[Scipy-svn] r2362 - in trunk/Lib/sparse: . tests

[scipy-svn at scipy.org]

Author: rc
Date: 2006-12-07 06:39:01 -0600 (Thu, 07 Dec 2006)
New Revision: 2362

Modified:
   trunk/Lib/sparse/sparse.py
   trunk/Lib/sparse/tests/test_sparse.py
Log:
coo._normalize() update by Nathan Bell (ticket 326)

Modified: trunk/Lib/sparse/sparse.py
===================================================================
--- trunk/Lib/sparse/sparse.py	2006-12-07 01:35:50 UTC (rev 2361)
+++ trunk/Lib/sparse/sparse.py	2006-12-07 12:39:01 UTC (rev 2362)
@@ -2276,19 +2276,20 @@
 
     def _normalize(self, rowfirst=False):
         if rowfirst:
-            l = zip(self.row, self.col, self.data)
-            l.sort()
-            row, col, data = list(itertools.izip(*l))
-            return data, row, col
+            #sort by increasing rows first, columns second
+            if getattr(self, '_is_normalized', None):
+                #columns already sorted, use stable sort for rows
+                P = numpy.argsort(self.row,kind='mergesort')        
+                return self.data[P], self.row[P], self.col[P]
+            else:
+                #nothing already sorted
+                P  = numpy.lexsort(keys=(self.col,self.row))                
+                return self.data[P], self.row[P], self.col[P]
         if getattr(self, '_is_normalized', None):
             return self.data, self.row, self.col
-        l = zip(self.col, self.row, self.data)
-        l.sort()
-        # This breaks when len(self.data) etc == 0.  Does this matter?
-        col, row, data = list(itertools.izip(*l))
-        self.col = asarray(col, intc)
-        self.row = asarray(row, intc)
-        self.data = array(data, self.dtype)
+        #sort by increasing rows first, columns second
+        P  = numpy.lexsort(keys=(self.row,self.col))        
+        self.data,self.row,self.col = self.data[P], self.row[P], self.col[P]
         setattr(self, '_is_normalized', 1)
         return self.data, self.row, self.col
 

Modified: trunk/Lib/sparse/tests/test_sparse.py
===================================================================
--- trunk/Lib/sparse/tests/test_sparse.py	2006-12-07 01:35:50 UTC (rev 2361)
+++ trunk/Lib/sparse/tests/test_sparse.py	2006-12-07 12:39:01 UTC (rev 2362)
@@ -21,8 +21,8 @@
 import random
 from numpy.testing import *
 set_package_path()
-from scipy.sparse import csc_matrix, csr_matrix, dok_matrix, spidentity, \
-        speye, lil_matrix
+from scipy.sparse import csc_matrix, csr_matrix, dok_matrix, coo_matrix, \
+     spidentity, speye, lil_matrix
 from scipy.linsolve import splu
 restore_path()
 
@@ -638,6 +638,38 @@
         b = array([[1, 0, 0], [0, 1, 0], [0, 0, 1]], dtype='d')
         assert_array_equal(a.toarray(), b)
 
+class test_coo(ScipyTestCase):
 
+    def check_normalize( self ):
+        
+        row  = numpy.array([2, 3, 1, 3, 0, 1, 3, 0, 2, 1, 2])
+        col  = numpy.array([0, 1, 0, 0, 1, 1, 2, 2, 2, 2, 1])
+        data = numpy.array([  6.,  10.,   3.,   9.,   1.,   4.,
+                              11.,   2.,   8.,   5.,   7.])
+
+        # coo.todense()
+        #    matrix([[  0.,   1.,   2.],
+        #            [  3.,   4.,   5.],
+        #            [  6.,   7.,   8.],
+        #            [  9.,  10.,  11.]])
+        coo = coo_matrix((data,(row,col)),(4,3))
+
+        ndata,nrow,ncol = coo._normalize(rowfirst=True)
+        assert(zip(nrow,ncol,ndata) == sorted(zip(row,col,data))) #should sort by rows, then cols
+        assert_array_equal(coo.data, data)                        #coo.data has not changed
+        assert_array_equal(coo.row, row)                          #coo.row has not changed
+        assert_array_equal(coo.col, col)                          #coo.col has not changed
+
+
+        ndata,nrow,ncol = coo._normalize(rowfirst=False)
+        assert(zip(ncol,nrow,ndata) == sorted(zip(col,row,data))) #should sort by cols, then rows
+        assert_array_equal(coo.data, ndata)                       #coo.data has changed
+        assert_array_equal(coo.row, nrow)                         #coo.row has changed
+        assert_array_equal(coo.col, ncol)                         #coo.col has changed
+
+        assert_array_equal(coo.tocsr().todense(), coo.todense())
+        assert_array_equal(coo.tocsc().todense(), coo.todense())
+
+
 if __name__ == "__main__":
     ScipyTest().run()