Update app/utils.py

app/utils.py

@@ -1,10 +1,13 @@
-# utils.py
 import os
+import pandas as pd
 from transformers import AutoModel, AutoTokenizer
 from PIL import Image, ImageEnhance, ImageFilter
 import torch
 import logging
 from transformers import BertTokenizer
+import nltk
+import requests
+import io
 
 logger = logging.getLogger(__name__)
 
@@ -21,9 +24,8 @@ class OCRModel:
         try:
             logger.info("Initializing OCR model...")
             
-            # Try different tokenizer approaches
+            # محاولة تحميل النموذج
             try:
-                # First try with the standard approach
                 self.tokenizer = AutoTokenizer.from_pretrained(
                     'stepfun-ai/GOT-OCR2_0',
                     trust_remote_code=True,
@@ -31,7 +33,6 @@ class OCRModel:
                 )
             except Exception as e:
                 logger.warning(f"Standard tokenizer failed, trying BertTokenizer: {str(e)}")
-                # Fall back to BertTokenizer if AutoTokenizer fails
                 self.tokenizer = BertTokenizer.from_pretrained(
                     'stepfun-ai/GOT-OCR2_0',
                     trust_remote_code=True
@@ -55,25 +56,24 @@ class OCRModel:
             raise
             
     def preprocess_image(self, image):
-        """Image preprocessing to improve text recognition quality"""
+        """تحسين جودة الصورة لتحسين استخراج النص"""
         try:
-            # Convert image to RGB if it is not already
             if image.mode != 'RGB':
                 image = image.convert('RGB')
 
-            # Improve contrast
+            # تحسين التباين
             enhancer = ImageEnhance.Contrast(image)
             image = enhancer.enhance(1.5)
 
-            # Improve Sharpness
+            # تحسين الحدة
             enhancer = ImageEnhance.Sharpness(image)
             image = enhancer.enhance(1.5)
 
-            # Improve Brightness
+            # تحسين السطوع
             enhancer = ImageEnhance.Brightness(image)
             image = enhancer.enhance(1.2)
 
-            # Apply a filter to soften the image a little.
+            # تطبيق فلتر لتليين الصورة
             image = image.filter(ImageFilter.SMOOTH)
 
             return image
@@ -81,38 +81,131 @@ class OCRModel:
             logger.error(f"Error in image preprocessing: {str(e)}", exc_info=True)
             raise
 
-    def process_image(self, image_stream):
+    def process_image(self, image):
         try:
             logger.info("Starting image processing")
             
-            # Save image temporarily because the model requires a file path.
-            temp_image_path = "temp_image.jpg"
-            
-            # Reset the start pointer for BytesIO
-            image_stream.seek(0)
-            
-            # Open and save the image temporarily.
-            image = Image.open(image_stream).convert('RGB')
+            # معالجة الصورة
             processed_image = self.preprocess_image(image)
+            
+            # حفظ الصورة مؤقتاً للتعامل مع النموذج
+            temp_image_path = "temp_ocr_image.jpg"
             processed_image.save(temp_image_path)
             
-            # ocr 
-            try:
-                result = self.model.chat(self.tokenizer, temp_image_path, ocr_type='format')
-                logger.info(f"Successfully extracted text: {result[:100]}...")
-                
-                # Delete temporary file
-                if os.path.exists(temp_image_path):
-                    os.remove(temp_image_path)
-                
-                return result.strip()
+            # استخراج النص
+            result = self.model.chat(self.tokenizer, temp_image_path, ocr_type='format')
+            logger.info(f"Successfully extracted text: {result[:100]}...")
+            
+            # حذف الملف المؤقت
+            if os.path.exists(temp_image_path):
+                os.remove(temp_image_path)
+            
+            return result.strip()
+            
+        except Exception as e:
+            logger.error(f"Error in OCR processing: {str(e)}", exc_info=True)
+            if os.path.exists(temp_image_path):
+                os.remove(temp_image_path)
+            return f"Error processing image: {str(e)}"
+
+class AllergyAnalyzer:
+    def __init__(self, dataset_path):
+        self.dataset_path = dataset_path
+        self.allergy_dict = self.load_allergy_data()
+        nltk.download('punkt', quiet=True)
+    
+    def load_allergy_data(self):
+        """تحميل بيانات الحساسيات من ملف Excel"""
+        try:
+            df = pd.read_excel(self.dataset_path)
+            allergy_dict = {}
+            
+            for index, row in df.iterrows():
+                allergy = row['Allergy']
+                ingredients = [ingredient for ingredient in row[1:] if pd.notna(ingredient)]
+                allergy_dict[allergy] = ingredients
                 
