Updating permutations and structure from Levi LOSH LJC code review

esda/local_geary.py

@@ -12,17 +12,17 @@
 )
 
 
-PERMUTATIONS = 999
-SIG = 0.05
-
-
 class Local_Geary(BaseEstimator):
+
     """Local Geary - Univariate"""
 
-    def __init__(self, connectivity=None, labels=False, sig=SIG,
-                 permutations=PERMUTATIONS, n_jobs=1, keep_simulations=True,
+    def __init__(self, connectivity=None, labels=False, sig=0.05,
+                 permutations=999, n_jobs=1, keep_simulations=True,
                  seed=None):
         """
+        Initialize a Local_Geary estimator
+        Arguments
+        ---------
         connectivity     : scipy.sparse matrix object
                            the connectivity structure describing
                            the relationships between observed units.
@@ -40,9 +40,11 @@ def __init__(self, connectivity=None, labels=False, sig=SIG,
                            Default significance threshold used for
                            creation of labels groups.
         permutations     : int
+                           (default=999)
                            number of random permutations for calculation
                            of pseudo p_values
         n_jobs           : int
+                           (default=1)
                            Number of cores to be used in the conditional
                            randomisation. If -1, all available cores are used.
         keep_simulations : Boolean
@@ -77,7 +79,7 @@ def __init__(self, connectivity=None, labels=False, sig=SIG,
         self.keep_simulations = keep_simulations
         self.seed = seed
 
-    def fit(self, x, n_jobs=1, permutations=999):
+    def fit(self, x):
         """
         Arguments
         ---------
@@ -109,10 +111,15 @@ def fit(self, x, n_jobs=1, permutations=999):
 
         w = self.connectivity
         w.transform = 'r'
+
+        permutations = self.permutations
+        sig = self.sig
+        n_jobs = self.n_jobs
+        seed = self.seed
 
         self.localG = self._statistic(x, w)
 
-        if self.permutations:
+        if permutations:
             self.p_sim, self.rlocalG = _crand_plus(
                 z=(x - np.mean(x))/np.std(x),
                 w=w,
@@ -131,16 +138,16 @@ def fit(self, x, n_jobs=1, permutations=999):
             # Outliers
             self.labs[(self.localG < Eij_mean) &
                       (y > y_mean) &
-                      (self.p_sim <= self.sig)] = 1
+                      (self.p_sim <= sig)] = 1
             # Clusters
             self.labs[(self.localG < Eij_mean) &
                       (y < y_mean) &
-                      (self.p_sim <= self.sig)] = 2
+                      (self.p_sim <= sig)] = 2
             # Other
             self.labs[(self.localG > Eij_mean) &
-                      (self.p_sim <= self.sig)] = 3
+                      (self.p_sim <= sig)] = 3
             # Non-significant
-            self.labs[self.p_sim > self.sig] = 4
+            self.labs[self.p_sim > sig] = 4
 
         del (self.keep_simulations, self.n_jobs,
              self.permutations, self.seed, self.rlocalG,
@@ -174,7 +181,6 @@ def _statistic(x, w):
 
 # Note: does not using the scaling parameter
 
-
 @_njit(fastmath=True)
 def _local_geary(i, z, permuted_ids, weights_i, scaling):
     zi, zrand = _prepare_univariate(i, z, permuted_ids, weights_i)

esda/local_geary_mv.py

@@ -7,20 +7,21 @@
 import libpysal as lp
 
 
-PERMUTATIONS = 999
-
-
 class Local_Geary_MV(BaseEstimator):
 
     """Local Geary - Multivariate"""
 
-    def __init__(self, connectivity=None, permutations=PERMUTATIONS):
+    def __init__(self, connectivity=None, permutations=999):
         """
+        Initialize a Local_Geary_MV estimator
+        Arguments
+        ---------
         connectivity     : scipy.sparse matrix object
                            the connectivity structure describing
                            the relationships between observed units.
                            Need not be row-standardized.
         permutations     : int
+                           (default=999)
                            number of random permutations for calculation
                            of pseudo p_values
         Attributes
@@ -36,7 +37,7 @@ def __init__(self, connectivity=None, permutations=PERMUTATIONS):
         self.connectivity = connectivity
         self.permutations = permutations
 
-    def fit(self, variables, permutations=999):
+    def fit(self, variables):
         """
         Arguments
         ---------
@@ -59,6 +60,11 @@ def fit(self, variables, permutations=999):
         >>> guerry = lp.examples.load_example('Guerry')
         >>> guerry_ds = gpd.read_file(guerry.get_path('Guerry.shp'))
         >>> w = libpysal.weights.Queen.from_dataframe(guerry_ds)
+        >>> import libpysal
+        >>> import geopandas as gpd
+        >>> guerry = lp.examples.load_example('Guerry')
+        >>> guerry_ds = gpd.read_file(guerry.get_path('Guerry.shp'))
+        >>> w = libpysal.weights.Queen.from_dataframe(guerry_ds)
         >>> x1 = guerry_ds['Donatns']
         >>> x2 = guerry_ds['Suicids']
         >>> lG_mv = Local_Geary(connectivity=w).fit([x1,x2])
@@ -72,6 +78,8 @@ def fit(self, variables, permutations=999):
 
         self.n = len(variables[0])
         self.w = w
+
+        permutations = self.permutations
 
         # Caclulate z-scores for input variables
         # to be used in _statistic and _crand