将任意长度的位置[4,1,2]的列表转换为嵌套列表的索引

时间:2011-07-02 16:58:16

标签: python recursion eval nested-lists generated-code

假设此列表

nestedList = ["a", "b", [1, 2, 3], "c",[4, 5, 6, [100, 200, 300]], "d"]

我有一个函数返回任意深度嵌套列表的位置列表。 的实施例

[2, 1] -> "2"
[5] -> "d"
[4, 3, 2] -> "300"

正如您所看到的那样,一开始并不清楚有多少级别的嵌套。

其他问题 对于列表修改,我想使用[:]或[4:]或[0:1]表示法。

对于人类而言,它很容易做到:只需添加尽可能多的索引位置。

nestedList[2][1]
nestedList[5]
nestedList[4][3][2]
nestedList[4][1:] = NewItem + nestedList[4][1:] #insert item
nestedList[2][1] = [] #remove item

然而,由于我必须将字符串附加在一起并稍后评估它,因此这种方法不会导致任何问题。显而易见的废话:)

处理具有未知数量的索引位置的嵌套列表的最佳方法是什么,并且仍然具有像普通列表一样处理它的功能(读取,修改,插入,删除)

我希望有答案。

P.S。列表必须保持嵌套。展平不是一种选择。

2 个答案:

答案 0 :(得分:7)

第一部分很容易。

>>> reduce(lambda x, y: x[y], [4, 3, 2], nestedList)
300

第二部分需要更多努力,但仍然可行。提示:

>>> a = [1, 2, 3]
>>> a[slice(1, None)] = [4, 5]
>>> a
[1, 4, 5]

答案 1 :(得分:1)

我终于有时间摆弄这个。我有点被带走了。这很长,但无论如何我都会粘贴它。我添加了set_iteminsertdeletefindfind_left方法,以及一些私有方法,以允许破坏游标的低级操作抽象。我还添加了一个move_cursor方法,该方法为超出范围或指向非顶层对象的索引元组抛出IndexError

基本上,它(应该)保证只要你只使用公共函数,光标总是指向顶级对象,插入和删除都发生在顶层。从这里开始,您应该能够安全地实施__getitem____setitem____delitem__等,甚至__getslice____setslice__

然而,有一些皱纹。光标始终指向顶层对象的限制使得迭代嵌套列表变得非常容易,就好像它是一个平面列表一样。但它也意味着光标不能指向较低级别的对象,因此仅使用insert就不会发生某些类型的插入。例如,假设您有三个列表:

>>> l1 = [1, 2, 3, 4]
>>> l2 = [5, 6, 7, 8]
>>> l3 = [l1, l2]
>>> l3
[[1, 2, 3, 4], [5, 6, 7, 8]]

现在将此嵌套结构放入NLI,移至5,然后尝试插入。

>>> nli = NestedListIter(l3)
>>> nli.find(5)
>>> nli.insert(9)
>>> nli.nested_list
[[1, 2, 3, 4], [9, 5, 6, 7, 8]]

正如您所看到的,您可以在l2中插入内容,但无法在l3中插入内容。事实上,要立即这样做,你必须使用一个私有函数,它以一种不愉快的方式打破游标抽象:

>>> nli._insert_at(nli.stack[:-1], 10)
>>> nli.nested_list
[[1, 2, 3, 4], 10, [9, 5, 6, 7, 8]]
>>> nli.get_item()
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "nestedlistiterator.py", line 130, in get_item
    return self._get_item_at(self.stack)
  File "nestedlistiterator.py", line 39, in _get_item_at
    item = item[i]
TypeError: 'int' object is unsubscriptable

确实有一些方法可以实施安全的公共insert_between_branches方法,但它们涉及的复杂程度比我现在关心的要多。

当尝试在4之后插入值时,会出现另一个问题。如您所见,您可以在l2之前将值插入5,但如果您将光标移至4insert,您很快就会意识到无法在4l1之后插入内容。

>>> nli.go_to_head()
>>> nli.find(4)
>>> nli.insert(11)
>>> nli.nested_list
[[1, 2, 3, 11, 4], 10, [9, 5, 6, 7, 8]]

从平面访问的角度来看,4之后插入和5之前插入是相同的,但从嵌套列表的角度来看,它们是不同的。由于insert实际上是left_insert,因此可以使用right_insert方法对此问题进行部分纠正(反过来,无法在l1的开头插入)。

通过允许光标指向较低级别的对象,可以更一般地处理这些问题,但这会使平面访问变得更加复杂。简而言之,任何纠正这些问题的尝试都将导致更大的复杂性,无论是在接口的平面还是嵌套侧。

(这就是为什么我仍然喜欢简单的enumerate_nested方法!在所有节点(而不仅仅是顶级节点)上具有值的正确树结构也可能更简单,更好。但这很有趣尽管如此。)

import collections

class NestedListIter(object):
    '''A mutable container that enables flat traversal of a nested tree of 
    lists. nested_list should contain only a list-like mutable sequence. 
    To preserve a clear demarcation between 'leaves' and 'branches', empty 
    sequences are not allowed as toplevel objects.'''
    def __init__(self, nested_list):
        if not nested_list:
            raise ValueError, 'nested_list must be a non-empty sequence'
        self.nested_list = nested_list # at some point, vet this to make sure
        self.go_to_head()              # it contains no empty sequences

    def _is_sequence(self, item=None):
        '''Private method to test whether an item is a non-string sequence.
        If item is None, test current item.'''
        if item is None:
            item = self._get_item_at(self.stack)
        return isinstance(item, collections.Sequence) and not isinstance(item, basestring)

