Grouping but the lists can contain duplicate elements

Question

I have an ordered list

L= [330.56, 330.6,330.65,330.7, .....]

and I want to group it by a certain tolerance -+, but I can have duplicates, so I need to iterate over every element ( see if there are other elements within element-tol and element +tol) and then delete if there is a complete sublist

The code I’m using look like this but there are no duplicates and only check next elements

def mw_grouper(iterable):
group = []
for item in iterable:
    if not group or item - group[0] <= 0.05:
        group.append(item)
    else:
        yield group
        group = [item]
if group:
    yield group

What I get is more of, I need to also check the previous elements

R = [[330.56, 330.6], [330.65], [330.7]]

The output I want

R =  [[330.56, 330.6],[330.56, 330.6, 330.65], [330.6,330.65,330.7], [330.7, 330.65]]

and then delete the sublists

  F =  [[330.56, 330.6, 330.65], [330.6,330.65,330.7]]

Asked By: ZAHR

||

Source

Answer 1

Ok, this should work, though I didn’t test it intensively, so there might be unforeseen errors.

Instead of deleting subgroups after the fact, I prevented the yielding of a subgroup of the last or next group.
I decided to use only one "sliding" list (group), but it would probably have been simpler (though more memory intensive in case of huge lists) to implement a list for each element of the iterable.

I may have overthought this and overlooked a simpler implementation, but that’s the best I’m able to produce at this time 😉

def group_by_tolerance(my_iterable, tolerance):
    active_index = 0   # The index (in group) of the next element "whose" group has to be yielded
    group = []
    last_released_group = []
    for number in my_iterable:
        if not group:
            group = [number]
            continue
        elif number - group[active_index] <= tolerance:
            group.append(number)
            continue
        while group and number - group[active_index] > tolerance:
            # check that this group is not a subgroup of the next one
            if active_index >= len(group)-1 or group[active_index+1] - group[0] > tolerance:
                # check that this group is not a subgroup of the previous one
                if len(last_released_group) < len(group) or group != last_released_group[-len(group)]:
                    last_released_group = group.copy()
                    yield last_released_group
            active_index += 1
            if active_index >= len(group):
                active_index = 0
                group = []
                continue
            while group[active_index] - group[0] > tolerance:
                group.pop(0)
                active_index -= 1
        group.append(number)
    if group:
        yield group
        
L= [330.56, 330.6,330.65,330.7]
        
print(list(group_by_tolerance(L,0.01)))
# [[330.56], [330.6], [330.65], [330.7]]

print(list(group_by_tolerance(L,0.051)))
# [[330.56, 330.6, 330.65], [330.6, 330.65, 330.7]]

print(list(group_by_tolerance(L,0.1)))
# [[330.56, 330.6, 330.65, 330.7]]

Here’s the code tweaked to return (index, value) instead of value only; you can then treat the output to get the values only, or the indices only:

def group_by_tolerance(my_iterable, tol):
    active_index = 0
    group = []
    last_released_group = []
    for i,number in enumerate(my_iterable):
        if not group:
            group = [(i,number)]
            continue
        elif number - group[active_index][1] <= tol:
            group.append((i,number))
            continue
        while group and number - group[active_index][1] > tol:
            # check that this group is not a subgroup of the next one
            if active_index >= len(group)-1 or group[active_index+1][1] - group[0][1] > tol:
                # check that this group is not a subgroup of the previous one
                if len(last_released_group) < len(group) or group != last_released_group[-len(group)]:
                    last_released_group = group.copy()
                    yield last_released_group
            active_index += 1
            if active_index >= len(group):
                active_index = 0
                group = []
                continue
            while group[active_index][1] - group[0][1] > tol:
                group.pop(0)
                active_index -= 1
        group.append((i,number))
    if group:
        yield group

print(list(group_by_tolerance(L,0.051)))
# [[(0, 330.56), (1, 330.6), (2, 330.65)], [(1, 330.6), (2, 330.65), (3, 330.7)]]

Answered By: Swifty

Grouping but the lists can contain duplicate elements

Question:

Answers: