Prioritize using longitude and latitude columns in building files whe…

…n present.

This saves time over parsing building polygon WKTs and deriving centroids from them.

Also, drop duplicate buildings during example generation.

PiperOrigin-RevId: 686532902

src/skai/buildings.py

@@ -41,21 +41,23 @@ def _read_buildings_csv(path: str) -> gpd.GeoDataFrame:
   """
   with tf.io.gfile.GFile(path, 'r') as csv_file:
     df = pd.read_csv(csv_file)
-  if 'geometry' in df.columns:
+  if 'longitude' in df.columns and 'latitude' in df.columns:
+    geometries = gpd.points_from_xy(df['longitude'], df['latitude'])
+    df.drop(columns=['longitude', 'latitude'], inplace=True)
+  elif 'geometry' in df.columns:
     logging.info('Parsing %d WKT strings. This could take a while.', len(df))
     geometries = gpd.GeoSeries.from_wkt(df['geometry'])
     df.drop(columns=['geometry'], inplace=True)
   elif 'wkt' in df.columns:
     logging.info('Parsing %d WKT strings. This could take a while.', len(df))
     geometries = gpd.GeoSeries.from_wkt(df['wkt'])
     df.drop(columns=['wkt'], inplace=True)
-  elif 'longitude' in df.columns and 'latitude' in df.columns:
-    geometries = gpd.points_from_xy(df['longitude'], df['latitude'])
-    df.drop(columns=['longitude', 'latitude'], inplace=True)
   else:
     raise ValueError(f'No geometry information found in file "{path}"')
 
-  return gpd.GeoDataFrame(df, geometry=geometries, crs=4326)
+  geometries = geometries.normalize()
+  gdf = gpd.GeoDataFrame(df, geometry=geometries, crs=4326)
+  return gdf.drop_duplicates()
 
 
 def convert_buildings_file(