Update app.py

app.py

@@ -3,282 +3,52 @@ import numpy as np
 import streamlit as st
 from PIL import Image
 import tensorflow as tf
-from tensorflow.keras.preprocessing.image import img_to_array
-from tensorflow.keras.applications.efficientnet import preprocess_input as eff_pre
 
-# ==================== PAGE CONFIG ====================
-st.set_page_config(
-    page_title="Fish Classification App",
-    page_icon="🐟",
-    layout="centered"
-)
+# Set page config
+st.set_page_config(page_title="Fish Classifier", page_icon="🐟")
 
-# ==================== TITLE & DESCRIPTION ====================
-st.markdown("""
-<div style="text-align: center;">
-    <h1>🐟 Fish Species Classifier</h1>
-    <p>Upload an image of a fish to identify its species using AI</p>
-</div>
-""", unsafe_allow_html=True)
+st.title("🐟 Fish Species Classifier")
 
-# ==================== CONSTANTS ====================
-IMG_SIZE = (256, 256)  # Must match training size
+# List all files to find the model
+st.write("Searching for model file...")
 
-CLASS_NAMES = [
-    "animal fish",
-    "animal fish bass",
-    "fish sea_food black_sea_sprat",
-    "fish sea_food gilt_head_bream",
-    "fish sea_food hourse_mackerel",
-    "fish sea_food red_mullet",
-    "fish sea_food red_sea_bream",
-    "fish sea_food sea_bass",
-    "fish sea_food shrimp",
-    "fish sea_food striped_red_mullet",
-    "fish sea_food trout"
+# Look for model in common locations
+possible_paths = [
+    "EfficientNetB0_head_finetuned (1).keras",
+    "EfficientNetB0_head_finetuned.keras",
+    "model.keras",
+    "fish_model.keras",
+    "EfficientNetB0.keras",
 ]
 
-# ==================== MODEL LOADING ====================
-@st.cache_resource
-def load_model_from_hf():
-    """
-    Load model from Hugging Face Space files
-    Two options: Direct download or from local if already in space
-    """
-    # Try multiple possible model locations
-    model_paths = [
-        "EfficientNetB0_head_finetuned.keras",  # Direct in space root
-        "fish_model/EfficientNetB0_head_finetuned.keras",  # In folder
-        "models/EfficientNetB0_head_finetuned.keras",  # Another possible location
-        "EfficientNetB0_head_finetuned (1).keras",  # Your uploaded filename
-    ]
-    
-    for model_path in model_paths:
-        if os.path.exists(model_path):
-            try:
-                st.info(f"Loading model from: {model_path}")
-                model = tf.keras.models.load_model(model_path, compile=False)
-                st.success(f"✅ Model loaded successfully from {model_path}")
-                return model
-            except Exception as e:
-                st.warning(f"Failed to load {model_path}: {str(e)[:100]}")
-    
-    # If not found locally, try to download from URL
-    try:
-        st.info("Model not found locally. Trying to download...")
-        
-        # You can add a direct download link if your model is hosted somewhere
-        # For example, if you uploaded to GitHub or other hosting
-        model_url = "YOUR_MODEL_URL_HERE"  # Replace with actual URL
-        
-        import urllib.request
-        local_filename = "downloaded_model.keras"
-        urllib.request.urlretrieve(model_url, local_filename)
-        
-        model = tf.keras.models.load_model(local_filename, compile=False)
-        st.success("✅ Model downloaded and loaded successfully!")
-        return model
-    except:
-        st.error("❌ Could not load model. Please ensure the model file is uploaded to this Space.")
-        return None
+found_model = None
+for path in possible_paths:
+    if os.path.exists(path):
+        st.success(f"✅ Found model at: {path}")
+        found_model = path
+        break
 
