def readlines(self,idx=None):
"""Read from file and split data into examples and labels"""
fp = open(self.filename,'r')
(dataname,issparse,alist,data) = arff.arffread(fp)
fp.close()
self.dataname = dataname
#if (alist[0][0]!='label'):
# sys.stderr.write('First column of ARFF file needs to be the label\n')
# sys.exit(-1)
if idx is None:
idx = range(len(data))
labels = [data[ix][0] for ix in idx]
labels = array(labels)
if self.extype == 'vec':
examples = [data[ix][1:] for ix in idx]
examples = array(examples).T
print '%d features, %d examples' % examples.shape
elif self.extype == 'seq':
examples = [data[ix][1] for ix in idx]
print 'sequence length = %d, %d examples' % (len(examples[0]),len(examples))
elif self.extype == 'mseq':
examples = [data[ix][1:] for ix in idx]
printstr = 'sequence lengths = '
for seq in examples[0]:
printstr += '%d, ' % len(seq)
printstr += '%d examples' % len(examples)
print printstr
return (examples, labels)
def arffread_sequence(filename):
"""Read an ARFF file containing a sequence dataset"""
import arff
f = open(filename,'r')
(dataname,issparse,alist,data) = arff.arffread(f)
f.close()
if (alist[0][0]!='label'):
sys.stderr.write('First column of ARFF file needs to be the label\n')
sys.exit(-1)
all_labels = [ex[0] for ex in data]
all_labels = array(all_labels)
all_examples = [ex[1].upper() for ex in data]
print 'sequence length = %d, %d examples' % (len(all_examples[0]),len(all_examples))
print '%d labels' % len(all_labels)
return all_examples, all_labels
def arffread_real(filename):
"""Read an ARFF file containing a vectorial dataset"""
import arff
f = open(filename,'r')
(dataname,issparse,alist,data) = arff.arffread(f)
f.close()
if (alist[0][0]!='label'):
sys.stderr.write('First column of ARFF file needs to be the label\n')
sys.exit(-1)
all_labels = [ex[0] for ex in data]
all_labels = array(all_labels)
all_examples = [ex[1:] for ex in data]
all_examples = transpose(array(all_examples))
print '%d features, %d examples' % all_examples.shape
print '%d labels' % len(all_labels)
return all_examples, all_labels
def arffread_sequence(filename):
"""Read an ARFF file containing a sequence dataset"""
import arff
f = open(filename, 'r')
(dataname, issparse, alist, data) = arff.arffread(f)
f.close()
if (alist[0][0] != 'label'):
sys.stderr.write('First column of ARFF file needs to be the label\n')
sys.exit(-1)
all_labels = [ex[0] for ex in data]
all_labels = array(all_labels)
all_examples = [ex[1].upper() for ex in data]
print 'sequence length = %d, %d examples' % (len(
all_examples[0]), len(all_examples))
print '%d labels' % len(all_labels)
return all_examples, all_labels
def arffread_real(filename):
"""Read an ARFF file containing a vectorial dataset"""
import arff
f = open(filename, 'r')
(dataname, issparse, alist, data) = arff.arffread(f)
f.close()
if (alist[0][0] != 'label'):
sys.stderr.write('First column of ARFF file needs to be the label\n')
sys.exit(-1)
all_labels = [ex[0] for ex in data]
all_labels = array(all_labels)
all_examples = [ex[1:] for ex in data]
all_examples = transpose(array(all_examples))
print '%d features, %d examples' % all_examples.shape
print '%d labels' % len(all_labels)
return all_examples, all_labels
import collections
import arff
import re
import string
name, is_sparse, attributes, rows = arff.arffread(open('winners_losers.arff', 'rb'))
#writer = csv.writer(open('winners_losers_augmented.csv', 'wb'))
def character_frequencies(input_string, allowed_character_set):
frequencies = {}
for character in allowed_character_set:
frequencies[character] = 0
for character in string.lower(input_string):
if(character in allowed_character_set):
frequencies[character] += 1
return frequencies
english_characters = [chr(i) for i in range(ord('a'), ord('z')+1)]
output_rows = []
name_sizes = []
for row in rows:
full_name = row[0]
classification = row[1]
def iris():
name, is_sparse, attributes, data = arff.arffread(open('iris.arff', 'rb'))
return ([[float(value) for value in row[0:4]] for row in data],
[attributes[0][0], attributes[1][0], attributes[2][0], attributes[3][0]])