Stylomepy

Stylomepy is a Python library for measuring the style of a text. It is available in Spanish and Engish languages. It can be used in the analysis of the statistics of a text, the readability index of it, the vocabulary richness, formality and coherence.

Installation

Stylomepy can be installed using pip and Python 3.0 is needed:

pip3 install stylomepy

Another way is to clone this repository and then install it:

git clone https://github.com/alvarodepablo/stylomepy
cd stylomepy
python3 setup.py install

Use of the library

When we have the text or texts we want to analyze, it is necessary to create a StyleMetrics object that will include the style features of that text:

import stylomepy
style = stylomepy.StyleMetrics(text, lang = 'en', lowercase = True, punct = True, contract = True, stopWords = False, blank = False, window = 500, limit = 0.72, segment=100, coherence= False, extractor = None)

text is the text to analyze
lang is the language of that text. Only English ('en') and Spanish ('es') are available
lowercase if True all words become lower
punct if True punctuation signs are removed
contract if True English contractions are expanded
stopWords if True stopwords are removed
blank if True some functions take blank spaces as characters
window is the window size of the MATTR algorithm (usually 500 or 100)
limit is the limit of the MTLD algorithm (usually 0.72)
segment is the size of the MSTTR algorithm segment (usually 100)
coherence if True the coherence of the text will be computed (extractor needed)
extractor is a word2vec extractor instance

TextStatistics

To analyze the statistics of the text:

style.getTextStatistics(index='index')

The available indexes are sentencesCount, charPerSent, charPerWord, wordsPerSent, diffWords, conjunctions, adjectives, verbs, nouns, adverbs, prepositions, determiners, pronouns, interjections, shortWords.

Readability Index

To analyze the readability difficulty of a text:

style.getReadabilityIndex(index='index')

The available spanish indexes are INFLESZ and Mu. The available english indexes are ARI, FleschRI and FogCount.

Vocabulary Richness

To analyze the lexical diversity of a text:

style.getVocabularyRichness(index='index')

The available indexes are TTR, MSTTR, MATTR, MTLD and HDD.

Formality

To analyze the formality score of a text:

style.getFormalityIndex()

To analyze the adjective score of a text:

style.getAdjScore()

Coherence

To analyze the coherence of a text, when the StyleMetrics object is created, coherence attribute needs to be True and a word2vec extractor instance has to be created.

from gsitk.features.word2vec import Word2VecFeatures
w2v_extractor = Word2VecFeatures(w2v_model_path="path", w2v_format='"format"', convolution=[1,0,0])

style = stylomepy.StyleMetrics(text, coherence = True, extractor = w2v_extractor)
style.getCoherenceIndex()

Another way of use

If you want to use the library analyzing a few statistics, you do not need to create a StyleMetrics object. You can use some of the developed functions separately. The next examples show this:

import stylomepy

adverbs = stylomepy.TextStatistics.wordClass(text, lang = 'en', lowercase = True, contract = False, punct = True, stopWords = False)[4]
shortWords = stylomepy.TextStatistics.shortWords(text, lang ='es', contract = True, stopWords = True)
lexDiversity = stylomepy.VocabularyRichness.MATTR(text, window = 100, lang = 'es', lowercase = True, punct = True, contract = True, stopWords = False)

Possible issues

When the library analyzes a Spanish text, an error could appear saying something like NLTK was unable to find the java file! Use software specific configuration paramaters or set the JAVAHOME environment variable.. If it happens, run the next script and then use the library.

import os
import nltk

java_path = "jdk_path"
os.environ['JAVAHOME'] = java_path

nltk.internals.config_java(java_path)

Where jdk_path is the Java JDK path in your system.