provide jupyter notebook for corpus generation

[proppy]

We should provide a notebook to document the corpus generation and allow developer to easily create alternative corpus.

This should cover:

• word extraction and filtering
• indexing
• text embeddeding model generation
• upload to database.

[proppy]

Interesting alternative corpus are available on https://www.ninjal.ac.jp/english/database/.

In particular http://chaki-data.ninjal.ac.jp/momotaro/momotaro-2015-11-10/ looks very interesting :)