Dockerfile

@@ -4,6 +4,17 @@ FROM python:3.13-slim
 # Ustawiamy katalog roboczy
 WORKDIR /app
 
+# Zależności systemowe do OCR, obrazów, tesseract i języka PL
+RUN apt-get update && apt-get install -y --no-install-recommends \
+    tesseract-ocr \
+    tesseract-ocr-pol \
+    libglib2.0-0 \
+    libsm6 \
+    libxrender1 \
+    libxext6 \
+    && apt-get clean \
+    && rm -rf /var/lib/apt/lists/*
+
 # Kopiujemy wymagania
 COPY requirements.txt requirements.txt
 

app.py

@@ -9,6 +9,9 @@ import psutil
 import secrets
 import hashlib
 
+import re
+import tempfile
+
 from pillow_heif import register_heif_opener
 
 from datetime import datetime, timedelta, UTC, timezone
@@ -49,6 +52,11 @@ from sqlalchemy import func, extract
 from collections import defaultdict, deque
 from functools import wraps
 
+# OCR
+from collections import Counter
+import pytesseract
+
+
 app = Flask(__name__)
 app.config.from_object(Config)
 register_heif_opener()  # pillow_heif dla HEIC
@@ -335,6 +343,89 @@ def _receipt_error(message):
     return redirect(request.referrer or url_for("main_page"))
 
 
+############# OCR ###########################
+
+
+def preprocess_image_for_tesseract(pil_image):
+    import cv2
+    import numpy as np
+    from PIL import Image
+
+    # Konwersja PIL.Image → NumPy grayscale
+    image = np.array(pil_image.convert("L"))
+
+    # Zwiększenie skali dla lepszej czytelności OCR
+    image = cv2.resize(image, None, fx=2.0, fy=2.0, interpolation=cv2.INTER_LINEAR)
+
+    # Adaptacyjne progowanie (lepsze niż THRESH_BINARY przy nierównym tle)
+    image = cv2.adaptiveThreshold(
+        image, 255, cv2.ADAPTIVE_THRESH_MEAN_C, cv2.THRESH_BINARY, blockSize=15, C=10
+    )
+
+    # Konwersja z powrotem na PIL.Image (dla pytesseract)
+    return Image.fromarray(image)
+
+
+def extract_total_tesseract(image):
+
+    text = pytesseract.image_to_string(image, lang="pol", config="--psm 6")
+    lines = text.splitlines()
+    candidates = []
+
+    fuzzy_regex = re.compile(r"[\dOo][.,:;g9zZ][\d]{2}")
+
+    for line in lines:
+        if not line.strip():
+            continue
+
+        matches = re.findall(r"\d{1,4}[.,]\d{2}", line)
+        for match in matches:
+            try:
+                val = float(match.replace(",", "."))
+                if 0.1 <= val <= 100000:
+                    candidates.append((val, line))
+            except:
+                continue
+
+        fuzzy_matches = fuzzy_regex.findall(line)
+        for match in fuzzy_matches:
+            cleaned = (
+                match.replace("O", "0")
+                .replace("o", "0")
+                .replace(":", ".")
+                .replace(";", ".")
+                .replace(",", ".")
+                .replace("g", "9")
+                .replace("z", "9")
+                .replace("Z", "9")
+            )
+            try:
+                val = float(cleaned)
+                if 0.1 <= val <= 100:
+                    candidates.append((val, line))
+            except:
+                continue
+
+    preferred = [
+        val
+        for val, line in candidates
+        if re.search(r"sum[aąo]?|razem|zapłaty", line.lower())
+    ]
+
+    if preferred:
+        max_val = round(max(preferred), 2)
+        return max_val, lines
+
+    if candidates:
+        max_val = round(max([val for val, _ in candidates]), 2)
+        return max_val, lines
+
+    return 0.0, lines
+
+
+############# END OCR #######################
+
+
 # zabezpieczenie logowani do systemu - błędne hasła
 def is_ip_blocked(ip):
     now = time.time()
@@ -1037,6 +1128,59 @@ def reorder_items():
     return jsonify(success=True)
 
 
+# OCR
+@app.route("/lists/<int:list_id>/analyze", methods=["POST"])
+@login_required
+def analyze_receipts_for_list(list_id):
+    list_obj = db.session.get(ShoppingList, list_id)
+    if not list_obj or list_obj.owner_id != current_user.id:
+        return jsonify({"error": "Brak dostępu"}), 403
+
+    receipt_objs = Receipt.query.filter_by(list_id=list_id).all()
+    results = []
+    total = 0.0
+
+    for receipt in receipt_objs:
+        filepath = os.path.join(app.config["UPLOAD_FOLDER"], receipt.filename)
+        if not os.path.exists(filepath):
+            continue
+
+        temp_path = None
+
+        try:
+            if filepath.lower().endswith(".webp"):
+
+                raw_image = Image.open(filepath).convert("RGB")
+                image = preprocess_image_for_tesseract(raw_image)
+            else:
+
+                raw_image = Image.open(filepath).convert("RGB")
+                image = preprocess_image_for_tesseract(raw_image)
+
+            value, lines = extract_total_tesseract(image)
+
+        except Exception as e:
+            print(f"OCR error for {receipt.filename}: {e}")
+            value = 0.0
+            lines = []
+
+        finally:
+            if temp_path and os.path.exists(temp_path):
+                os.unlink(temp_path)
+
+        results.append(
+            {
+                "id": receipt.id,
+                "filename": receipt.filename,
+                "amount": round(value, 2),
+                "debug_text": lines,
+            }
+        )
+        total += value
+
+    return jsonify({"results": results, "total": round(total, 2)})
+
+
 @app.route("/admin")
 @login_required
 @admin_required

