Update app.py

app.py

@@ -1,113 +1,28 @@
 import streamlit as st
-from PIL import Image
+from transformers import pipeline
 import os
-import sys
-
-# Try multiple import approaches for transformers
-try:
-    # For newer versions of transformers
-    from transformers import pipeline
-    TRANSFORMERS_AVAILABLE = True
-    print("✅ Transformers imported successfully")
-except ImportError as e:
-    st.error(f"❌ Transformers import error: {e}")
-    TRANSFORMERS_AVAILABLE = False
-
-# Page configuration
-st.set_page_config(
-    page_title="Jerry - Deepfake Detector",
-    page_icon="🕵️",
-    layout="wide",
-    initial_sidebar_state="collapsed"
-)
+from PIL import Image
 
-# Custom CSS for better styling
-st.markdown("""
-<style>
-    .main-header {
-        font-size: 3rem;
-        color: #1f77b4;
-        text-align: center;
-        margin-bottom: 1rem;
-    }
-    .sub-header {
-        font-size: 1.2rem;
-        color: #666;
-        text-align: center;
-        margin-bottom: 2rem;
-    }
-    .verdict-real {
-        padding: 20px;
-        background-color: #d4edda;
-        border: 2px solid #c3e6cb;
-        border-radius: 10px;
-        color: #155724;
-        font-size: 1.5rem;
-        font-weight: bold;
-        text-align: center;
-        margin: 20px 0;
-    }
-    .verdict-fake {
-        padding: 20px;
-        background-color: #f8d7da;
-        border: 2px solid #f5c6cb;
-        border-radius: 10px;
-        color: #721c24;
-        font-size: 1.5rem;
-        font-weight: bold;
-        text-align: center;
-        margin: 20px 0;
-    }
-    .confidence-bar {
-        height: 25px;
-        border-radius: 5px;
-        margin: 10px 0;
-    }
-    .metric-card {
-        padding: 15px;
-        border-radius: 10px;
-        background-color: #f8f9fa;
-        border-left: 4px solid #1f77b4;
-        margin: 10px 0;
-    }
-    .warning-box {
-        padding: 20px;
-        background-color: #fff3cd;
-        border: 2px solid #ffeaa7;
-        border-radius: 10px;
-        color: #856404;
-        margin: 20px 0;
-    }
-</style>
-""", unsafe_allow_html=True)
+# Global variable to store the pipeline
+deepfake_pipe = None
 
-@st.cache_resource
 def load_model():
     """Load the model once and cache it"""
-    if not TRANSFORMERS_AVAILABLE:
-        return None
-        
-    try:
-        st.info("🔄 Loading Deepfake Detection Model... This might take a moment.")
-        pipe = pipeline(
+    global deepfake_pipe
+    if deepfake_pipe is None:
+        st.info("Loading Deepfake Detection Model...")
+        deepfake_pipe = pipeline(
             "image-classification", 
             model="prithivMLmods/Deep-Fake-Detector-v2-Model"
         )
-        st.success("✅ Model loaded successfully!")
-        return pipe
-    except Exception as e:
-        st.error(f"❌ Failed to load model: {str(e)}")
-        return None
+        st.success("Model loaded successfully!")
+    return deepfake_pipe
 
 def predict_deepfake(image):
     """Predict if image is deepfake or real"""
-    if not TRANSFORMERS_AVAILABLE:
-        return None, "Transformers library not available. Please check dependencies."
-    
     try:
+        # Load model if not already loaded
         pipe = load_model()
-        if pipe is None:
-            return None, "Model failed to load. Please check your internet connection and try again."
         
         # Make prediction
         results = pipe(image)
@@ -115,208 +30,79 @@ def predict_deepfake(image):
         # Format results
         prediction = {result['label']: result['score'] for result in results}
         
-        # Validate prediction structure
-        if not prediction:
-            return None, "No prediction results returned. Please try with a different image."
-            
-        return prediction, None
+        # Determine final verdict
+        deepfake_score = prediction.get('Deepfake', 0)
+        realism_score = prediction.get('Realism', 0)
+        
+        if deepfake_score > realism_score:
+            verdict = f"🚨 DEEPFAKE DETECTED ({deepfake_score:.2%} confidence)"
+            color = "red"
+        else:
+            verdict = f"✅ REAL IMAGE ({realism_score:.2%} confidence)"
+            color = "green"
+        
+        return prediction, verdict, color
         
     except Exception as e:
-        return None, f"Prediction failed: {str(e)}"
+        return {"Error": 1.0}, f"❌ Prediction failed: {str(e)}", "red"
 
