split repexp to strict and console.log

[pmario]

Hi,
This is a really cool library.

But the split regexp in trie.js

words = words.split(/[^a-zA-Z]+/);

IMO is a little bit to strict for eg: german language, where we have öäü ..
It would be nice to have the possibility to overwrite your defaults with a config object, that contains a regexp eg: /[^a-zA-ZöäüÖÄÜ]+/

And there is a console.log which you forgot to comment.

kind regards

[mckoss]

Thanks. I just noticed your comment. I assume these are Unicode symbols then. Strictly speaking, I can relax words further as long as they are not characters that I've reserved for punctuation in the PackedTrie string format (currently, ':', '!', and ';').

So, I could split on the traditional \s (space characters) plus my reserved characters. That would then allow contractions into the dictionary as well (like can't).

words = words.split(/[\s:!;]+/);

I'll put the console.log inside an if DEBUG statement...