-            except Exception as e:
-                logger.error(f"Error in OCR processing: {str(e)}", exc_info=True)
-                if os.path.exists(temp_image_path):
-                    os.remove(temp_image_path)
-                raise
+            return allergy_dict
+        except Exception as e:
+            logger.error(f"Error loading allergy data: {str(e)}", exc_info=True)
+            return {}
+    
+    def tokenize_text(self, text):
+        """تقسيم النص إلى كلمات"""
+        tokens = nltk.word_tokenize(text)
+        return [w.lower() for w in tokens if w.isalpha()]
+    
+    def check_database_allergens(self, token, user_allergens):
+        """التحقق من وجود الحساسيات في قاعدة البيانات"""
+        results = []
+        for allergy in user_allergens:
+            if allergy in self.allergy_dict and token in self.allergy_dict[allergy]:
+                results.append(allergy)
+        return results
+    
+    def check_claude_allergens(self, token, allergy, api_key):
+        """الاستعلام من Claude API عن الحساسيات"""
+        prompt = f"""
+You are a professional food safety expert. Analyze if '{token}' contains or is derived from {allergy}. 
+
+Respond ONLY with 'Yes' or 'No'. No explanations.
+"""
+        
+        url = "https://api.anthropic.com/v1/messages"
+        headers = {
+            "x-api-key": api_key,
+            "content-type": "application/json",
+            "anthropic-version": "2023-06-01"
+        }
+        
+        data = {
+            "model": "claude-3-opus-20240229",
+            "messages": [{"role": "user", "content": prompt}],
+            "max_tokens": 10
+        }
+        
+        try:
+            response = requests.post(url, json=data, headers=headers)
+            json_response = response.json()
+            
+            if "content" in json_response and isinstance(json_response["content"], list):
+                return json_response["content"][0]["text"].strip().lower() == 'yes'
+            return False
             
         except Exception as e:
-            logger.error(f"Error in image processing: {str(e)}", exc_info=True)
-            return f"Error processing image: {str(e)}"
+            logger.error(f"Error querying Claude API: {str(e)}")
+            return False
+    
+    def analyze_text(self, text, user_allergens, claude_api_key=None):
+        """تحليل النص للكشف عن الحساسيات"""
+        detected_allergens = set()
+        database_matches = {}
+        claude_matches = {}
+        tokens = self.tokenize_text(text)
+        
+        for token in tokens:
+            # التحقق من قاعدة البيانات أولاً
+            db_results = self.check_database_allergens(token, user_allergens)
+            
+            if db_results:
+                for allergy in db_results:
+                    detected_allergens.add(allergy)
+                    database_matches[allergy] = database_matches.get(allergy, []) + [token]
+            else:
+                # إذا لم توجد في قاعدة البيانات، نستخدم Claude API
+                if claude_api_key:
+                    for allergy in user_allergens:
+                        if self.check_claude_allergens(token, allergy, claude_api_key):
+                            detected_allergens.add(allergy)
+                            claude_matches[allergy] = claude_matches.get(allergy, []) + [token]
+        
+        return {
+            "detected_allergens": list(detected_allergens),
+            "database_matches": database_matches,
+            "claude_matches": claude_matches,
+            "analyzed_tokens": tokens
+        }
+    
+    def get_allergen_list(self):
+        """الحصول على قائمة الحساسيات المعروفة"""
+        return list(self.allergy_dict.keys())