-def safe_get_prediction(prediction, key, default=0.0):
-    """Safely get value from prediction dictionary"""
-    if prediction is None:
-        return default
-    return prediction.get(key, default)
-
-def show_demo_results():
-    """Show demo results when transformers is not available"""
-    st.markdown("### Confidence Scores")
-    
-    # Realism score
-    col_real, col_real_pct = st.columns([3, 1])
-    with col_real:
-        st.progress(0.78, text="Realism: 78%")
-    with col_real_pct:
-        st.metric("Realism", "78.0%")
-    
-    # Deepfake score
-    col_fake, col_fake_pct = st.columns([3, 1])
-    with col_fake:
-        st.progress(0.22, text="Deepfake: 22%")
-    with col_fake_pct:
-        st.metric("Deepfake", "22.0%")
-    
-    # Verdict
-    st.markdown("### 🎯 Verdict")
-    st.markdown('<div class="verdict-real">✅ REAL IMAGE<br><small>Confidence: 78.0%</small></div>', unsafe_allow_html=True)
+# Streamlit app configuration
+st.set_page_config(
+    page_title="Jerry - Deepfake Detector",
+    page_icon="🕵️",
+    layout="wide"
+)
 
 def main():
-    # Header
-    st.markdown('<h1 class="main-header">🕵️ Jerry - Deepfake Detection Tool</h1>', unsafe_allow_html=True)
-    st.markdown('<p class="sub-header">Upload an image to check if it\'s real or AI-generated!</p>', unsafe_allow_html=True)
-    
-    # Show warning if transformers is not available
-    if not TRANSFORMERS_AVAILABLE:
-        st.markdown("""
-        <div class="warning-box">
-            <h3>⚠️ Dependency Issue Detected</h3>
-            <p>The transformers library could not be imported properly. This is usually due to version compatibility issues.</p>
-            <p><strong>Solution:</strong> Please check that you're using compatible versions of transformers and torch.</p>
-        </div>
-        """, unsafe_allow_html=True)
-    
-    # Sidebar for information
-    with st.sidebar:
-        st.header("ℹ️ About")
-        st.markdown("""
-        **How it works:**
-        - Upload any facial image
-        - Jerry analyzes it using AI
-        - Get instant results with confidence scores
-        
-        **Detection Categories:**
-        - 🟢 **Realism**: Likely a real human face
-        - 🔴 **Deepfake**: AI-generated or manipulated
-        """)
-        
-        st.header("⚙️ Technical Info")
-        st.write(f"**Transformers Available:** {'✅ Yes' if TRANSFORMERS_AVAILABLE else '❌ No'}")
-        st.write("**Model:** Deep-Fake-Detector-v2")
-        
-        if not TRANSFORMERS_AVAILABLE:
-            st.header("🔧 Fix Required")
-            st.markdown("""
-            **Install compatible versions:**
-            ```bash
-            pip install transformers==4.30.0
-            pip install torch==2.0.0
-            pip install pillow==10.0.0
-            ```
-            """)
-        
-        st.markdown("---")
-        st.markdown("Built with ❤️ using Streamlit")
+    st.title("🕵️ Jerry - Deepfake Detection Tool")
+    st.markdown("**Upload an image to check if it's real or AI-generated!**")
     
-    # Main content area
-    col1, col2 = st.columns([1, 1])
+    # Create two columns
+    col1, col2 = st.columns(2)
     
     with col1:
-        st.subheader("📤 Upload Image")
-        
-        # File uploader
+        st.subheader("Upload Image")
         uploaded_file = st.file_uploader(
-            "Choose an image file",
-            type=['jpg', 'jpeg', 'png', 'bmp'],
-            help="Upload a facial image for deepfake detection"
+            "Choose an image",
+            type=['png', 'jpg', 'jpeg'],
+            label_visibility="collapsed"
         )
         
         if uploaded_file is not None:
-            try:
-                # Display the uploaded image
-                image = Image.open(uploaded_file)
-                st.image(image, caption="Uploaded Image", use_column_width=True)
-                
-                # Analyze button
-                if st.button("🔍 Analyze Image", type="primary", use_container_width=True):
-                    if not TRANSFORMERS_AVAILABLE:
-                        st.error("❌ Cannot analyze - transformers library not available. Please check the sidebar for installation instructions.")
-                    else:
-                        with st.spinner("🔄 Analyzing image... This may take a few seconds."):
-                            prediction, error = predict_deepfake(image)
-                            
-                            if error:
-                                st.error(f"❌ {error}")
-                            else:
-                                # Store results in session state
-                                st.session_state.prediction = prediction
-                                st.session_state.image = image
-                                st.session_state.analysis_done = True
-                                
-                                # Show success message
-                                st.success("✅ Analysis complete!")
-            except Exception as e:
-                st.error(f"❌ Error processing image: {str(e)}")
+            # Display uploaded image
+            image = Image.open(uploaded_file)
+            st.image(image, caption="Uploaded Image", use_column_width=True)
     
