added ptp() and associated case test

theano/tensor/basic.py

@@ -5000,3 +5000,27 @@ def stacklists(arg):
         return stack(*map(stacklists, arg))
     else:
         return arg
+
+def ptp(a, axis=None):
+    """
+    Range of values (maximum - minimum) along an axis.
+
+    The name of the function comes from the acronym for peak to peak.	
+
+    :param a : Input tensor.
+
+    :param axis : Axis along which to find the peaks. By default, flatten the array.
+
+    :return : A new array holding the result, unless out was specified, in which case a reference to out is returned.
+    """
+    
+    a = as_tensor_variable(a)
+    
+    out = max (a, axis)-min(a, axis)
+   
+    return out
+    
+    
+    
+    
+

theano/tensor/tests/test_basic.py

@@ -45,7 +45,7 @@ from theano.tensor import (_shared, wvector, bvector, autocast_float_as,
         dtensor3, SpecifyShape, Mean,
         itensor3, Tile, switch, Diagonal, Diag,
         nonzero, flatnonzero, nonzero_values,
-        stacklists, DimShuffle, hessian)
+        stacklists, DimShuffle, hessian,ptp)
 
 from theano.tests import unittest_tools as utt
 
@@ -6777,6 +6777,62 @@ def test_norm():
     f = theano.function([x], n)
     assert numpy.allclose(f([1, 1]), numpy.sqrt(2))
 
+class test_ptp(unittest.TestCase):
+    def test_scalar(self):
+        """
+        Should return 0 for all scalar
+        """
+        x = scalar('x')
+        p = ptp(x)
+        f = theano.function([x], p)
+        self.assertTrue(f(rand()*20000-10000)==0)
+        
+    def test_vector(self):
+        
+        x = vector('x')
+        p = ptp(x,0)
+        f = theano.function([x], p)
+        
+        y = rand_ranged(-1000, 1000, {100})
+        result = f(y)
+        maxLessMin=[numpy.amax(y)-numpy.amin(y)]
+        
+        self.assertTrue(result==maxLessMin)
+    
+    def test_matrix_first_axis(self):
+        
+        x = matrix('x')
+        p = ptp(x,1)
+        f = theano.function([x], p)
+        
+        y = rand_ranged(-1000, 1000, [100,100])
+        result = f(y)
+        maxLessMin=[numpy.amax(i)-numpy.amin(i) for i in y]
+        
+        self.assertTrue(numpy.array_equal(result,maxLessMin))
+        
+    def test_matrix_second_axis(self):
+        x = matrix('x')
+        p = ptp(x,0)
+        f = theano.function([x], p)
+        
+        y = rand_ranged(-1000, 1000, [100,100])
+        result = f(y)
+        y = numpy.swapaxes(y, 0,1)
+        maxLessMin=[numpy.amax(i)-numpy.amin(i) for i in y]
+        
+        self.assertTrue(numpy.array_equal(result,maxLessMin))
+        
+    def test_matrix_neg_axis(self):
+        x = matrix('x')
+        p = ptp(x,-1)
+        f = theano.function([x], p)
+        
+        y = rand_ranged(-1000, 1000, [100,100])
+        result = f(y)
+        maxLessMin=[numpy.amax(i)-numpy.amin(i) for i in y]
+        
+	
 if __name__ == '__main__':
 
     t = TestInferShape('setUp')