Need some assistance with Algo C4.5, got a recursion when building a decision tree

Question

I need some help with C4.5 algo. I got one task to write that algo by myself and build a decision tree on a train data and then test it.

Using data from UCI Machine Learning Repository. I wrote some code, got recursion child in child and it built wrong tree.

Here is how i build tree

def build_tree(parent):
    for sublist in parent.split_list:
        child = Node()
        child.parent_atr = parent.atr
        child.parent_type = sublist[parent.atr][0]

        if entropy(sublist) == 0.0:
            child.class_type = sublist[CLASS][0]
            parent.children.append(child)
        else:
            atr, split_list = get_best_attribute(sublist.copy().drop([parent.atr], axis=1))

            if atr is None:
                child.class_type = round(sum(sublist[CLASS]/len(sublist)))
                parent.children.append(child)
            else:
                child.atr = atr
                child.split_list = split_list
                parent.children.append(child)
                return build_tree(child)

The Node class:

class Node:
    children = list()
    split_list = pd.DataFrame()

    def __int__(self, param_name="", parent_type="", parent_atr="", class_type=None):
        self.atr = param_name
        self.parent_type = parent_type
        self.parent_atr = parent_atr
        self.class_type = class_type

and get_best_attribute is

def get_best_attribute(data):
    ratios = dict()
    dict_list = dict()

    if len(data.columns) == 1:
        return None, data

    for column_name in data:
        if column_name == CLASS:
            continue
        uniq_values = data[column_name].unique()
        subarray_list = list()
        for value in uniq_values:
            subarray_list.append(data.loc[data[column_name] == value].copy())
        ratios[column_name] = gain_ratio(data, subarray_list)
        dict_list[column_name] = subarray_list
    ans = max(ratios.items(), key=operator.itemgetter(1))[0]
    return ans, dict_list[ans]

I checkd the formulas (entropy and e.t.c), they work correctly.

When debugging got this recurion.

What I'm doing wrong?