-# Load the model
-model = load_model_from_hf()
-
-if model is None:
-    st.warning("""
-    **Setup Instructions:**
-    1. Upload your `EfficientNetB0_head_finetuned.keras` file to this Space
-    2. Click the **"Files"** tab in the top navigation
-    3. Click **"Add file"** → **"Upload file"**
-    4. Select your model file and upload
-    5. The app will automatically detect it
-    """)
+if not found_model:
+    st.error("❌ Model file not found!")
+    st.write("Please upload your model file to the Space.")
     st.stop()
 
-# ==================== IMAGE PREPROCESSING ====================
-def prepare_image(pil_img):
-    """Preprocess image for EfficientNet model"""
-    # Convert to RGB and resize
-    pil_img = pil_img.convert("RGB")
-    pil_img = pil_img.resize((IMG_SIZE[1], IMG_SIZE[0]))  # (width, height)
-    
-    # Convert to array and expand dimensions
-    arr = img_to_array(pil_img)          # → (H, W, 3)
-    arr = np.expand_dims(arr, axis=0)    # → (1, H, W, 3)
-    
-    # Apply EfficientNet preprocessing
-    arr = eff_pre(arr)
-    return arr
-
-# ==================== PREDICTION ====================
-def predict_top1(img):
-    """Get top-1 prediction"""
-    x = prepare_image(img)
-    preds = model.predict(x, verbose=0)[0]  # vector of 11 probabilities
-    top_index = np.argmax(preds)           # best class index
-    
-    return CLASS_NAMES[top_index], float(preds[top_index])
-
-def predict_top3(img):
-    """Get top-3 predictions"""
-    x = prepare_image(img)
-    preds = model.predict(x, verbose=0)[0]
-    
-    # Get top 3 indices
-    top_indices = np.argsort(preds)[-3:][::-1]
-    
-    results = []
-    for idx in top_indices:
-        results.append({
-            'class': CLASS_NAMES[idx],
-            'confidence': float(preds[idx])
-        })
-    
-    return results
-
-# ==================== SIDEBAR ====================
-with st.sidebar:
-    st.markdown("### ℹ️ About")
-    st.markdown("""
-    This app classifies fish species using a deep learning model.
-    
-    **Supported Classes:**
-    - Animal Fish
-    - Bass
-    - Black Sea Sprat
-    - Gilt-head Bream
-    - Horse Mackerel
-    - Red Mullet
-    - Red Sea Bream
-    - Sea Bass
-    - Shrimp
-    - Striped Red Mullet
-    - Trout
-    """)
-    
-    st.markdown("### ⚙️ Settings")
-    show_top3 = st.checkbox("Show top 3 predictions", value=True)
-
-# ==================== MAIN INTERFACE ====================
-st.markdown("### 📤 Upload Fish Image")
-uploaded_file = st.file_uploader(
-    "Choose an image...", 
-    type=["jpg", "jpeg", "png", "bmp"],
-    help="Upload an image of a fish for classification"
-)
-
-col1, col2 = st.columns([1, 1])
-with col1:
-    # Example images section
-    st.markdown("### 📸 Example Images")
-    st.markdown("Try with these examples:")
-    
-    example_images = {
-        "Fish 1": "https://images.unsplash.com/photo-1578662996442-48f60103fc96?w=400",
-        "Fish 2": "https://images.unsplash.com/photo-1550358864-518f202c02ba?w-400",
-        "Fish 3": "https://images.unsplash.com/photo-1560279966-8ff2f6c81d26?w=400"
-    }
-    
-    for name, url in example_images.items():
-        if st.button(f"Use {name}"):
-            # In real implementation, you'd download and use the image
-            st.info(f"Would use {name} example (implement image download)")
-
-with col2:
-    # Model info
-    st.markdown("### 🧠 Model Info")
-    st.info(f"""
-    **Model:** EfficientNetB0 Fine-tuned
-    **Input Size:** {IMG_SIZE[0]}x{IMG_SIZE[1]}
-    **Classes:** {len(CLASS_NAMES)} species
-    **File:** EfficientNetB0_head_finetuned.keras
-    """)
+# Try to load the model
+try:
+    model = tf.keras.models.load_model(found_model, compile=False)
+    st.success("✅ Model loaded successfully!")
+except Exception as e:
+    st.error(f"❌ Failed to load model: {str(e)[:200]}")
+    st.stop()
 
