Add PyArray_Acorrelate function

PyArray_Correlate function did not compute the usual definition for
correlation: it assumed correlate(a, b) == correlate(b, a), and did not
use conjugate for the 2nd argument. As some people may rely on the
current behavior, we add a PyArray_Acorrelate function and acorrelate
python function to implement the usual definition.

Squashed commit of the following:

commit 1e1e0ff38968934825ad64cddd36fcb21e378954
Author: David Cournapeau <cournape@gmail.com>
Date:   Tue Jun 2 18:59:57 2009 +0900

    Refactor [A]Correlate, so that they share most of the implementation.

commit ffbd54ecf539df508363a725ec5f7529f71a9bb4
Author: David Cournapeau <cournape@gmail.com>
Date:   Tue Jun 2 18:35:21 2009 +0900

    Handle conjugate in correlate.

commit df47fddd8e5142758504dece39b0716b8f46434f
Author: David Cournapeau <cournape@gmail.com>
Date:   Tue Jun 2 17:42:38 2009 +0900

    Correlate and correlate tests better shared.

commit f48de4450b05260e9174cf224302a1c8a1f60331
Author: David Cournapeau <cournape@gmail.com>
Date:   Tue Jun 2 17:39:15 2009 +0900

    Add acorrelate function at the python level.

commit a37cc7ff144b0e6b64f8fd0a8cae4c9e561f4a1c
Author: David Cournapeau <cournape@gmail.com>
Date:   Tue Jun 2 17:21:20 2009 +0900

    Update test for correlate - we will test for inverted output for acorrelate.

commit 4140a30f1b6daf95ce2fa8fa7bd26f827b2fcfd1
Author: David Cournapeau <cournape@gmail.com>
Date:   Tue Jun 2 17:19:14 2009 +0900

    Add PyArray_Acorrelate function in the C API.

commit 082f9d4b0205a9a4f0112d8cfd3852b3907b6e09
Author: David Cournapeau <cournape@gmail.com>
Date:   Mon Jun 1 12:04:12 2009 +0900

    Add simple test for correlate.

commit 7179187953d7cb396a7557a418e6fbcaeddc4211
Author: David Cournapeau <cournape@gmail.com>
Date:   Mon Jun 1 11:57:46 2009 +0900

    Invert output array if we swapped input arrays.

1 files changed, 37 insertions, 2 deletions

diff --git a/numpy/core/numeric.py b/numpy/core/numeric.py
index 26ef27dc9..50e3fd75b 100644
--- a/numpy/core/numeric.py
+++ b/numpy/core/numeric.py
@@ -6,7 +6,7 @@ __all__ = ['newaxis', 'ndarray', 'flatiter', 'ufunc',
            'set_numeric_ops', 'can_cast',
            'asarray', 'asanyarray', 'ascontiguousarray', 'asfortranarray',
            'isfortran', 'empty_like', 'zeros_like',
-           'correlate', 'convolve', 'inner', 'dot', 'outer', 'vdot',
+           'correlate', 'acorrelate', 'convolve', 'inner', 'dot', 'outer', 'vdot',
            'alterdot', 'restoredot', 'roll', 'rollaxis', 'cross', 'tensordot',
            'array2string', 'get_printoptions', 'set_printoptions',
            'array_repr', 'array_str', 'set_string_function',
@@ -579,11 +579,46 @@ def correlate(a,v,mode='valid'):
     --------
     convolve : Discrete, linear convolution of two
                one-dimensional sequences.
-
+    acorrelate: Discrete correlation following the usual signal processing
+               definition for complex arrays, and without assuming that
+               correlate(a, b) == correlate(b, a)
     """
     mode = _mode_from_name(mode)
     return multiarray.correlate(a,v,mode)
 
+def acorrelate(a, v, mode='valid'):
+    """
+    Discrete, linear correlation of two 1-dimensional sequences.
+
+    This function computes the correlation as generally defined in signal
+    processing texts:
+
+    z[k] = sum_n a[n] * conj(v[n+k])
+
+    with a and v sequences being zero-padded where necessary and conj being the
+    conjugate.
+
+    Parameters
+    ----------
+    a, v : array_like
+        Input sequences.
+    mode : {'valid', 'same', 'full'}, optional
+        Refer to the `convolve` docstring.  Note that the default
+        is `valid`, unlike `convolve`, which uses `full`.
+
+    Note
+    ----
+    This is the function which corresponds to matlab xcorr.
+
+    See Also
+    --------
+    convolve : Discrete, linear convolution of two
+               one-dimensional sequences.
+    correlate: Deprecated function to compute correlation
+    """
+    mode = _mode_from_name(mode)
+    return multiarray.acorrelate(a, v, mode)
+
 
 def convolve(a,v,mode='full'):
     """