Add a step to create an evenly-spaced version of categorical values

pdgstaging/ConfigManager.py

@@ -122,6 +122,13 @@ class ConfigManager():
             - simplify_tolerance : float
                 The tolerance to use when simplifying the input polygons.
                 Defaults to 0.0001. Set to None to skip simplification.
+            - categorical_values: dict
+                A dictionary mapping from the names of peroperties in the input
+                dataset representing categorical values to the order that the
+                values can be sorted into in order to be represented with
+                evenly-spaced values. This will be stored in a new column in
+                the staged dataset, with the prefix 'staged_normalized_' added
+                to the property's original name.
 
         - Tiling & rasterization options.
             - tms_id : str

pdgstaging/TileStager.py

@@ -135,6 +135,7 @@ def stage(self, path):
             gdf = self.set_crs(gdf)
             self.grid = self.make_tms_grid(gdf)
             gdf = self.add_properties(gdf, path)
+            gdf = self.normalize_categorical_values(gdf)
             self.save_tiles(gdf)
         else:
             logger.warning(f'No features in {path}')
@@ -422,6 +423,34 @@ def assign_tile(self, gdf):
             predicate='intersects',
             as_tile=True)
 
+    def normalize_categorical_values(self, gdf):
+        """
+            Add a column converting each 'categorical_value' defined in the
+            config to an evenly-spaced set of values.
+
+            Parameters
+            ----------
+            gdf: GeoDataFrame
+                The GeoDataFrame to add normalize the values for.
+
+            Returns
+            -------
+            GeoDataFrame
+                The GeoDataFrame with the normalized values added as a new
+                column.
+        """
+        categorical_values = self.config.get('categorical_values')
+        for categorical_value in categorical_values.items():
+            prop = categorical_value[0]
+            order = categorical_value[1]
+            value_map = {}
+            i = 1
+            for element in order:
+                value_map[element] = i
+                i += 1
+            gdf['staged_normalized_' + prop] = gdf[prop].map(value_map)
+        return gdf
+
     def make_tms_grid(self, gdf):
         """
             Create a GeoDataFrame of rectangles that creates a TMS grid that