快排实现仿order by多字段排序

class OrderBy(object):

    def __init__(self, sequence, *condition, **extra_condition):
        """
        排序初始化条件
        condition为优先排序条件,序列内元素必须为字典类型
        extra_condition为额外的条件因素,当condition不存在时,额外条件才会生效
        :param sequence:
        :param condition:
        :param extra_condition:
        """
        self.array = sequence
        self.condition = list(condition)
        self.extra_condition = extra_condition
        self.has_condition = bool(self.condition)

    def partition(self, left, right, key=None, desc=False, func=None):
        """
        将区间数据进行排序
        分区比较,选择一个基准,根据排序规则,正序:比基准大的放右边,比基准小的放左边
        :param left: 序列区间的开始位置
        :param right: 序列区间的结束位置
        :param key: 对于字典对象,如果指定key,则以key对象比较否则当None值处理
        :param desc: 比较规则为正序或倒序
        :param func: 对值进行处理特殊处理的函数
        :return:
        """
        pivot = self.array[left]
        if isinstance(pivot, dict):
            if key is not None:
                pivot = pivot.get(key)
        if callable(func):
            pivot = func(pivot)
        low = left
        while low < right:
            while low < right:
                _left = self.array[low]
                _right = self.array[right]
                if isinstance(_left, dict):
                    _left = _left.get(key)
                if isinstance(_right, dict):
                    _right = _right.get(key)
                if callable(func):
                    _left = func(_left)
                    _right = func(_right)
                if desc:
                    # 倒序,右边值与基准值都不为空,且右边值小于基准值,左移
                    if _right and pivot and _right < pivot:
                        right -= 1
                    # 倒序,右边值为空,左移
                    elif not _right:
                        right -= 1
                    # 倒序,左边值与基准值都不为空,且左边值大于等于基准值,右移
                    elif _left and pivot and _left >= pivot:
                        low += 1
                    # 倒序,基准值为空,左边值不为空,右移
                    elif _left and not pivot:
                        low += 1
                    else:
                        break
                else:
                    # 正序,基准为空,右边值不为空,左移
                    if _right and not pivot:
                        right -= 1
                    # 正序,右边值与基准都不为空,且右边值大于基准值,左移
                    elif _right and pivot and _right > pivot:
                        right -= 1
                    # 正序,左边值与基准都不为空,且左边值小于等于基准值,右移
                    elif _left and pivot and _left <= pivot:
                        low += 1
                    # 正序,左边值为空，右移
                    elif not _left:
                        low += 1
                    else:
                        break
            if low < right:
                temp = self.array[low]
                self.array[low] = self.array[right]
                self.array[right] = temp
        self.array[left], self.array[low] = self.array[low], self.array[left]
        return low

    def quick_sort(self, left=0, right=None, key=None, desc=False, func=None):
        """
        快速排序算法
        :param left: 区间起始位置
        :param right: 区间结束位置
        :param key: 字典元素使用指定键名值排序
        :param desc: 是否倒序
        :param func: 对于排序值,支持使用函数处理
        :return:
        """
        if right is None:
            right = len(self.array) - 1
        if left < right:
            pivot_position = self.partition(left, right, key, desc, func)
            self.quick_sort(left, pivot_position - 1, key, desc, func)
            self.quick_sort(pivot_position + 1, right, key, desc, func)

    def sort(self, **condition):
        if self.has_condition:
            if not self.condition:
                return
            _condition = self.condition.pop(0)
            if isinstance(_condition, dict):
                condition['key'] = _condition.get('key')
                condition['desc'] = _condition.get('desc', False)
        else:
            condition.update(**self.extra_condition)
        left = condition.get('left')
        right = condition.get('right')
        if not left:
            left = 0
            condition['left'] = left
        if not right:
            right = len(self.array) - 1
            condition['right'] = right
        self.quick_sort(**condition)
        self.sub_sort(left, right, condition.get('key'))

    def sub_sort(self, left, right, key=None, next_index=0):
        """
        标记当前位置begin,及下一个位置end
        当begin位置的元素与end位置元素不等时,当begin!=end - 1时,则进行区间排序
        :param left: 区间起始位置
        :param right: 区间结束位置
        :param key: 当前排序键名
        :param next_index: 下一个条件位置
        :return:
        """
        condition_size = len(self.condition)
        if not condition_size > next_index:
            return
        begin = left
        for end in range(left, right + 1):
            _left = self.array[begin]
            _right = self.array[end]
            if isinstance(_left, dict):
                _left = _left.get(key)
            if isinstance(_right, dict):
                _right = _right.get(key)
            # 当上一个值与当前值不相等,则进入二次排序
            # 当上一个值与当前值相等,且当前位置等于边界位置,且还有下一个排序条件,则进入二次排序
            if _left != _right or (
                    end == right and condition_size >= next_index + 1):
                condition = self.condition[next_index]
                _key = condition.get('key')
                desc = condition.get('desc')
                func = condition.get('func')
                if end == right:
                    _end = end
                else:
                    _end = end - 1
                self.quick_sort(begin, _end, _key, desc, func)
                self.sub_sort(begin, end, _key, next_index + 1)
                begin = end

if __name__ == '__main__':
    a = [dict(age=18, money=200, name='z1'),
         dict(age=16, money=200, name='z2'),
         dict(age=16, money=200, name='z3'),
         dict(age=16, money=100, name='z4'),
         dict(age=16, money=200, name='z5')]
    order_by = OrderBy(a, dict(key='age', desc=False),
                       dict(key='money', desc=True),
                       dict(key='name', desc=True))
    print(a)
    order_by.sort()
    print(a)