Replace functions related to indirect correlation matrix

pairing/pairing.py

@@ -38,41 +38,30 @@ def generate_direct_correlation(trj, cutoff=1.0):
     return direct_corr
 
 
-def _generate_indirect_connectivity(direct_corr):
+def generate_indirect_connectivity(direct_corr):
     """
-    Genrate indirect correlation matrix from a direct correlation matrix
-
     Parameters
     ----------
-    direct_corr : numpy.ndarray, dtype=np.int32
-        Direct correlation matrix from which an indirect correlation matrix
-        will be generated.
+    direct_corr: np.ndarray, dtype=np.int32
+    direct correlation matrix
 
     Returns
     -------
-    indirect_corr : numpy.ndarray, dtype=np.int32
-        Indirect corrlation matrix
+    indirect: np.ndarray, dtype=np.int32
+    indirect connectivity matrix
     """
-
     c = deepcopy(direct_corr)
-    size = np.shape(direct_corr)
-    if size[0] != size[1]:
-        raise ValueError('Direct correlation matrix must be square')
-    length = size[0]
-
-    for combo in itertools.combinations([_ for _ in range(length)], 2):
-        for i in range(length):
-            vals = [c[i, combo[0]], c[i, combo[1]]]
-            if vals == [0, 0]:
-                continue
-            elif vals == [1, 1]:
-                intersect = _find_intersection(c[:, combo[0]], c[:, combo[1]])
-                c[:, combo[0]] = intersect
-                c[:, combo[1]] = intersect
-                continue
-
-    indirect_corr = c
-    return indirect_corr
+    for row in c:
+        ones = np.where(row == 1)[0]
+        if len(ones) == 1:
+            continue
+        else:
+            intersect = np.maximum.reduce(c[:, ones].T)
+            for ele in ones:
+                c[:, ele] = intersect
+    indirect = c
+
+    return indirect
 
 
 def generate_clusters(indirect):
@@ -114,68 +103,3 @@ def analyze_clusters(clusters):
     avg = np.mean(cluster_sizes)
     stdev = np.std(cluster_sizes)
     return avg, stdev
-
-
-def _find_intersection(a, b):
-    """
-    Find set intersection of two arrays
-
-    Parameters
-    ----------
-    a : array-like
-        First array to compare
-    b : array-like
-        Second array to compare
-
-    Returns
-    -------
-    intersection : array-like
-        Set intersection of a and b
-    """
-
-    intersection = np.maximum(a, b)
-    return intersection
-
-
-def _check_validity(c_I):
-    """
-    Check validity of indirect connectivity matrix
-
-    Parameters
-    ----------
-    c_I : np.ndarray
-    indirect connectivity matrix to test
-
-    Returns
-    -------
-    Boolean 'True' or 'False'
-    """
-
-    test_indirect = _generate_indirect_connectivity(c_I)
-    return (test_indirect == c_I).all()
-
-
-def new_generate_indirect(direct_corr):
-    """
-    Iteratively call '_generate_indirect_connectivity' and
-    '_check_validity' to generate valid indirect correlation
-    matrices
-
-    Parameters
-    ----------
-    direct_corr : np.ndarray
-    direct correlation matrix
-
-    Returns
-    _______
-    new_indirect : np.ndarray
-    indirect connectivity matrix
-    """
-
-    new_indirect = _generate_indirect_connectivity(
-            direct_corr)
-    while not _check_validity(new_indirect):
-        new_indirect = _generate_indirect_connectivity(
-                new_indirect)
-
-    return new_indirect

pairing/tests/test_pairing.py

@@ -35,41 +35,21 @@ def test_sevick1988():
                       [0, 0, 0, 1, 0],
                       [1, 1, 1, 0, 1]], dtype=np.int32)
 
-    assert (c_I == pairing.pairing._generate_indirect_connectivity(c_D)).all()
-
-
-def test_check_validity_pass():
-    c_I = np.asarray([[1, 1, 1, 0, 1],
-                      [1, 1, 1, 0, 1],
-                      [1, 1, 1, 0, 1],
-                      [0, 0, 0, 1, 0],
-                      [1, 1, 1, 0, 1]], dtype=np.int32)
-
-    assert pairing.pairing._check_validity(c_I)
-
-
-def test_check_validity_fail():
-    c_intermediate = np.asarray([[1, 0, 0, 0, 1],
-                                 [0, 1, 1, 0, 0],
-                                 [1, 1, 1, 0, 1],
-                                 [0, 0, 0, 1, 0],
-                                 [1, 0, 1, 0, 1]], dtype=np.int32)
-
-    assert not pairing.pairing._check_validity(c_intermediate)
+    assert (c_I == pairing.generate_indirect_connectivity(c_D)).all()
 
 
 def test_40_atoms():
     trj = md.load(get_fn('sevick1988.gro'))
     direct = pairing.generate_direct_correlation(trj, cutoff=0.8)
-    indirect = pairing.pairing._generate_indirect_connectivity(direct)
+    indirect = pairing.generate_indirect_connectivity(direct)
 
     assert indirect.dtype == np.int32
 
 
 def test_indirect_matrix_reduction():
     trj = md.load(get_fn('sevick1988.gro'))
     direct = pairing.generate_direct_correlation(trj, cutoff=0.8)
-    indirect = pairing.pairing._generate_indirect_connectivity(direct)
+    indirect = pairing.generate_indirect_connectivity(direct)
 
     c_R = np.asarray([[0, 1],
                       [0, 1],
@@ -83,20 +63,7 @@ def test_indirect_matrix_reduction():
 def test_cluster_analysis():
     trj = md.load(get_fn('sevick1988.gro'))
     direct = pairing.generate_direct_correlation(trj, cutoff=0.8)
-    indirect = pairing.pairing._generate_indirect_connectivity(direct)
+    indirect = pairing.generate_indirect_connectivity(direct)
     reduction = pairing.generate_clusters(indirect)
 
     assert pairing.analyze_clusters(reduction) == (2.5, 1.5)
-
-
-def test_new_indirect():
-    ref = np.asarray([[1, 1, 1, 0, 1],
-                      [1, 1, 1, 0, 1],
-                      [1, 1, 1, 0, 1],
-                      [0, 0, 0, 1, 0],
-                      [1, 1, 1, 0, 1]], dtype=np.int32)
-    trj = md.load(get_fn('sevick1988.gro'))
-    direct = pairing.generate_direct_correlation(trj, cutoff=0.8)
-    indirect = pairing.new_generate_indirect(direct)
-
-    assert (indirect == ref).all()