from nltk.corpus import movie_reviews
import analisador
def word_feats(words):
rdata = ""
for word in words:
rdata += word
return rdata
negids = movie_reviews.fileids('neg')
posids = movie_reviews.fileids('pos')
neg_movies = [(word_feats(movie_reviews.words(fileids=[f])), 'neg') for f in negids]
pos_movies = [(word_feats(movie_reviews.words(fileids=[f])), 'pos') for f in posids]
testfeats = neg_movies[750:1000] + pos_movies[750:1000]
acertos = 0
for a in testfeats:
if analisador.sentiment(a[0])['label'] == a[1]:
acertos += 1
print 'accuracy:', acertos/float(len(testfeats))
import socket
import analisador
HOST = 'localhost'
PORT = 7001
s = socket.socket(socket.AF_INET, socket.SOCK_STREAM)
s.bind((HOST, PORT))
s.listen(1)
print 'Analisador 1 inicializado na porta', PORT
while 1:
conn, addr = s.accept()
data = conn.recv(1024)
if not data: continue
result = analisador.sentiment(data)
conn.sendall(str(result))
conn.close()
import analisador
f = open('C:\Users\Paulo Oliveira\Desktop\\teste feats.txt', 'r')
count = 0
total = 0
for l in f:
total += 1
data = l.split('.:.')
try:
sentiment = analisador.sentiment(data[1])
if sentiment['label'] == data[0].strip():
count += 1
else:
print sentiment['prob'], data[0].strip(), data[1].strip()
except:
continue
print total, count, count/float(total) * 100
