Huffman Trees

This repository contains a simple Python implementation of Huffman trees.

Implementation Notes

Constructs trees in an arbitrary base, inserting dummy symbols to be as efficient as possible
tests.py tests that decoding an encoded message works properly and that the expected word length is within one bit of entropy by assembling a string of random symbols.
symbols.yaml is a config file where you can put the symbols you want to encode (symbols should be single characters). The weights can be coefficients. They do not have to be probabilities; the code will normalize them into probabilities. You can also pick the word length for the test message there.