def save_result(self):
'''Save classify result.'''
now = str(datetime.now()).replace(':', '-')
file_opt = options = {}
options['defaultextension'] = ''
options['initialfile'] = 'ff_' + now[:19] + '.pickle'
options['filetypes'] = [('pickle file', '.pickle')]
options['initialdir'] = getcwd() + '\\data'
options['title'] = 'Save Result'
options['parent'] = self.parent
# create dir if not exist
if not path.exists(options['initialdir']):
makedirs(options['initialdir'])
file_name = tkFD.asksaveasfilename(**file_opt)
self.text_output.delete('1.0', END)
if file_name:
if self.FFF._factor_finder.save(file_name):
self.text_output.insert(END, 'Saving result successful...' + '\n\n')
util.debug('saved')
else:
self.text_output.insert(END, 'Saving result failed...' + '\n\n')
util.debug('failed to save result')
pass
def load_result(self):
'''Load classify result.'''
file_opt = options = {}
options['defaultextension'] = ''
options['filetypes'] = [('pickle file', '.pickle')]
options['initialdir'] = getcwd() + '\\data'
options['title'] = 'Load Result'
options['parent'] = self.parent
# create dir if not exist
if not path.exists(options['initialdir']):
makedirs(options['initialdir'])
file_name = tkFD.askopenfilename(**file_opt)
if file_name:
print file_name
flag, self.FFF._factor_finder = self.FFF._factor_finder.load(
file_name)
self.text_output.delete('1.0', END)
# print to the output text
if flag:
util.debug('Loaded')
self.text_output.insert(END, 'Result loaded...' + '\n\n')
else:
self.text_output.insert(END, 'Failed to load result...')
util.debug('Failed to load result')
def load_result(self):
'''Load classify result.'''
file_opt = options = {}
options['defaultextension'] = ''
options['filetypes'] = [('pickle file', '.pickle')]
options['initialdir'] = getcwd() + '\\data'
options['title'] = 'Load Result'
options['parent'] = self.parent
# create dir if not exist
if not path.exists(options['initialdir']):
makedirs(options['initialdir'])
file_name = tkFD.askopenfilename(**file_opt)
if file_name:
print file_name
flag, self.FFF._factor_finder = self.FFF._factor_finder.load(file_name)
self.text_output.delete('1.0', END)
# print to the output text
if flag:
util.debug('Loaded')
self.text_output.insert(END, 'Result loaded...' + '\n\n')
else:
self.text_output.insert(END, 'Failed to load result...')
util.debug('Failed to load result')
def save_classifier(self):
'''Save classifier.'''
now = str(datetime.now()).replace(':', '-')
file_opt = options = {}
options['defaultextension'] = ''
options['initialfile'] = 'c_' + now[:19] + '.pickle'
options['filetypes'] = [('pickle file', '.pickle')]
options['initialdir'] = getcwd() + '\\data'
options['title'] = 'Save classifier'
options['parent'] = self.parent
# create dir if not exist
if not path.exists(options['initialdir']):
makedirs(options['initialdir'])
file_name = tkFD.asksaveasfilename(**file_opt)
self.text_output.delete('1.0', END)
if file_name:
if self.FFF._classifier.save(file_name):
self.text_output.insert(
END, 'Saving classifier successful...' + '\n\n')
util.debug('saved')
else:
self.text_output.insert(END,
'Saving classifier failed...' + '\n\n')
util.debug('failed to save classifier')
def extract_topic(self):
'''Get when the breakpoints happened. Show the graph if it's selected.'''
self.text_output.delete('1.0', END)
# check if the data is not empty
if self.FFF._factor_finder.list_tweet == None:
self.text_output.insert(END, 'Failed to extract topic, list tweet is empty' + '\n')
else:
# retrieve parameter
try:
# date
start_time = datetime.strptime(self.ent_start_date.get() + ' ' + self.ent_start_time.get(), '%d-%m-%Y %H:%M:%S')
end_time = datetime.strptime(self.ent_end_date.get() + ' ' + self.ent_end_time.get(), '%d-%m-%Y %H:%M:%S')
duration_hour = int(self.ent_duration.get())
decay_factor = float(self.ent_decay_factor.get())
show_graph = bool(self.show_graph_val.get())
except Exception, e:
util.debug('retrieve parameter error')
self.text_output.insert(END, 'Parameter error' + '\n')