     with col2:
-        st.subheader("📊 Results")
-        
-        if not TRANSFORMERS_AVAILABLE:
-            st.info("🔧 Transformers library not available. Showing demo results.")
-            show_demo_results()
-            
-        elif st.session_state.get('analysis_done') and 'prediction' in st.session_state:
-            prediction = st.session_state.prediction
-            
-            if prediction is not None:
-                # Safely extract scores using helper function
-                deepfake_score = safe_get_prediction(prediction, 'Deepfake', 0)
-                realism_score = safe_get_prediction(prediction, 'Realism', 0)
-                
-                # Display confidence scores with progress bars
-                st.markdown("### Confidence Scores")
-                
-                # Realism score
-                col_real, col_real_pct = st.columns([3, 1])
-                with col_real:
-                    st.progress(float(realism_score), text=f"Realism: {realism_score:.1%}")
-                with col_real_pct:
-                    st.metric("Realism", f"{realism_score:.1%}")
-                
-                # Deepfake score
-                col_fake, col_fake_pct = st.columns([3, 1])
-                with col_fake:
-                    st.progress(float(deepfake_score), text=f"Deepfake: {deepfake_score:.1%}")
-                with col_fake_pct:
-                    st.metric("Deepfake", f"{deepfake_score:.1%}")
-                
-                # Verdict
-                st.markdown("### 🎯 Verdict")
-                if realism_score > deepfake_score:
-                    st.markdown(f'<div class="verdict-real">✅ REAL IMAGE<br><small>Confidence: {realism_score:.2%}</small></div>', unsafe_allow_html=True)
-                else:
-                    st.markdown(f'<div class="verdict-fake">🚨 DEEPFAKE DETECTED<br><small>Confidence: {deepfake_score:.2%}</small></div>', unsafe_allow_html=True)
-                
-                # Detailed prediction breakdown
-                st.markdown("### 📈 Detailed Analysis")
-                for label, score in prediction.items():
-                    col_label, col_score, col_bar = st.columns([2, 1, 3])
-                    with col_label:
-                        st.write(f"**{label}**")
-                    with col_score:
-                        st.write(f"{float(score):.1%}")
-                    with col_bar:
-                        st.progress(float(score))
-            else:
-                st.error("❌ No valid prediction results available. Please try again.")
-        
-        else:
-            # Placeholder before analysis
-            st.info("👆 Upload an image and click 'Analyze Image' to see results here.")
-            
-            # Sample results preview
-            st.markdown("### Example Output:")
-            st.progress(0.85, text="Realism: 85%")
-            st.progress(0.15, text="Deepfake: 15%")
-            st.markdown('<div class="verdict-real">✅ REAL IMAGE</div>', unsafe_allow_html=True)
+        st.subheader("Detection Results")
+        if uploaded_file is not None:
+            if st.button("🔍 Analyze Image", type="primary"):
+                with st.spinner("Analyzing image..."):
+                    # Convert uploaded file to filepath for pipeline
+                    prediction, verdict, color = predict_deepfake(uploaded_file)
+                    
+                    # Display results
+                    st.write("**Prediction Scores:**")
+                    for label, score in prediction.items():
+                        st.write(f"{label}: {score:.2%}")
+                    
+                    # Display verdict with colored text
+                    st.markdown(f"<p style='color:{color};font-size:18px;'><b>{verdict}</b></p>", unsafe_allow_html=True)
     
-    # Footer
-    st.markdown("---")
-    st.markdown("""
-    <div style='text-align: center; color: #666;'>
-        <p>For best results, use clear facial images. Performance may vary with image quality.</p>
-        <p><small>If you encounter errors, try refreshing the page and ensuring you have a stable internet connection.</small></p>
-    </div>
-    """, unsafe_allow_html=True)
+    # Additional information
+    st.markdown(
+        """
+        ---
+        **How it works:**
+        - Upload any facial image
+        - Jerry analyzes it and gives appropriate verdict
+        - Returns confidence scores for "Deepfake" vs "Realism"
+        
+        **Note:** The model loads only once when first used, then runs quickly for subsequent predictions!
+        """
+    )
 
 if __name__ == "__main__":
-    # Initialize session state
-    if 'prediction' not in st.session_state:
-        st.session_state.prediction = None
-    if 'image' not in st.session_state:
-        st.session_state.image = None
-    if 'analysis_done' not in st.session_state:
-        st.session_state.analysis_done = False
-    
+    print("🚀 Starting Jerry - Deepfake Detection App...")
+    print("Model will load on first prediction...")
     main()