pypos 0.1.2

A part of speech tagger based on Hidden Markov models

PyPOS - Python Part-of-Speech tagger

This is  a  project, which allows its  users to assign part of speech tags to words in a  sentence .
dt   vbz dt nn     , wdt   vbz    prp$ nns   to vb     nn   in nn     nns  to nns   in dt nn       .

PyPOS uses Hidden Markov Models and Viterbi decoding to determine the most likely sequence of POS tags for a given sequence of words.

Usage

Installation

Requires Python 3.6 or higher

pip3 install pypos

Training

from pypos import PartOfSpeechTagger, PartOfSpeechDataset
tagger = PartOfSpeechTagger()
ds = PartOfSpeechDataset.load('train.txt')
tagger.fit(ds).save('tagger.p')

Tagging

from pypos import PartOfSpeechTagger
tagger = PartOfSpeechTagger.load('tagger.p')

# Reproducing the results shown above:
sentence = 'This is a project, which allows its users to assign part of speech tags to words in a sentence.'
tokens = tagger.tokenize(sentence)
tags = tagger.tag(sentence, human_readable=False)