    def _is_in_range(self, index_tuple=None):
        '''Private method to test whether an index is in range. 
        If index is None, test current index.'''
        if index_tuple is None:
            index_tuple = self.stack
        if any(x < 0 for x in index_tuple):
            return False
        try:
            self._get_item_at(index_tuple)
        except IndexError:
            return False
        else:
            return True

    def _get_item_at(self, index_tuple):
        '''Private method to get item at an arbitrary index, with no bounds checking.'''
        item = self.nested_list
        for i in index_tuple:
            item = item[i]
        return item

    def _set_item_at(self, index_tuple, value):
        '''Private method to set item at an arbitrary index, with no bounds checking.
        Throws a ValueError if value is an empty non-string sequence.'''
        if self._is_sequence(value) and not value:
            raise ValueError, "Cannot set an empty list!"
        containing_list = self._get_item_at(index_tuple[:-1])
        containing_list[index_tuple[-1]] = value

    def _insert_at(self, index_tuple, value):
        '''Private method to insert item at an arbitrary index, with no bounds checking.
        Throws a ValueError if value is an empty non-string sequence.'''
        if self._is_sequence(value) and not value:
            raise ValueError, "Cannot insert an empty list!"
        containing_list = self._get_item_at(index_tuple[:-1])
        containing_list.insert(index_tuple[-1], value)

    def _delete_at(self, index_tuple):
        '''Private method to delete item at an arbitrary index, with no bounds checking.
        Recursively deletes a resulting branch of empty lists.'''
        containing_list = self._get_item_at(index_tuple[:-1])
        del containing_list[index_tuple[-1]]
        if not self._get_item_at(index_tuple[:-1]):
            self._delete_at(index_tuple[:-1])

    def _increment_stack(self):
        '''Private method that tires to increment the top value of the stack.
        Returns True on success, False on failure (empty stack).'''
        try:
            self.stack[-1] += 1
        except IndexError:
            return False
        else: 
            return True

    def _decrement_stack(self):
        '''Private method that tries to decrement the top value of the stack.
        Returns True on success, False on failure (empty stack).'''
        try:
            self.stack[-1] -= 1
        except IndexError:
            return False
        else:
            return True

    def go_to_head(self):
        '''Move the cursor to the head of the nested list.'''
        self.stack = []
        while self._is_sequence():
            self.stack.append(0)

    def go_to_tail(self):
        self.stack = []
        '''Move the cursor to the tail of the nested list.'''
        while self._is_sequence():
            self.stack.append(len(self.get_item()) - 1)

    def right(self):
        '''Move cursor one step right in the nested list.'''
        while self._increment_stack() and not self._is_in_range():
            self.stack.pop()
        if not self.stack:
            self.go_to_tail()
            return False
        while self._is_sequence():
            self.stack.append(0)
        return True

    def left(self):
        '''Move cursor one step left in the nested list.'''
        while self._decrement_stack() and not self._is_in_range():
            self.stack.pop()
        if not self.stack:
            self.go_to_head()
            return False
        while self._is_sequence():
            self.stack.append(len(self.get_item()) - 1)
        return True

    def move_cursor(self, index_tuple):
        '''Move cursor to the location indicated by index_tuple.
        Raises an error if index_tuple is out of range or doesn't correspond
        to a toplevel object.'''
        item = self._get_item_at(index_tuple)
        if self._is_sequence(item):
            raise IndexError, 'index_tuple must point to a toplevel object'

    def get_item(self):
        '''Get the item at the cursor location.'''
        return self._get_item_at(self.stack)

    def set_item(self, value):
        '''Set the item a the cursor locaiton.'''
        return self._set_item_at(self.stack, value)

    def insert(self, value):
        '''Insert an item at the cursor location. If value is a sequence, 
        cursor moves to the first toplevel object in value after insertion. 
        Otherwise, cursor does not move.'''
        temp_stack = self.stack[:]
        self.left()
        self._insert_at(temp_stack, value)
        self.right()

    def delete(self):
        '''Deete an item at the cursor location. Cursor does not move.'''
        temp_stack = self.stack[:]
        self.left()
        self._delete_at(temp_stack)
        self.right()

    def iterate(self):
        '''Iterate over the values in nested_list in sequence'''
        self.go_to_head()
        yield self.get_item()
        while self.right():
            yield self.get_item()

    def iterate_left(self):
        '''Iterate over the values in nested_list in reverse.'''
        self.go_to_tail()
        yield self.get_item()
        while self.left():
            yield self.get_item()

    def find(self, value):
        '''Search for value in nested_list; move cursor to first location of value.'''
        for i in self.iterate():
            if i == value:
                break

    def find_left(self, value):
        '''Search for value backwards in nested_list; move cursor to last location of value.'''
        for i in self.iterate_left():
            if i == value:
                break

def _NLI_Test():
    l = [1, 2, 3, ['a', 'b', 'c'], 4, ['d', 'e', [100, 200, 300]], 5, ['a', 'b', 'c'], 6]
    nli = NestedListIter(l)
    print nli.nested_list
    for i in nli.iterate():
        print i,
    print
    for i in nli.iterate_left():
        print i,
    print

    nli.go_to_head()
    for i in range(5):
        nli.right()
    nli.insert('cow')
    nli.insert(['c', ['o', 'w']])
    print nli.nested_list
    nli.find('cow')
    print nli.get_item()
    nli.delete()
    print nli.nested_list
    nli.find('c')
    nli.delete()
    print nli.nested_list
    nli.find_left('w')
    nli.delete()
    nli.find('o')
    nli.delete()
    print nli.nested_list
    print nli.nested_list == l
    nli.find(100)
    nli.set_item(100.1)
    print nli.nested_list

if __name__ == '__main__':
    _NLI_Test()