aboutsummaryrefslogtreecommitdiff
path: root/gn3/computations/slink.py
blob: 3156b3150ac651eeadb88f065db2710eaccbb0d7 (plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
class LengthError(BaseException):
    pass

class MirrorError(BaseException):
    pass

def raise_valueerror_if_data_is_not_lists_or_tuples(lists):
    """Check that `lists` is a list of lists: If not, raise an exception."""
    def is_list_or_tuple(item):
        return type(item) == type([]) or type(item) == type(tuple)

    if (not is_list_or_tuple(lists)) or (not all(map(is_list_or_tuple, lists))):
        raise ValueError("Expected list or tuple")

def raise_valueerror_if_lists_empty(lists):
    """Check that the list and its direct children are not empty."""
    def empty(lst):
        return len(lst) == 0
    if (empty(lists)) or not all(map(lambda x: not empty(x), lists)):
        raise ValueError("List/Tuple should NOT be empty!")

def raise_lengtherror_if_child_lists_are_not_same_as_parent(lists):
    def len_is_same_as_parent(lst):
        return len(lst) == len(lists)
    if not all(map(len_is_same_as_parent, lists)):
        raise LengthError("All children lists should be same length as the parent.")

def raise_valueerror_if_child_list_distance_from_itself_is_not_zero(lists):
    def get_child_distance(child):
        idx = lists.index(child)
        return lists[idx][idx]
    def distance_is_zero(dist):
        return dist == 0
    children_distances = map(get_child_distance, lists)
    if not all(map(distance_is_zero, children_distances)):
        raise ValueError("Distance of each child list/tuple from itself should be zero!")

def raise_mirrorerror_of_distances_one_way_are_not_same_other_way(lists):
    """Check that the distance from A to B, is the same as the distance from B to A.
If the two distances are different, throw an exception."""
    for i in range(len(lists)):
        for j in range(len(lists)):
            if lists[i][j] != lists[j][i]:
                raise MirrorError(
                    ("Distance from one child({}) to the other ({}) "
                     "should be the same in both directions.").format(
                         lists[i][j], lists[j][i]))

def raise_valueerror_on_negative_distances(lists):
    """Check that distances between 'somethings' are all positive, otherwise,
raise an exception."""
    def zero_or_positive(val):
        return val >= 0;
    # flatten lists
    flattened = [distance for child in lists for distance in child]
    if not all(map(zero_or_positive, flattened)):
        raise ValueError("Distances should be positive.")

def nearest(lists, i, j):
    """Computes some form of distance.
This is 'copied' over from genenetwork1, from https://github.com/genenetwork/genenetwork1/blob/master/web/webqtl/heatmap/slink.py#L42-L64.

This description should be updated once the form/type of 'distance' identified."""

    #### Guard Functions: Should we do this a different way? ####
    raise_valueerror_if_data_is_not_lists_or_tuples(lists)
    raise_valueerror_if_lists_empty(lists)
    raise_lengtherror_if_child_lists_are_not_same_as_parent(lists)
    raise_valueerror_if_child_list_distance_from_itself_is_not_zero(lists)
    raise_mirrorerror_of_distances_one_way_are_not_same_other_way(lists)
    raise_valueerror_on_negative_distances(lists)
    #### END: Guard Functions ####
    if type(i) == int and type(j) == int: # From member i to member j
        return lists[i][j]
    elif type(i) == int and (type(j) in [list, tuple]):
        return min(map(lambda j_new: nearest(lists, i, j_new), j))
    elif type(j) == int and (type(i) in [list, tuple]):
        return min(map(lambda i_new: nearest(lists, i_new, j), i))