Update kode untuk fitur baru

app.py

@@ -2,6 +2,10 @@ import streamlit as st
 import pickle
 import requests
 from bs4 import BeautifulSoup
+import easyocr
+import numpy as np
+from PIL import Image
+import cv2
 
 # === Custom CSS for better styling ===
 def load_css():
@@ -40,6 +44,16 @@ def load_css():
         margin: 1rem 0;
     }
     
+    /* Camera card styling */
+    .camera-card {
+        background: linear-gradient(135deg, #f8f9fa, #e9ecef);
+        padding: 1.5rem;
+        border-radius: 10px;
+        box-shadow: 0 2px 4px rgba(0, 0, 0, 0.1);
+        border-left: 4px solid #28a745;
+        margin: 1rem 0;
+    }
+    
     /* Results styling - matching the image design */
     .allergen-result {
         padding: 1rem 1.5rem;
@@ -75,6 +89,23 @@ def load_css():
         text-align: center;
     }
     
+    /* OCR result styling */
+    .ocr-result {
+        background: #f8f9fa;
+        padding: 1rem 1.5rem;
+        margin: 1rem 0;
+        border-radius: 10px;
+        border-left: 4px solid #17a2b8;
+        box-shadow: 0 2px 4px rgba(0, 0, 0, 0.05);
+        line-height: 1.6;
+        font-size: 1rem;
+    }
+    
+    .ocr-result strong {
+        color: #495057;
+        font-weight: 600;
+    }
+    
     /* Button styling */
     .stButton > button {
         background: linear-gradient(135deg, #667eea 0%, #764ba2 100%);
@@ -92,6 +123,11 @@ def load_css():
         box-shadow: 0 6px 8px rgba(0, 0, 0, 0.15);
     }
     
+    /* Camera button styling */
+    .camera-button {
+        background: linear-gradient(135deg, #28a745 0%, #20c997 100%) !important;
+    }
+    
     /* Radio button styling */
     .stRadio > div {
         background: white;
@@ -155,6 +191,17 @@ def load_css():
     </style>
     """, unsafe_allow_html=True)
 
+# === Load EasyOCR Reader ===
+@st.cache_resource
+def load_ocr_reader():
+    """Load EasyOCR reader with Indonesian and English language support"""
+    try:
+        reader = easyocr.Reader(['id', 'en'], gpu=False)  # Indonesian and English
+        return reader
+    except Exception as e:
+        st.error(f"❌ Gagal memuat EasyOCR: {str(e)}")
+        return None
+
 # === Load TF-IDF Vectorizer ===
 @st.cache_resource
 def load_vectorizer():
@@ -167,6 +214,37 @@ def load_model():
     with open("saved_models/XGBoost_model.pkl", "rb") as f:
         return pickle.load(f)
 
+# === OCR Text Extraction ===
+def extract_text_from_image(image, reader):
+    """Extract text from image using EasyOCR"""
+    try:
+        # Convert PIL image to numpy array
+        if isinstance(image, Image.Image):
+            image_array = np.array(image)
+        else:
+            image_array = image
+        
+        # Perform OCR
+        results = reader.readtext(image_array)
+        
+        # Extract text from results
+        extracted_texts = []
+        confidence_scores = []
+        
+        for (bbox, text, confidence) in results:
+            if confidence > 0.3:  # Filter out low confidence text
+                extracted_texts.append(text)
+                confidence_scores.append(confidence)
+        
+        # Join all extracted text
+        full_text = " ".join(extracted_texts)
+        avg_confidence = np.mean(confidence_scores) if confidence_scores else 0
+        
+        return full_text, extracted_texts, avg_confidence
+    
+    except Exception as e:
+        return "", [], 0
+
 # === Prediksi ===
 def predict_allergen(model, vectorizer, input_text):
     X_input = vectorizer.transform([input_text])
@@ -228,6 +306,27 @@ def get_ingredients_from_cookpad(url):
     except Exception as e:
         return None, f"Terjadi kesalahan: {str(e)}"
 
