smarkov

Simple, lightweight and easy to read implementation of Markov chains and HMMs.

This is a toy project, don't expect any exciting speeds or robustness.

Happy hacking!

Installing

pip3 install git+git://github.com/greenify/smarkov.git

Hacking

git clone https://github.com/greenify/smarkov
cd smarkov
python3 setup.py develop

Train with a corpus

from smarkov import Markov
chain = Markov(["AGACAGACGAC"])

Attributes

corpus: given corpus (a corpus_entry needs to be a tuple or array)
order: maximal order to look back for a given state (default 1) tokenize: function how to split an element of the corpus (e.g sentences into words)

Generate text from a chain

print("".join(chain.generate_text()))

Generate_text() generates exactly one element from the Markov chain. In other words: It goes in the Markov chain the universal start state to universal end state.

More Examples

See examples

Coming

Documentation how to use it with HMM.

License

MIT