DESCRIPTION.txt

Feet is a tool that extract terms from a text according to dictionaries.
A specific dictionary can be the categorization of terms that belong to the same entity family. So we will also mention a dictionary as an entity.

Everything that Feet needs resides in a distributed in-memory database system [Redis](https://www.redis.io).
So It's fast and scalable.

The underlying natural language processing is provided by Python NLTK package.
Japanese has been implemented thanks to the Mecab POS tagger.

On-going work for next version:

* project packaging
* much more test cases
* implement various technics for disambiguition of entities by semantics
analysis including the use of thesauri and machine learning.
* compliant to JSON API interface
* definition of relationships between entities according to grammatical patterns including but not only the definition of intents, facts etc.
* additional languages support: Chinese and Korean

As a bonus, there is a function to extract dates from text in Japanese. Next
version may feature a multi-lingual dates extraction function.

Last but not the least this is WIP so rely on it at your own risk. We should
get a reliable version within a few weeks, along with new exciting features.