Update app.py

app.py

@@ -1,9 +1,9 @@
 """
-Biomass Prediction Gradio App with Feature Engineering
+Biomass Prediction Gradio App with Exact 99 Features
 Author: najahpokkiri
 Date: 2025-05-17
 
-This version ensures proper feature engineering to match the model's expected 99 features.
+This updated version ensures the model receives exactly the 99 expected features.
 """
 import os
 import sys
@@ -75,10 +75,15 @@ class BiomassPredictorApp:
                 
                 logger.info(f"Package keys: {list(self.package.keys())}")
                 logger.info(f"Model expects {n_features} features")
+                
+                # Verify feature count is 99
+                if n_features != 99:
+                    logger.warning(f"Warning: Model expects {n_features} features, not the expected 99. This may cause issues.")
+                
             except Exception as e:
                 logger.error(f"Error loading package file: {e}")
                 # Fallback to default values
-                n_features = 99  # Default to expected feature count
+                n_features = 99  # We know there are 99 features
                 self.feature_names = [f"feature_{i}" for i in range(n_features)]
                 
                 # Create a minimal package with essential components
@@ -156,28 +161,22 @@ class BiomassPredictorApp:
                 logger.info(f"Processing image: {height}x{width} pixels, {image.shape[0]} bands")
                 
                 # Validate minimum band count
-                if image.shape[0] < 59:
-                    return None, f"Error: Image has {image.shape[0]} bands, but model expects at least 59 bands."
+                if image.shape[0] < 1:
+                    return None, f"Error: Image has no bands. Please use multi-band satellite imagery."
                 
                 # Generate all features using feature engineering
-                logger.info("Generating features from bands...")
+                logger.info("Generating all 99 features from bands...")
                 feature_matrix, valid_mask, generated_features = extract_all_features(image)
                 
-                # Check if we have the expected number of features
-                n_expected_features = self.package['n_features']
-                if feature_matrix.shape[1] != n_expected_features:
-                    logger.warning(f"Generated {feature_matrix.shape[1]} features, but model expects {n_expected_features}")
-                    # If too many features, truncate
-                    if feature_matrix.shape[1] > n_expected_features:
-                        logger.info(f"Truncating to first {n_expected_features} features")
-                        feature_matrix = feature_matrix[:, :n_expected_features]
-                    # If too few features, can't proceed
-                    else:
-                        return None, f"Error: Generated only {feature_matrix.shape[1]} features, but model expects {n_expected_features}."
+                # Verify we have exactly 99 features
+                if feature_matrix.shape[1] != 99:
+                    logger.error(f"Error: Generated {feature_matrix.shape[1]} features, but model expects 99.")
+                    return None, f"Error: Generated {feature_matrix.shape[1]} features, but model expects 99."
                 
                 # Apply feature scaling if available
                 try:
                     if 'scaler' in self.package and self.package['scaler'] is not None:
+                        logger.info("Applying feature scaling...")
                         feature_matrix = self.package['scaler'].transform(feature_matrix)
                 except Exception as e:
                     logger.warning(f"Error applying scaler: {e}. Using original features.")
@@ -220,7 +219,8 @@ class BiomassPredictorApp:
                             predictions[y_idx, x_idx] = pred
                         
                         # Log progress
-                        logger.info(f"Processed {end_idx}/{len(valid_y)} pixels")
+                        if (i // batch_size) % 5 == 0 or end_idx == len(valid_y):
+                            logger.info(f"Processed {end_idx}/{len(valid_y)} pixels")
                 
                 # Create visualization
                 logger.info("Creating visualization...")
@@ -339,9 +339,8 @@ class BiomassPredictorApp:
             Upload a multi-band satellite image to predict above-ground biomass (AGB) across the landscape.
             
             **Requirements:**
-            - Image must be a GeoTIFF with at least 59 spectral bands
-            - The app will automatically generate additional features needed by the model
-            - For best results, image bands should match the training data order
+            - Image must be a GeoTIFF with spectral bands
+            - For best results, use imagery with at least 59 bands or similar to training data
             """)
             
             with gr.Row():
@@ -378,18 +377,17 @@ class BiomassPredictorApp:
                 ### Model Details
                 
                 - Architecture: StableResNet
-                - Input: Multi-spectral satellite imagery (59 bands + 40 engineered features)
+                - Input: Multi-spectral satellite imagery
                 - Output: Above-ground biomass (Mg/ha)
                 - Creator: najahpokkiri
                 - Date: 2025-05-17
                 
                 ### How It Works
                 
-                1. The app extracts 59 bands from the input satellite image
-                2. Additional features are generated (spectral indices, texture, PCA components)
-                3. The combined features are processed through the StableResNet model
-                4. The model outputs a biomass prediction for each pixel
-                5. Results are visualized as a heatmap or RGB overlay
+                1. The model extracts features from each pixel in the satellite image
+                2. These features include spectral bands, vegetation indices, texture metrics, and more
+                3. The model outputs a biomass prediction for each pixel
+                4. Results are visualized as a heatmap or RGB overlay
                 """)
             
             # Add a warning if model failed to load
@@ -414,7 +412,7 @@ def launch_app():
         # Create interface
         interface = app.create_interface()
         
-        # Launch interface
+        # Launch interface - Important: no share=True in Hugging Face Spaces
         interface.launch()
     except Exception as e:
         logger.error(f"Error launching app: {e}")