column means a new list appended below, leaving these as the
  * builds that wrote them left them.
  */
 private val V1_COLUMNS: List<TransactionCsvColumn> = listOf(
     TransactionCsvColumn.Id,
     TransactionCsvColumn.WorkspaceId,
     TransactionCsvColumn.AccountId,
     TransactionCsvColumn.Type,
     TransactionCsvColumn.Status,
     TransactionCsvColumn.AmountMinor,
     TransactionCsvColumn.Currency,
     TransactionCsvColumn.CategoryId,
     TransactionCsvColumn.Note,
     TransactionCsvColumn.OperationAt,
     TransactionCsvColumn.OperationDate,
     TransactionCsvColumn.CreatedAt,
     TransactionCsvColumn.UpdatedAt,
     TransactionCsvColumn.TransferId,
 )
 
 private val V2_COLUMNS: List<TransactionCsvColumn> = V1_COLUMNS + listOf(
     TransactionCsvColumn.Merchant,
     TransactionCsvColumn.Tags,
     TransactionCsvColumn.RecurringRuleId,
 )
 
 private val V3_COLUMNS: List<TransactionCsvColumn> = V2_COLUMNS + listOf(
     TransactionCsvColumn.SplitId,
 )