fix: preserve overlapping import deduplication

Reviewed-on: #40

CHANGELOG.md

@@ -1,5 +1,11 @@
 # Changelog
 
+## [Backend 0.10.3] - 2026-08-24
+
+### Fixed
+
+- Treat VTB savings-account deposits and closures as internal transfers when importing PDF-to-JSON statements.
+
 ## [Backend 0.10.2] - 2026-08-24
 
 ### Fixed

backend/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@family-budget/backend",
-  "version": "0.10.2",
+  "version": "0.10.4",
   "private": true,
   "scripts": {
     "dev": "tsx watch src/app.ts",
@@ -15,6 +15,7 @@
     "test:transactions": "tsx src/services/transactions.test.ts",
     "test:analytics:db": "NODE_ENV=test tsx src/services/analytics.integration.test.ts",
     "test:import:db": "NODE_ENV=test tsx src/services/import.integration.test.ts",
+    "test:import:direction": "tsx src/services/import.test.ts",
     "test:llm": "tsx src/scripts/testLlm.ts"
   },
   "dependencies": {

backend/src/services/import.integration.test.ts

@@ -8,6 +8,21 @@ const makeStatement = (sourceIds: string[]) => ({
   transactions: sourceIds.map((sourceId) => ({ operationAt: '2026-08-20T10:00:00+03:00', amountSigned: 100, commission: 0, description: 'Пополнение', sourceId })),
 });
 
+const duplicateStatement = {
+  schemaVersion: '1.0', bank: 'TEST',
+  statement: { accountNumber: 'fingerprint-test', currency: 'RUB', openingBalance: 0, closingBalance: 200, exportedAt: '2026-08-20T12:00:00+03:00' },
+  transactions: Array.from({ length: 2 }, () => ({ operationAt: '2026-08-20T00:00:00+03:00', amountSigned: 100, commission: 0, description: 'Пополнение' })),
+};
+const overlapStatement = {
+  ...duplicateStatement,
+  statement: { ...duplicateStatement.statement, accountNumber: 'overlap-test' },
+};
+const singleStatement = {
+  ...overlapStatement,
+  statement: { ...overlapStatement.statement, closingBalance: 100 },
+  transactions: overlapStatement.transactions.slice(0, 1),
+};
+
 async function run(): Promise<void> {
   try {
     await importStatement(makeStatement(['first', 'first-second']));
@@ -18,11 +33,32 @@ async function run(): Promise<void> {
     await importStatement(makeStatement(['second']));
     const result = await pool.query('SELECT account_type, status FROM accounts WHERE id = $1', [account.rows[0].id]);
     assert.deepEqual(result.rows[0], { account_type: 'savings', status: 'closed' });
+    const duplicateResult = await importStatement(duplicateStatement);
+    if ('status' in duplicateResult) throw new Error(duplicateResult.message);
+    assert.deepEqual(duplicateResult, {
+      accountId: duplicateResult.accountId,
+      isNewAccount: true,
+      accountNumberMasked: 'fingerp******test',
+      imported: 2,
+      duplicatesSkipped: 0,
+      totalInFile: 2,
+    });
+    await importStatement(singleStatement);
+    const overlapResult = await importStatement(overlapStatement);
+    if ('status' in overlapResult) throw new Error(overlapResult.message);
+    assert.deepEqual(overlapResult, {
+      accountId: overlapResult.accountId,
+      isNewAccount: false,
+      accountNumberMasked: 'overla******test',
+      imported: 1,
+      duplicatesSkipped: 1,
+      totalInFile: 2,
+    });
     console.log('import metadata SQL: OK');
   } finally {
-    await pool.query('DELETE FROM transactions WHERE account_id IN (SELECT id FROM accounts WHERE bank = \'TEST\' AND account_number = \'metadata-test\')');
-    await pool.query("DELETE FROM imports WHERE account_id IN (SELECT id FROM accounts WHERE bank = 'TEST' AND account_number = 'metadata-test')");
-    await pool.query("DELETE FROM accounts WHERE bank = 'TEST' AND account_number = 'metadata-test'");
+    await pool.query("DELETE FROM transactions WHERE account_id IN (SELECT id FROM accounts WHERE bank = 'TEST' AND account_number IN ('metadata-test', 'fingerprint-test', 'overlap-test'))");
+    await pool.query("DELETE FROM imports WHERE account_id IN (SELECT id FROM accounts WHERE bank = 'TEST' AND account_number IN ('metadata-test', 'fingerprint-test', 'overlap-test'))");
+    await pool.query("DELETE FROM accounts WHERE bank = 'TEST' AND account_number IN ('metadata-test', 'fingerprint-test', 'overlap-test')");
     await pool.end();
   }
 }

