A simple POC of sentiment analisys using a small portuguese corpus

textblob_sentan_pt.py

"""
A simple POC of sentiment analisys using a small portuguese corpus
- The corpus can be downloaded from here http://www.linguateca.pt/Repositorio/ReLi/ReLi-Lex.rar
please visit the page before download http://www.linguateca.pt/Repositorio/ReLi/

"""
from textblob import TextBlob
from textblob.classifiers import NaiveBayesClassifier
import os
import re

base_path = 'ReLi-Lex'
train = []
files = [os.path.join(base_path, f) for f in os.listdir(base_path)]

for file in files:
    t = 'pos' if '_Positivos' in file else 'neg'
    with open(file, 'r') as content_file:
        content = content_file.read()
        all = re.findall('\[.*?\]',content)
        for w in all:
            train.append((w[1:-1], t))

cl = NaiveBayesClassifier(train)
blob = TextBlob("A cerveja é boa. Mas a ressaca é horrível.", classifier=cl)
for s in blob.sentences:
    print("[%s] %s" % (s.classify(), s))

#[pos] A cerveja é boa.
#[neg] Mas a ressaca é horrível.