From 71d32d72ef887ec6eeba0621b9c9badf5d6ff0e8 Mon Sep 17 00:00:00 2001
From: Pierre HUBERT <pierre.git@communiquons.org>
Date: Sun, 20 Jul 2025 18:07:22 +0200
Subject: [PATCH] Can extract date of expenses

---
 .../lib/routes/scan/scan_screen.dart          | 10 ++--
 moneymgr_mobile/lib/utils/ocr_utils.dart      | 51 ++++++++++++++++---
 2 files changed, 46 insertions(+), 15 deletions(-)

diff --git a/moneymgr_mobile/lib/routes/scan/scan_screen.dart b/moneymgr_mobile/lib/routes/scan/scan_screen.dart
index 5b1f008..884ef21 100644
--- a/moneymgr_mobile/lib/routes/scan/scan_screen.dart
+++ b/moneymgr_mobile/lib/routes/scan/scan_screen.dart
@@ -13,10 +13,10 @@ part 'scan_screen.g.dart';
 
 /// Scan a document & return generated PDF as byte file
 @riverpod
-Future<(Uint8List?, double?)> _scanDocument(Ref ref) async {
+Future<(Uint8List?, BaseExpenseInfo?)> _scanDocument(Ref ref) async {
   final pdf = await scanDocAsPDF();
   final img = await renderPdf(pdfBytes: pdf);
-  final amount = await extractTotalFromBill(img);
+  final amount = await extractInfoFromBill(img);
   return (pdf, amount);
 }
 
@@ -42,11 +42,7 @@ class ScanScreen extends HookConsumerWidget {
       child: switch (scanDocProvider) {
         AsyncData(:final value) when value.$1 != null => ExpenseEditor(
           file: value.$1!,
-          initialData: BaseExpenseInfo(
-            label: null,
-            cost: value.$2 ?? 0.0,
-            time: DateTime.now(),
-          ),
+          initialData: value.$2,
           onFinished: (expense) async {
             await expenses.add(
               info: expense,
diff --git a/moneymgr_mobile/lib/utils/ocr_utils.dart b/moneymgr_mobile/lib/utils/ocr_utils.dart
index bc6fe6e..f454f3c 100644
--- a/moneymgr_mobile/lib/utils/ocr_utils.dart
+++ b/moneymgr_mobile/lib/utils/ocr_utils.dart
@@ -5,9 +5,10 @@ import 'dart:ui' as ui;
 import 'package:flutter/material.dart';
 import 'package:google_mlkit_text_recognition/google_mlkit_text_recognition.dart';
 import 'package:logging/logging.dart';
+import 'package:moneymgr_mobile/services/storage/expenses.dart';
 
-/// Attempt to extract total amount from invoice image
-Future<double?> extractTotalFromBill(Uint8List imgBuff) async {
+/// Attempt to extract information from invoice image
+Future<BaseExpenseInfo?> extractInfoFromBill(Uint8List imgBuff) async {
   final decodedImage = await decodeImageFromList(imgBuff);
 
   final byteData = await decodedImage.toByteData(
@@ -25,20 +26,54 @@ Future<double?> extractTotalFromBill(Uint8List imgBuff) async {
 
   Logger.root.fine("Expense text: ${extractionResult.text}");
 
-  // Check for highest amount on invoice
-  final regexp = RegExp(
+  // Check for highestCost amount on invoice
+  final costRegexp = RegExp(
     r'([0-9]+([ ]*(\\.|,)[ ]*[0-9]{1,2}){0,1})([ \\t\\n]*(EUR|eur|€)|E)',
     multiLine: true,
     caseSensitive: false,
   );
-  var highest = 0.0;
-  for (final match in regexp.allMatches(extractionResult.text)) {
+  var highestCost = 0.0;
+  for (final match in costRegexp.allMatches(extractionResult.text)) {
     if (match.groupCount == 0) continue;
 
     // Process only numeric value
     final value = (match.group(1) ?? "").replaceAll(",", ".");
-    highest = max(highest, double.tryParse(value) ?? 0.0);
+    highestCost = max(highestCost, double.tryParse(value) ?? 0.0);
   }
 
-  return highest == 0.0 ? null : highest;
+  // Check for highestCost amount on invoice
+  final dateRegexp = RegExp(
+    r'([0-3][0-9])(\/|-)([0-1][0-9])(\/|-)(20[0-9]{2})',
+    multiLine: true,
+    caseSensitive: false,
+  );
+  final currDate = DateTime.now();
+  DateTime? newest;
+  for (final match in dateRegexp.allMatches(extractionResult.text)) {
+    if (match.groupCount < 5) continue;
+
+    try {
+      final date = DateTime(
+        int.tryParse(match.group(5)!) ?? currDate.year,
+        int.tryParse(match.group(3)!) ?? currDate.month,
+        int.tryParse(match.group(1)!) ?? currDate.day,
+      );
+
+      if (newest == null) {
+        newest = date;
+      } else {
+        newest = DateTime.fromMillisecondsSinceEpoch(
+          max(newest.millisecondsSinceEpoch, date.millisecondsSinceEpoch),
+        );
+      }
+    } catch (e, s) {
+      Logger.root.warning("Failed to parse date! $e$s");
+    }
+  }
+
+  return BaseExpenseInfo(
+    label: null,
+    cost: highestCost,
+    time: newest?.isBefore(currDate) ?? false ? newest! : currDate,
+  );
 }