[NeetCode 150] Word Ladder

Word Ladder

You are given two words, beginWord and endWord, and also a list of words wordList. All of the given words are of the same length, consisting of lowercase English letters, and are all distinct.

Your goal is to transform beginWord into endWord by following the rules:

You may transform beginWord to any word within wordList, provided that at exactly one position the words have a different character, and the rest of the positions have the same characters.

You may repeat the previous step with the new word that you obtain, and you may do this as many times as needed.

Return the minimum number of words within the transformation sequence needed to obtain the endWord, or 0 if no such sequence exists.

Example 1:

复制代码
Input: beginWord = "cat", endWord = "sag", wordList = ["bat","bag","sag","dag","dot"]

Output: 4

Explanation: The transformation sequence is "cat" -> "bat" -> "bag" -> "sag".

Example 2:

复制代码
Input: beginWord = "cat", endWord = "sag", wordList = ["bat","bag","sat","dag","dot"]

Output: 0

Explanation: There is no possible transformation sequence from "cat" to "sag" since the word "sag" is not in the wordList.

Constraints:

复制代码
1 <= beginWord.length <= 10
1 <= wordList.length <= 100

Solution

The "distance" of every transformation is 1 so it is OK to apply BFS for searching the shortest path. Because it will take O ( wordList.length 2 × beginWord.length 2 ) O(\text{wordList.length}^2\times\text{beginWord.length}^2) O(wordList.length2×beginWord.length2) to build up the graph inevitably, more advanced shortest path algorithm is not necessary.

At first, we put begin word into BFS queue and set the initial distance of 1. Then we keep getting the top word from queue and check whether it can reach out other unvisited words. If so, we add these new words into queue and set their corresponding distance to current distance+1. When we reach the end word during this process, we can return early. If we cannot reach end word after the queue is empty, it means the end word is not reachable.

Code

python 复制代码
class Solution:
    def ladderLength(self, beginWord: str, endWord: str, wordList: List[str]) -> int:
        if beginWord == endWord:
            return 1
        vis_flag = {word: False for word in wordList}
        # dis = {word: 10000000 for word in wordList}
        vis_flag[beginWord] = True
        vis_flag[endWord] = False
        from queue import Queue
        bfs_queue = Queue()
        bfs_queue.put((beginWord, 1))
        def check(a, b):
            if a == b:
                return False
            cnt = 0
            for i in range(len(a)):
                if a[i] != b[i]:
                    cnt += 1
            return cnt == 1
        while not bfs_queue.empty():
            cur = bfs_queue.get()
            for word in wordList:
                if not vis_flag[word] and check(cur[0], word):
                    if word == endWord:
                        return cur[1] + 1
                    vis_flag[word] = True
                    bfs_queue.put((word, cur[1]+1))
        return 0
        
相关推荐
qq_5139704434 分钟前
力扣 hot100 Day56
算法·leetcode
mortimer2 小时前
安装NVIDIA Parakeet时,我遇到的两个Pip“小插曲”
python·github
@昵称不存在3 小时前
Flask input 和datalist结合
后端·python·flask
赵英英俊3 小时前
Python day25
python
东林牧之3 小时前
Django+celery异步:拿来即用,可移植性高
后端·python·django
爱喝矿泉水的猛男4 小时前
非定长滑动窗口(持续更新)
算法·leetcode·职场和发展
何双新4 小时前
基于Tornado的WebSocket实时聊天系统:从零到一构建与解析
python·websocket·tornado
YuTaoShao4 小时前
【LeetCode 热题 100】131. 分割回文串——回溯
java·算法·leetcode·深度优先
AntBlack4 小时前
从小不学好 ,影刀 + ddddocr 实现图片验证码认证自动化
后端·python·计算机视觉
凪卄12135 小时前
图像预处理 二
人工智能·python·深度学习·计算机视觉·pycharm