hahmed1 / trie-apps

Excursions with my favorite data structure

Trie-apps

Excursions with my favorite data structure

A basic implementation of autocomplete using a Trie. Provide a large body of text, and the program builds a weighted Trie, where the weights are based on the frequency of occurence of each word. The returned list is sorted by weight in descending order.

usage: auto-complete.py [-h] -text TEXT -prefix PREFIX [-count COUNT]

arguments:

-h, --help show this help message and exit

-text TEXT, -t TEXT the path to the source text with which we will build the word-frequency trie

-prefix PREFIX, -p PREFIX the prefix (i.e. key) to the trie

-count COUNT, -c COUNT the number of words to return in the suggestion list

The demo below creates a weighted Trie out of all the words in the the French classic The Count of Monte Cristo, and then suggests 10 words starting with the prefix "the". Most of the execution time is spent in building the Trie (the book has about half a million words, about 40K of which are unique). The actual lookup/autocomplete is fast once the Trie is built.