backend/src/services/import.test.ts

@@ -0,0 +1,12 @@
+import assert from 'node:assert/strict';
+import { computeFingerprint, determineDirection } from './import';
+
+assert.equal(determineDirection(1, 'Перечисление средств на счет N 123 со счета N 456'), 'transfer');
+assert.equal(determineDirection(1, 'Перечисление средств на вклад N 123'), 'transfer');
+assert.equal(determineDirection(-1, 'Перечисление суммы вклада при закрытии'), 'transfer');
+assert.equal(determineDirection(-1, 'Оплата покупки'), 'expense');
+assert.equal(determineDirection(1, 'Выплата процентов'), 'income');
+const duplicateTransaction = { operationAt: '2026-08-20T00:00:00+03:00', amountSigned: 100, commission: 0, description: 'Пополнение' };
+assert.notEqual(computeFingerprint('fingerprint-test', duplicateTransaction, 0), computeFingerprint('fingerprint-test', duplicateTransaction, 1));
+
+console.log('import direction: OK');

backend/src/services/import.ts

@@ -6,14 +6,18 @@ import type { StatementFile, ImportStatementResponse } from '@family-budget/shar
 const TRANSFER_PHRASES = [
   'перевод между своими счетами',
   'перевод средств на счет',
+  'перечисление средств на счет',
+  'перечисление средств на вклад',
+  'перечисление суммы вклада при закрытии',
   'внутри втб',
 ];
 const CASHBACK_KEYWORD = 'зачисление';
 const UUID_RE = /^[0-9a-f]{8}-[0-9a-f]{4}-[1-5][0-9a-f]{3}-[89ab][0-9a-f]{3}-[0-9a-f]{12}$/i;
 
-function computeFingerprint(
+export function computeFingerprint(
   accountNumber: string,
   tx: { operationAt: string; amountSigned: number; commission: number; description: string; sourceId?: string },
+  sourcePosition?: number,
 ): string {
   if (tx.sourceId) {
     const raw = [accountNumber, tx.sourceId.trim()].join('|');
@@ -26,12 +30,13 @@ function computeFingerprint(
     String(tx.amountSigned),
     String(tx.commission),
     tx.description.trim(),
+    ...(sourcePosition === undefined ? [] : [String(sourcePosition)]),
   ].join('|');
   const hash = crypto.createHash('sha256').update(raw, 'utf-8').digest('hex');
   return `sha256:${hash}`;
 }
 
-function determineDirection(amountSigned: number, description: string): string {
+export function determineDirection(amountSigned: number, description: string): string {
   const lower = description.toLowerCase();
   for (const phrase of TRANSFER_PHRASES) {
     if (lower.includes(phrase)) return 'transfer';
@@ -136,7 +141,7 @@ function validateSemantics(data: StatementFile): ValidationError | null {
   const operationIds = new Set<string>();
   for (let i = 0; i < data.transactions.length; i++) {
     const fp = computeFingerprint(data.statement.accountNumber, data.transactions[i]);
-    if (fps.has(fp)) {
+    if (data.transactions[i].sourceId && fps.has(fp)) {
       return { status: 422, error: 'VALIDATION_ERROR', message: `Duplicate fingerprint found within file at transaction index ${i}` };
     }
     fps.add(fp);
@@ -218,9 +223,15 @@ export async function importStatement(
 
     // Insert transactions
     const insertedIds: number[] = [];
+    const fallbackFingerprintOccurrences = new Map<string, number>();
 
     for (const [sourcePosition, tx] of data.transactions.entries()) {
-      const fp = computeFingerprint(data.statement.accountNumber, tx);
+      const fallbackFingerprint = computeFingerprint(data.statement.accountNumber, tx);
+      const occurrence = fallbackFingerprintOccurrences.get(fallbackFingerprint) ?? 0;
+      fallbackFingerprintOccurrences.set(fallbackFingerprint, occurrence + 1);
+      const fp = !tx.sourceId && occurrence > 0
+        ? computeFingerprint(data.statement.accountNumber, tx, occurrence)
+        : fallbackFingerprint;
       const isCashbackCommissionImport =
         tx.amountSigned === 0 &&
         tx.commission > 0 &&

docs/backlog/api_import.md

@@ -91,7 +91,7 @@
 
 - `statement.currency` соответствует допустимому коду валюты (MVP: `"RUB"`).
 - `operationAt` у всех транзакций — валидная дата (парсится без ошибок).
-- Отсутствуют дубликаты fingerprint внутри одного файла.
+- Повторяющиеся `sourceId` внутри одного файла отклоняются; одинаковые операции без `sourceId` различаются по позиции в массиве `transactions`.
 
 Ответ при ошибке:
 
@@ -120,10 +120,11 @@
 Для каждой транзакции вычисляется SHA-256 от полей, соединённых разделителем `|`:
 
 ```text
-accountNumber|operationAt|amountSigned|commission|normalizedDescription
+accountNumber|operationAt|amountSigned|commission|normalizedDescription[|sourcePosition]
 ```
 
 - `normalizedDescription` — `description` после `trim`.
+- `sourcePosition` — порядковый номер повторяющейся операции в массиве `transactions`; добавляется, только если одинаковые операции без `sourceId` повторяются в одном файле.
 - Суммы подставляются в том виде, в котором пришли в JSON (числовое представление).
 - Разделитель `|` исключает коллизии при склейке полей разной длины.