从 trie 中获取包含特定字母的单词

Question

I'd like to retrieve words from a trie that contain specific letters. For example: List all words that contain the letters [a, g]. If my trie has the words ["APPLE", "EGG", "CAR", "BLUE", "AGRICULTURE", "DONE"] it would return "AGRICULTURE".

This is a very simple trie implementation

def load_trie(words):
    root = {}
    for word in words:
        curr_node = root
        for letter in word:
            curr_node = curr_node.setdefault(letter, {})
        curr_node.setdefault('', True)
    return root

with open('sowpods') as word_list:
    words = [word.strip().upper() for word in word_list]
    
TRIE = load_trie(words)

If I can check for words containing specific letters, it would also be nice to look for words that don't contain specific letters.

Answer 1

@Mark offers a helpful remark about the dual problem:

For words the don't contain the letter you can prune the branches with those keys.

Now, how could we make a trie, or any tree, well adapted to the primal problem? Let's see. A standard answer for the Anagram Problem is to store sorted letter sets

for word in vocabulary:
    set_to_word[sorted(word)].append(word)

and then

set_to_word.get(sorted(target_word))

will reveal all corresponding anagrams.

To adapt this to a trie, we want frequent letters near the root. Here is one plausible frequency ordering of the alphabet:

ETAOINSRHLDCUMFPGWYBV KXJQZ

Rather than the sorted( ... ) permutation, permute words by etaoin, and take advantage of the fact that hits will often result in a hit near the root. In this scheme "tears" would map to "etasr".

For the dual problem, simply store the complement set of letters, obeying the same ordering. So "tears" maps to a 21-character string, or perhaps some prefix truncation would suffice.