def reg_bussvc(list_line, n_token, command):
# check if the token match the regular expression for bus service or not
list_ftr = []
for i in range(0, len(list_line), 3):
split_first = 0
split_second = 0
if i % 3 == 0:
split_first = list_line[i].lower().strip().split('\t')
j = i + 1
if j % 3 == 1:
split_second = list_line[j].strip().split('\t')
ftr = ''
for k in range(0, len(split_first)):
token = split_first[k].strip()
if pattern_token_bussvc(token, command) is True:
ftr += '1'
else:
range_k = range_text_index(k, len(split_first), n_token)
token_text = ''
for m in range(range_k[1], range_k[2] + 1):
token_text = token_text + ' ' + split_first[m]
token_text = token_text.strip()
if pattern_tokenText_bussvc(token, token_text, command):
ftr += '1'
else:
ftr += '0'
# print len(split_first), split_first
# print len(ftr), ftr
print i / 3
list_ftr.append(ftr)
return list_ftr
def extract_ftr_gt_road_busstop(list_line, command, n_token):
list_dict = load_dict(command)
list_ftr = []
cnt = 0
for i in range(0, len(list_line), 3):
split_first = 0
split_second = 0
if i % 3 == 0:
split_first = list_line[i].strip().split('\t')
j = i + 1
if j % 3 == 1:
split_second = list_line[j].strip().split('\t')
for k in range(0, len(split_second)):
label = 0
if command == 'road':
label = 2
elif command == 'busstop':
label = 3
if label == 0: # quit if we don't have the correct command
print 'Give the correct command'
quit()
if (int(split_second[k] == label)) or (split_first[k].strip().lower() in list_dict):
word = split_first[k].lower() + ' ' # take the word which in svc dictionary
range_k = range_text_index(k, len(split_second), n_token)
ftr_text = ''
for m in range(range_k[1], range_k[2] + 1):
ftr_text = ftr_text + ' ' + split_first[m]
ftr_text = ftr_text.strip()
if int(split_second[k]) == label:
list_ftr.append(str(cnt) + '\t' + 'TRUE' + '\t' + word + '\t' + ftr_text)
else:
list_ftr.append(str(cnt) + '\t' + 'FALSE' + '\t' + word + '\t' + ftr_text)
cnt += 1
for value in list_ftr:
print value
print 'Length of list features is: %i' % len(list_ftr)
def extract_ftr_gt_svc(list_line, load_svc, n_token):
list_ftr = []
cnt = 0
cnt_true_svc = 0
for i in range(0, len(list_line), 3):
split_first = 0
split_second = 0
if i % 3 == 0:
split_first = list_line[i].strip().split('\t')
j = i + 1
if j % 3 == 1:
split_second = list_line[j].strip().split('\t')
for k in range(0, len(split_second)):
if split_first[k] in load_svc: # mean bus svc
svc = split_first[k].lower() + ' ' # take the word which in svc dictionary
range_k = range_text_index(k, len(split_second), n_token)
ftr_text = ''
for m in range(range_k[1], range_k[2] + 1):
ftr_text = ftr_text + ' ' + split_first[m]
ftr_text = ftr_text.strip()
if int(split_second[k]) == 1:
list_ftr.append(str(cnt) + '\t' + 'TRUE' + '\t' + svc + '\t' + ftr_text)
elif int(split_second[k]) == 0:
list_ftr.append(str(cnt) + '\t' + 'FALSE' + '\t' + svc + '\t' + ftr_text)
if int(split_second[k]) == 1:
cnt_true_svc += 1
cnt += 1
for value in list_ftr:
print value
print 'Length of list features is: %i' % len(list_ftr)
print 'Length of bus service labeling TRUE is: %i' % cnt_true_svc