Merge pull request #237 from knaaptime/floatfix

numpy float is deprecated

Author: knaaptime

segregation/dynamics/segregation_profile.py

@@ -1,10 +1,11 @@
 """Compute multiscalar segregation profiles."""
 
 import warnings
+from warnings import warn
 
 import numpy as np
 import pandas as pd
-from libpysal.weights import Kernel
+from libpysal.graph import Graph
 from pyproj.crs import CRS
 
 
@@ -19,7 +20,8 @@ def compute_multiscalar_profile(
     decay="linear",
     function="triangular",
     precompute=True,
-    **kwargs
+    check_crs=True,
+    **kwargs,
 ):
     """Compute multiscalar segregation profile.
 
@@ -56,11 +58,13 @@ def compute_multiscalar_profile(
     precompute: bool
         Whether the pandana.Network instance should precompute the range
         queries. This is True by default
+    check_crs:
+        warn if geodataframe is not stored in WGS coordinates. Can help debug
+        errors when computed index is zero indicating a poor merge. Default is True
     **kwargs : dict
         additional keyword arguments passed to each index (e.g. for setting a random
         seed in indices like ModifiedGini or ModifiedDissm)
 
-
     Returns
     -------
     pandas.Series
@@ -89,13 +93,16 @@ def compute_multiscalar_profile(
     with warnings.catch_warnings():
         warnings.simplefilter("ignore")
         if network:
-            if not gdf.crs.equals(CRS(4326)):
-                gdf = gdf.to_crs(epsg=4326)
+            distances = np.array(distances).astype(float)
+            if check_crs:
+                if not gdf.crs.equals(CRS(4326)):
+                    warn(
+                        "GeoDataFrame is in CRS {gdf.crs}. Ensure your network is also stored in this system"
+                    )
             if precompute:
                 maxdist = max(distances)
                 network.precompute(maxdist)
             for distance in distances:
-                distance = np.float(distance)
                 if group_pop_var:
                     idx = segregation_index(
                         gdf,
@@ -105,7 +112,7 @@ def compute_multiscalar_profile(
                         decay=decay,
                         distance=distance,
                         precompute=False,
-                        **kwargs
+                        **kwargs,
                     )
                 elif groups:
                     idx = segregation_index(
@@ -115,20 +122,33 @@ def compute_multiscalar_profile(
                         decay=decay,
                         distance=distance,
                         precompute=False,
-                        **kwargs
+                        **kwargs,
                     )
 
                 indices[distance] = idx.statistic
         else:
             for distance in distances:
-                w = Kernel.from_dataframe(gdf, bandwidth=distance, function=function)
+                funcs = [
+                    "triangular",
+                    "parabolic",
+                    "gaussian",
+                    "bisquare",
+                    "cosine",
+                    "boxcar",
+                    "identity",
+                ]
+                if function not in funcs:
+                    raise ValueError(f"function must be one of {funcs}")
+                w = Graph.build_kernel(
+                    gdf.set_geometry(gdf.centroid), bandwidth=distance, kernel=function
+                ).to_W()
                 if group_pop_var:
                     idx = segregation_index(
                         gdf,
                         group_pop_var=group_pop_var,
                         total_pop_var=total_pop_var,
                         w=w,
-                        **kwargs
+                        **kwargs,
                     )
                 else:
                     idx = segregation_index(gdf, groups, w=w, **kwargs)

segregation/inference/comparative.py

@@ -329,7 +329,7 @@ def _prepare_random_label(seg_class_1, seg_class_2):
         data_1.loc[:, (seg_class_1.group_pop_var, seg_class_1.total_pop_var)] = (
             data_1.loc[:, (seg_class_1.group_pop_var, seg_class_1.total_pop_var)]
             .round(0)
-            .astype(int)
+            .astype(float)
         )
 
         # random permutation needs the columns to have the same names
@@ -345,7 +345,7 @@ def _prepare_random_label(seg_class_1, seg_class_2):
         data_2.loc[:, (seg_class_2.group_pop_var, seg_class_2.total_pop_var)] = (
             data_2.loc[:, (seg_class_2.group_pop_var, seg_class_2.total_pop_var)]
             .round(0)
-            .astype(int)
+            .astype(float)
         )
         data_2 = data_2[
             [
@@ -362,8 +362,8 @@ def _prepare_random_label(seg_class_1, seg_class_2):
         groups_list = seg_class_1.groups
 
         for i in range(len(groups_list)):
-            data_1[groups_list[i]] = round(data_1[groups_list[i]]).astype(int)
-            data_2[groups_list[i]] = round(data_2[groups_list[i]]).astype(int)
+            data_1[groups_list[i]] = round(data_1[groups_list[i]]).astype(float)
+            data_2[groups_list[i]] = round(data_2[groups_list[i]]).astype(float)
 
         if seg_class_1.groups != seg_class_2.groups:
             raise ValueError("MultiGroup groups should be the same")