-# ==================== PREDICTION SECTION ====================
-if uploaded_file is not None:
-    # Display uploaded image
-    image = Image.open(uploaded_file)
-    
-    st.markdown("---")
-    col1, col2 = st.columns(2)
-    
-    with col1:
-        st.image(image, caption="Uploaded Image", use_container_width=True)
-    
-    with col2:
-        st.markdown("### 🔍 Analysis")
-        
-        if st.button("🔍 Classify Image", type="primary", use_container_width=True):
-            with st.spinner("Analyzing image..."):
-                # Get predictions
-                top1_class, top1_conf = predict_top1(image)
-                
-                # Display top-1 result
-                st.markdown(f"### 🎯 **Prediction:** {top1_class}")
-                st.markdown(f"#### **Confidence:** {top1_conf*100:.2f}%")
-                
-                # Confidence bar
-                st.progress(float(top1_conf))
-                
-                # Show top-3 if enabled
-                if show_top3:
-                    st.markdown("### 📊 Top 3 Predictions")
-                    top3_results = predict_top3(image)
-                    
-                    for i, result in enumerate(top3_results):
-                        col_a, col_b = st.columns([3, 2])
-                        with col_a:
-                            st.write(f"{i+1}. {result['class']}")
-                        with col_b:
-                            st.write(f"{result['confidence']*100:.1f}%")
-                        
-                        # Progress bar for each prediction
-                        st.progress(float(result['confidence']))
-    
-    # Download button for results
-    st.markdown("---")
-    if st.button("📥 Download Prediction Results"):
-        results_text = f"Fish Classification Results\n"
-        results_text += f"Image: {uploaded_file.name}\n"
-        results_text += f"Top Prediction: {top1_class} ({top1_conf*100:.2f}%)\n\n"
-        
-        if show_top3:
-            results_text += "Top 3 Predictions:\n"
-            top3 = predict_top3(image)
-            for i, r in enumerate(top3):
-                results_text += f"{i+1}. {r['class']}: {r['confidence']*100:.2f}%\n"
-        
-        # Create download
-        st.download_button(
-            label="Download as Text File",
-            data=results_text,
-            file_name="fish_classification_results.txt",
-            mime="text/plain"
-        )
+# Simple interface
+uploaded = st.file_uploader("Upload fish image", type=["jpg", "png", "jpeg"])
 
-else:
-    # Welcome message when no image uploaded
-    st.markdown("---")
-    st.info("👈 **Upload an image using the file uploader above to get started!**")
+if uploaded:
+    img = Image.open(uploaded)
+    st.image(img, caption="Uploaded Image")
     
-    # Quick tips
-    with st.expander("💡 Tips for Best Results", expanded=True):
-        st.markdown("""
-        1. **Clear Images:** Use well-lit, clear photos
-        2. **Focus on Fish:** Ensure the fish is clearly visible
-        3. **Multiple Angles:** Side views work best
-        4. **File Types:** JPG, PNG, BMP supported
-        5. **Size:** Images are resized to 256x256 pixels
-        
-        The model was trained on 11 fish species using EfficientNetB0 architecture.
-        """)
-
-# ==================== FOOTER ====================
-st.markdown("---")
-st.markdown("""
-<div style="text-align: center; color: gray;">
-    <p>Built with TensorFlow & Streamlit | Model: EfficientNetB0 Fine-tuned</p>
-    <p>Upload your <code>EfficientNetB0_head_finetuned.keras</code> file to get started!</p>
-</div>
-""", unsafe_allow_html=True)
+    if st.button("Classify"):
+        st.write("Processing...")
+        # Add your prediction code here
+        st.success("Classification complete!")