From 68e9c85c8ee2c2f3c646c0aeb3ae6d4f69077245 Mon Sep 17 00:00:00 2001
From: "sarah.rautenbach" <sarah.rautenbach@deltares.nl>
Date: Tue, 5 Mar 2024 08:30:38 +0100
Subject: [PATCH] define function to split composite area based on aggregation

---
 hydromt_fiat/workflows/aggregation_areas.py | 58 +++++++++++----------
 1 file changed, 30 insertions(+), 28 deletions(-)

diff --git a/hydromt_fiat/workflows/aggregation_areas.py b/hydromt_fiat/workflows/aggregation_areas.py
index b783a5a4..0fb2ed9e 100644
--- a/hydromt_fiat/workflows/aggregation_areas.py
+++ b/hydromt_fiat/workflows/aggregation_areas.py
@@ -69,34 +69,7 @@ def join_exposure_aggregation_multiple_areas(
         exposure_gdf.drop_duplicates(subset="Object ID", keep="first", inplace=True)
         
         if new_composite_area[0] is True:
-            if aggregated[attribute_name].apply(lambda x: len(x) >= 2).any():
-                # If new composite area falls in multiple aggregation zones should do an overlay function and where they overlay, split the composite area into
-                # polygons and assign the aggregation area
-                new_exposure_gdf = exposure_gdf.rename(columns = {attribute_name: "pot"})
-                res_intersection = new_exposure_gdf.overlay(aggregation_gdf[[attribute_name, 'geometry']], how='intersection')
-                res_intersection.drop(["index_right", "pot"], axis = 1, inplace = True)
-
-                # exposure area - res_intersection = left over shape in no zone 
-                exposure_outside_aggregation = new_exposure_gdf.overlay(res_intersection[["geometry"]], how = "symmetric_difference")
-                exposure_outside_aggregation.drop(["index_right", "pot"], axis = 1, inplace = True)
-                exposure_outside_aggregation = exposure_outside_aggregation.dropna()
-
-                # Combine divided objects of new composite area within aggregation zone and outside #TODO create OBJECT IDs
-                exposure_gdf = pd.concat([res_intersection, exposure_outside_aggregation], ignore_index=True)
-                idx_duplicates = exposure_gdf.index[exposure_gdf.duplicated("Object ID") == True]
-                exposure_gdf["Object ID"] = exposure_gdf["Object ID"].astype(int)
-                exposure_gdf.loc[idx_duplicates, "Object ID"] = np.random.choice(range(exposure_gdf["Object ID"].values.max() + 1 ,exposure_gdf["Object ID"].values.max() +1 + len(idx_duplicates)), size=len(idx_duplicates), replace=False)
-                        
-                # Create an empty GeoDataFrame and append the exposure data
-                try:
-                    new_exposure_aggregation
-                except NameError:
-                    data = pd.DataFrame(columns=['geometry'])
-                    final_exposure = gpd.GeoDataFrame(data, geometry='geometry')
-                    new_exposure_aggregation = pd.concat([final_exposure, exposure_gdf], ignore_index=True) 
-                    exposure_gdf = new_exposure_aggregation
-                else:
-                    exposure_gdf = pd.concat([new_exposure_aggregation, exposure_gdf], ignore_index=True)            
+            split_composite_area(exposure_gdf, aggregated, aggregation_gdf, attribute_name)         
         else:
             exposure_gdf.drop(columns=attribute_name, inplace=True)
             exposure_gdf = exposure_gdf.merge(aggregated, on="Object ID")
@@ -116,6 +89,35 @@ def join_exposure_aggregation_multiple_areas(
 
     return exposure_gdf
 
+def split_composite_area(exposure_gdf, aggregated, aggregation_gdf, attribute_name):
+    if aggregated[attribute_name].apply(lambda x: len(x) >= 2).any():
+    # If new composite area falls in multiple aggregation zones should do an overlay function and where they overlay, split the composite area into
+        # polygons and assign the aggregation area
+        new_exposure_gdf = exposure_gdf.rename(columns = {attribute_name: "pot"})
+        res_intersection = new_exposure_gdf.overlay(aggregation_gdf[[attribute_name, 'geometry']], how='intersection')
+        res_intersection.drop(["index_right", "pot"], axis = 1, inplace = True)
+
+        # exposure area - res_intersection = left over shape in no zone 
+        exposure_outside_aggregation = new_exposure_gdf.overlay(res_intersection[["geometry"]], how = "symmetric_difference")
+        exposure_outside_aggregation.drop(["index_right", "pot"], axis = 1, inplace = True)
+        exposure_outside_aggregation = exposure_outside_aggregation.dropna()
+
+        # Combine divided objects of new composite area within aggregation zone and outside #TODO create OBJECT IDs
+        exposure_gdf = pd.concat([res_intersection, exposure_outside_aggregation], ignore_index=True)
+        idx_duplicates = exposure_gdf.index[exposure_gdf.duplicated("Object ID") == True]
+        exposure_gdf["Object ID"] = exposure_gdf["Object ID"].astype(int)
+        exposure_gdf.loc[idx_duplicates, "Object ID"] = np.random.choice(range(exposure_gdf["Object ID"].values.max() + 1 ,exposure_gdf["Object ID"].values.max() +1 + len(idx_duplicates)), size=len(idx_duplicates), replace=False)
+                
+        # Create an empty GeoDataFrame and append the exposure data
+    try:
+        new_exposure_aggregation
+    except NameError:
+        data = pd.DataFrame(columns=['geometry'])
+        final_exposure = gpd.GeoDataFrame(data, geometry='geometry')
+        new_exposure_aggregation = pd.concat([final_exposure, exposure_gdf], ignore_index=True) 
+        exposure_gdf = new_exposure_aggregation
+    else:
+        exposure_gdf = pd.concat([new_exposure_aggregation, exposure_gdf], ignore_index=True) 
 
 def join_exposure_aggregation_areas(
     exposure_gdf: gpd.GeoDataFrame,