Support expanding of compound words into separate tokens

[tituomin]

I have patched our version of the plugin (based on v0.3.0) and added a configuration parameter expandCompounds to optionally support expanding of compound words (yhdyssanat) into separate tokens.

https://github.com/City-of-Helsinki/elasticsearch-analysis-voikko/commit/9a6bd8165e4de3a7d4f8bbe0993ebaec17197f94

I would like to get this feature into master and upstream, if you find it desirable. I can port it to master myself, but currently we are using 0.3.0.

We have found that extracting the parts of compound words is highly desirable in the index analysis stage, for several reasons:

• users often misspell compound words and write them separately
• often parts of compound words (for example "terveys" in "terveysasema") are meaningful and relevant even separated from the compound word

[komu]

Sounds great, if you'll open a PR I'll look forward into merging it.

[tituomin]

@komu here is my attempt at a PR.