Update app.py

app.py

@@ -229,27 +229,66 @@ def aggregate_data_yearly(collection):
     yearly_images = ee.List(grouped_by_year.map(calculate_yearly_mean))
     return ee.ImageCollection(yearly_images)
 
-# Cloud percentage calculation
-def calculate_cloud_percentage(image, cloud_band='QA60'):
-    qa60 = image.select(cloud_band)
-    opaque_clouds = qa60.bitwiseAnd(1 << 10)
-    cirrus_clouds = qa60.bitwiseAnd(1 << 11)
-    cloud_mask = opaque_clouds.Or(cirrus_clouds)
-    total_pixels = qa60.reduceRegion(
-        reducer=ee.Reducer.count(),
-        geometry=image.geometry(),
-        scale=60,
-        maxPixels=1e13
-    ).get(cloud_band)
-    cloudy_pixels = cloud_mask.reduceRegion(
-        reducer=ee.Reducer.sum(),
-        geometry=image.geometry(),
-        scale=60,
-        maxPixels=1e13
-    ).get(cloud_band)
-    if total_pixels == 0:
-        return 0
-    return ee.Number(cloudy_pixels).divide(ee.Number(total_pixels)).multiply(100)
+# # Cloud percentage calculation
+# def calculate_cloud_percentage(image, cloud_band='QA60'):
+#     qa60 = image.select(cloud_band)
+#     opaque_clouds = qa60.bitwiseAnd(1 << 10)
+#     cirrus_clouds = qa60.bitwiseAnd(1 << 11)
+#     cloud_mask = opaque_clouds.Or(cirrus_clouds)
+#     total_pixels = qa60.reduceRegion(
+#         reducer=ee.Reducer.count(),
+#         geometry=image.geometry(),
+#         scale=60,
+#         maxPixels=1e13
+#     ).get(cloud_band)
+#     cloudy_pixels = cloud_mask.reduceRegion(
+#         reducer=ee.Reducer.sum(),
+#         geometry=image.geometry(),
+#         scale=60,
+#         maxPixels=1e13
+#     ).get(cloud_band)
+#     if total_pixels == 0:
+#         return 0
+#     return ee.Number(cloudy_pixels).divide(ee.Number(total_pixels)).multiply(100)
+
+def filter_images_by_cloud_percentage(collection, max_cloud_percentage):
+    def calculate_cloud_percentage(image):
+        # Extract the QA60 band
+        qa60 = image.select('QA60')
+        
+        # Identify opaque and cirrus clouds
+        opaque_clouds = qa60.bitwiseAnd(1 << 10)
+        cirrus_clouds = qa60.bitwiseAnd(1 << 11)
+        cloud_mask = opaque_clouds.Or(cirrus_clouds)
+        
+        # Count total pixels and cloudy pixels
+        total_pixels = qa60.reduceRegion(
+            reducer=ee.Reducer.count(),
+            geometry=image.geometry(),
+            scale=60,
+            maxPixels=1e13
+        ).get('QA60')
+        
+        cloudy_pixels = cloud_mask.reduceRegion(
+            reducer=ee.Reducer.sum(),
+            geometry=image.geometry(),
+            scale=60,
+            maxPixels=1e13
+        ).get('QA60')
+        
+        # Calculate cloud percentage
+        cloud_percentage = ee.Number(cloudy_pixels).divide(ee.Number(total_pixels)).multiply(100)
+        
+        # Return the cloud percentage as a property of the image
+        return image.set('cloud_percentage', cloud_percentage)
+
+    # Map the cloud percentage calculation over the collection
+    collection_with_clouds = collection.map(calculate_cloud_percentage)
+    
+    # Filter out images with cloud percentage exceeding the threshold
+    filtered_collection = collection_with_clouds.filter(ee.Filter.lt('cloud_percentage', max_cloud_percentage))
+    
+    return filtered_collection
 
 # Preprocessing function
 def preprocess_collection(collection, pixel_cloud_threshold):
@@ -370,7 +409,6 @@ def process_single_geometry(row, start_date_str, end_date_str, dataset_id, selec
             st.error(f"Error retrieving value for {location_name}: {e}")
     return aggregated_results
 
-# Process aggregation
 # Process aggregation
 def process_aggregation(locations_df, start_date_str, end_date_str, dataset_id, selected_bands, reducer_choice, shape_type, aggregation_period, original_lat_col, original_lon_col, custom_formula="", kernel_size=None, include_boundary=None, tile_cloud_threshold=0, pixel_cloud_threshold=0, user_scale=None):
     aggregated_results = []
@@ -379,14 +417,15 @@ def process_aggregation(locations_df, start_date_str, end_date_str, dataset_id,
     progress_text = st.empty()
     start_time = time.time()
 
-    # Log the original collection size for debugging
+    # Fetch the original collection
     raw_collection = ee.ImageCollection(dataset_id) \
         .filterDate(ee.Date(start_date_str), ee.Date(end_date_str))
+
     st.write(f"Original Collection Size: {raw_collection.size().getInfo()}")
 
-    # Apply cloud filtering globally if applicable
+    # Apply cloud filtering at the image level
     if pixel_cloud_threshold > 0:
-        raw_collection = preprocess_collection(raw_collection, pixel_cloud_threshold)
+        raw_collection = filter_images_by_cloud_percentage(raw_collection, pixel_cloud_threshold)
         st.write(f"Filtered Collection Size (After Cloud Masking): {raw_collection.size().getInfo()}")
 
     # Use ThreadPoolExecutor to process each geometry