+# === Display OCR Results ===
+def display_ocr_results(extracted_text, text_list, confidence):
+    """Display OCR extraction results"""
+    st.markdown("### 📖 Hasil Ekstraksi Teks")
+    
+    if extracted_text.strip():
+        st.markdown(f'''
+        <div class="ocr-result">
+            <strong>📝 Teks yang Terdeteksi:</strong><br>
+            {extracted_text}
+        </div>
+        ''', unsafe_allow_html=True)
+        
+        # Show confidence and individual text elements
+        with st.expander(f"📊 Detail OCR (Confidence: {confidence:.2f})", expanded=False):
+            st.markdown("**Teks Individual yang Terdeteksi:**")
+            for i, text in enumerate(text_list, 1):
+                st.write(f"{i}. {text}")
+    else:
+        st.warning("⚠️ Tidak ada teks yang dapat diekstrak dari gambar. Pastikan gambar memiliki teks yang jelas dan terbaca.")
+
 # === Display results with custom styling matching the image ===
 def display_results(results, probabilities, labels):
     st.markdown("### 🎯 Hasil Analisis Alergen")
@@ -307,7 +406,7 @@ def main():
     st.markdown("""
     <div class="main-header">
         <h1>🥘 Deteksi Alergen Makanan</h1>
-        <p>Analisis kandungan alergen dalam resep makanan dengan teknologi AI</p>
+        <p>Analisis kandungan alergen dalam resep makanan dengan teknologi AI & OCR</p>
     </div>
     """, unsafe_allow_html=True)
     
@@ -325,9 +424,17 @@ def main():
         
         st.markdown("### 💡 Tips Penggunaan")
         st.markdown("""
+        **Input Manual:**
         - Masukkan bahan dengan detail
         - Gunakan nama bahan dalam bahasa Indonesia
-        - Untuk URL Cookpad, pastikan link valid
+        
+        **Kamera OCR:**
+        - Pastikan teks terlihat jelas
+        - Gunakan pencahayaan yang baik
+        - Hindari blur atau teks terpotong
+        
+        **URL Cookpad:**
+        - Pastikan link valid
         - Maksimal 20 URL per analisis
         """)
     
@@ -339,7 +446,7 @@ def main():
         st.markdown("### 🔧 Pilih Metode Input")
         input_mode = st.radio(
             "",
-            ["📝 Input Manual", "🔗 URL Cookpad"],
+            ["📝 Input Manual", "📷 Kamera OCR", "🔗 URL Cookpad"],
             horizontal=True
         )
         
@@ -384,6 +491,52 @@ def main():
                         st.success("✅ Analisis selesai!")
                         display_results(results, probs, labels)
         
+        elif input_mode == "📷 Kamera OCR":
+            st.markdown("### 📷 Deteksi Alergen dari Gambar")
+            
+            # Info card for camera
+            st.markdown("""
+            <div class="camera-card">
+                <strong>📷 Petunjuk Kamera:</strong> Ambil foto langsung dari daftar bahan, kemasan makanan, 
+                atau resep. Pastikan teks terlihat jelas dan pencahayaan memadai untuk hasil OCR terbaik.
+            </div>
+            """, unsafe_allow_html=True)
+            
+            # Camera input
+            camera_image = st.camera_input("📸 Ambil foto dengan kamera")
+            
+            if camera_image is not None:
+                # Display the captured image
+                st.image(camera_image, caption="📷 Gambar yang diambil", use_column_width=True)
+                
+                # Load OCR reader
+                with st.spinner("🔄 Memuat OCR engine..."):
+                    reader = load_ocr_reader()
+                
+                if reader is None:
+                    st.error("❌ Gagal memuat OCR engine. Pastikan EasyOCR telah terinstall.")
+                else:
+                    col_btn1, col_btn2, col_btn3 = st.columns([2, 2, 2])
+                    with col_btn2:
+                        if st.button("🔍 Ekstrak Teks & Analisis", use_container_width=True, key="ocr_analyze"):
+                            # Extract text from image
+                            with st.spinner("📖 Mengekstrak teks dari gambar..."):
+                                extracted_text, text_list, confidence = extract_text_from_image(camera_image, reader)
+                            
+                            if extracted_text.strip():
+                                # Display OCR results
+                                display_ocr_results(extracted_text, text_list, confidence)
+                                
+                                # Analyze allergens
+                                with st.spinner("🔄 Menganalisis alergen..."):
+                                    pred, probs = predict_allergen(model, vectorizer, extracted_text)
+                                    results = dict(zip(labels, pred))
+                                
+                                st.success("✅ Analisis selesai!")
+                                display_results(results, probs, labels)
+                            else:
+                                st.warning("⚠️ Tidak dapat mengekstrak teks dari gambar. Coba ambil foto dengan pencahayaan yang lebih baik dan pastikan teks terlihat jelas.")
+        
         elif input_mode == "🔗 URL Cookpad":
             st.markdown("### 🔗 Analisis dari URL Cookpad")
             
@@ -458,7 +611,7 @@ def main():
     # Footer
     st.markdown("""
     <div class="footer">
-        <p>🔬 Powered by XGBoost & TF-IDF | Made with ❤️ using Streamlit</p>
+        <p>🔬 Powered by XGBoost, TF-IDF & EasyOCR | Made with ❤️ using Streamlit</p>
     </div>
     """, unsafe_allow_html=True)