config.py

@@ -10,6 +10,13 @@ class Config:
     DEFAULT_ADMIN_PASSWORD = os.environ.get("DEFAULT_ADMIN_PASSWORD", "admin123")
     UPLOAD_FOLDER = os.environ.get("UPLOAD_FOLDER", "uploads")
     AUTHORIZED_COOKIE_VALUE = os.environ.get("AUTHORIZED_COOKIE_VALUE", "cookievalue")
-    AUTH_COOKIE_MAX_AGE = int(os.environ.get("AUTH_COOKIE_MAX_AGE", 86400))
+    try:
+        AUTH_COOKIE_MAX_AGE = int(os.environ.get("AUTH_COOKIE_MAX_AGE", "86400") or "86400")
+    except ValueError:
+        AUTH_COOKIE_MAX_AGE = 86400
+
     HEALTHCHECK_TOKEN = os.environ.get("HEALTHCHECK_TOKEN", "alamapsaikota1234")
-    SESSION_TIMEOUT_MINUTES = int(os.environ.get("SESSION_TIMEOUT_MINUTES", 10080))
+    try:
+        SESSION_TIMEOUT_MINUTES = int(os.environ.get("SESSION_TIMEOUT_MINUTES", "10080") or "10080")
+    except ValueError:
+        SESSION_TIMEOUT_MINUTES = 10080

requirements.txt

@@ -7,4 +7,7 @@ eventlet
 Werkzeug
 Pillow
 psutil
-pillow-heif
+pillow-heif
+
+pytesseract
+opencv-python-headless

static/js/receipt_analysis.js

@@ -0,0 +1,63 @@
+document.addEventListener("DOMContentLoaded", () => {
+    const analyzeBtn = document.getElementById("analyzeBtn");
+    if (analyzeBtn) {
+        analyzeBtn.addEventListener("click", () => analyzeReceipts(LIST_ID));
+    }
+});
+
+async function analyzeReceipts(listId) {
+    const resultsDiv = document.getElementById("analysisResults");
+    resultsDiv.innerHTML = `<div class="text-info">⏳ Trwa analiza paragonów...</div>`;
+
+    const start = performance.now(); // ⏱ START
+
+    try {
+        const res = await fetch(`/lists/${listId}/analyze`, { method: "POST" });
+        const data = await res.json();
+
+        const duration = ((performance.now() - start) / 1000).toFixed(2); // ⏱ STOP
+
+        let html = `<p><b>📊 Łącznie wykryto:</b> ${data.total.toFixed(2)} PLN</p>`;
+        html += `<p class="text-secondary"><small>⏱ Czas analizy OCR: ${duration} sek.</small></p>`;
+
+        data.results.forEach((r, i) => {
+            html += `
+        <div class="mb-2">
+          <span class="text-light">${r.filename}</span>:
+          <input type="number" id="amount-${i}" value="${r.amount}" step="0.01" class="form-control d-inline-block bg-dark text-white border-light rounded" style="width: 120px;">
+          <button onclick="emitExpense(${i})" class="btn btn-sm btn-outline-success ms-2">➕ Dodaj</button>
+        </div>`;
+        });
+
+        if (data.results.length > 1) {
+            html += `<button onclick="emitAllExpenses(${data.results.length})" class="btn btn-success mt-3">➕ Dodaj wszystkie</button>`;
+        }
+
+        resultsDiv.innerHTML = html;
+        window._ocr_results = data.results;
+
+    } catch (err) {
+        resultsDiv.innerHTML = `<div class="text-danger">❌ Wystąpił błąd podczas analizy.</div>`;
+        console.error(err);
+    }
+}
+
+
+function emitExpense(i) {
+    const r = window._ocr_results[i];
+    const val = parseFloat(document.getElementById(`amount-${i}`).value);
+    if (!isNaN(val) && val > 0) {
+        socket.emit('add_expense', {
+            list_id: LIST_ID,
+            amount: val
+
+        });
+        document.getElementById(`amount-${i}`).disabled = true;
+    }
+}
+
+function emitAllExpenses(n) {
+    for (let i = 0; i < n; i++) {
+        emitExpense(i);
+    }
+}

templates/list_share.html

@@ -106,6 +106,21 @@
 <div class="collapse" id="receiptSection">
   {% set receipt_pattern = 'list_' ~ list.id %}
 
+  {% if receipt_files %}
+  <hr>
+  <div class="mt-3 p-3 border border-secondary rounded bg-dark text-white" id="receiptAnalysisBlock">
+    <h5>🧠 Analiza paragonów (OCR)</h5>
+    <p class="text-small">System spróbuje automatycznie rozpoznać kwoty z dodanych paragonów.</p>
+
+    <button id="analyzeBtn" class="btn btn-outline-info mb-3">
+      🔍 Zleć analizę OCR
+    </button>
+
+    <div id="analysisResults" class="mt-2"></div>
+  </div>
+  {% endif %}
+
+
   <h5 class="mt-4">📸 Paragony dodane do tej listy</h5>
 
   <div class="row g-3 mt-2" id="receiptGallery">
@@ -192,7 +207,7 @@
 <script src="{{ url_for('static_bp.serve_js', filename='clickable_row.js') }}"></script>
 <script src="{{ url_for('static_bp.serve_js', filename='receipt_section.js') }}"></script>
 <script src="{{ url_for('static_bp.serve_js', filename='receipt_upload.js') }}"></script>
-
+<script src="{{ url_for('static_bp.serve_js', filename='receipt_analysis.js') }}"></script>
 <script>
   setupList({{ list.id }}, '{{ current_user.username if current_user.is_authenticated else 'Gość' }}');
 </script>