Tokenizer: A tokenizer for Icelandic text

Tokenization is a necessary first step in many natural language processing
tasks, such as word counting, parsing, spell checking, corpus generation, and
statistical analysis of text.