# to avoid long variable name, just call it div_sent short of divide_sentiment, pffttt...
div_sent = self.FFF._factor_finder.divide_sentiment_time(start_time, end_time, duration_hour, decay_factor)
break_point = self.FFF._factor_finder.get_break_points()
# print retval_divide_sentiment and retval_break_point to output text
self.text_output.insert(END, 'Topic Extraction : ' + '\n\n')
# topic extraction
# topics = self.FFF._factor_finder.get_break_point_topics()
topics_pos = self.FFF._factor_finder.get_all_topics(5, decay_factor, 1)
topics_neg = self.FFF._factor_finder.get_all_topics(5, decay_factor, -1)
self.text_output.insert(END, 'Topic Extraction parameters : ' + '\n')
datetime.now().strftime('%d-%m-%Y')
self.text_output.insert(END, 'Start time : ' + str(start_time) + '\n')
self.text_output.insert(END, 'End time : ' + str(end_time) + '\n')
self.text_output.insert(END, 'Duration : ' + str(duration_hour) + '\n')
self.text_output.insert(END, 'Discounted Cumulative Factor : ' + str(decay_factor) + '\n\n')
self.text_output.insert(END, 'No\tStart time\t\t Num\t\tSentiment\tKumulatif\n')
i = 0
for idx in div_sent:
#self.text_output.insert(END, str(i + 1) + '\t' + str(div_sent[idx]['start_time']) + '\t' + str(div_sent[idx]['end_time'])[:13] + '\t' + str(len(div_sent[idx]['list_tweet'])) + '\t\t' + str(div_sent[idx]['sentiment'])[:6] + '\t' + str(div_sent[idx]['cum_sentiment'])[:6] +'\n')
self.text_output.insert(END, str(i + 1) + '\t' + str(div_sent[idx]['start_time']) + '\t\t' + str(len(div_sent[idx]['list_tweet'])) + '\t\t' + str(div_sent[idx]['sentiment'])[:6] + '\t' + str(div_sent[idx]['cum_sentiment'])[:6] +'\n')
self.text_output.insert(END, '\tPositif Topics :\t'+ ', '.join(topics_pos[idx]) + '\n')
self.text_output.insert(END, '\tNegatif Topics :\t'+ ', '.join(topics_neg[idx]) + '\n')
i += 1
# show graph or not
if show_graph:
self.text_output.insert(END, '\nShowing graph' + '\n')
self.text_output.insert(END, 'Close pop up to continue' + '\n')
self.FFF._factor_finder.plot_graph()
def test(model, fusion_datasets, configs, load_weight_path=False, save_path=None):
model.eval()
if load_weight_path:
assert configs['TEST']['weight_path'] != 'None', 'Test Need To Resume Chekpoint'
weight_path = configs['TEST']['weight_path']
checkpoint = torch.load(weight_path)
model.load_state_dict(checkpoint['model'].state_dict())
is_use_gpu = torch.cuda.is_available()
test_dataloader = DataLoader(fusion_datasets, batch_size=configs['TEST']['batch_size'], shuffle=False)
test_num_iter = len(test_dataloader)
dtransforms = transforms.Compose([transforms.ToPILImage()])
with tqdm(total=test_num_iter) as test_bar:
for iter, data in enumerate(test_dataloader):
if is_use_gpu:
model = model.cuda()
data = {sensor: data[sensor].cuda() for sensor in data}
fusion_image = model(data)
input_imgs, fusion_imgs = debug(configs['MODEL'], configs['TEST_DATASET'], data, fusion_image)
input_imgs = [input_imgs[sensor] for sensor in configs['MODEL']['input_sensors']]
imgs = input_imgs + [fusion_imgs]
imgs = torch.cat(imgs, dim=3)
for batch in range(imgs.shape[0]):
if save_path is None:
save_path = configs['TEST']['save_path']
name = os.path.join(save_path, str(len(os.listdir(save_path))))
img = imgs[batch].cpu()
img = dtransforms(img)
img.save(f'{name}.jpg')
test_bar.update(1)
def load_classifier(self):
'''Load classifier.'''
file_opt = options = {}
options['defaultextension'] = ''
options['filetypes'] = [('pickle file', '.pickle')]
options['initialdir'] = getcwd() + '\\data'
options['title'] = 'Load classifier'
options['parent'] = self.parent
# create dir if not exist
if not path.exists(options['initialdir']):
makedirs(options['initialdir'])
file_name = tkFD.askopenfilename(**file_opt)
if file_name:
print file_name
flag, self.FFF._classifier = self.FFF._classifier.load(file_name)
self.text_output.delete('1.0', END)
# print to the output text
if flag:
util.debug('Loaded')
self.text_output.insert(END, 'Classifier loaded...' + '\n\n')
self.text_output.insert(END, 'Preprocess Parameter : ' + '\n')
self.text_output.insert(END, 'Fold Case : ' + str(bool(self.FFF._classifier.dict_param['fold_case'])) + '\n')
self.text_output.insert(END, 'Remove RT : ' + str(bool(self.FFF._classifier.dict_param['remove_RT'])) +'\n')
self.text_output.insert(END, 'Remove Hashtag : ' + str(bool(self.FFF._classifier.dict_param['remove_hashtag'])) + '\n')
self.text_output.insert(END, 'Remove Username : '******'remove_username'])) + '\n')
self.text_output.insert(END, 'Convert Number : ' + str(bool(self.FFF._classifier.dict_param['convert_number'])) + '\n')
self.text_output.insert(END, 'Clean Number : ' + str(bool(self.FFF._classifier.dict_param['clean_number'])) + '\n')
self.text_output.insert(END, 'Convert Emoticon : ' + str(bool(self.FFF._classifier.dict_param['convert_emoticon'])) + '\n')
self.text_output.insert(END, 'Remove Punctuation : ' + str(bool(self.FFF._classifier.dict_param['remove_punctuation_string'])) + '\n')
self.text_output.insert(END, 'Convert Word : ' + str(bool(self.FFF._classifier.dict_param['convert_word'])) + '\n')
self.text_output.insert(END, 'Remove Stop Word : ' + str(bool(self.FFF._classifier.dict_param['remove_stop_words'])) + '\n')
self.text_output.insert(END, 'Convert Negation : ' + str(bool(self.FFF._classifier.dict_param['convert_negation'])) + '\n\n')
self.text_output.insert(END, 'Minimal Occur : ' + str(self.FFF._classifier.min_occur) + '\n')
else:
self.text_output.insert(END, 'Failed to load classifier...')
util.debug('Failed to load')
def train(model, train_datasets, test_datasets, configs):
if not os.path.exists(
os.path.join(configs['PROJECT']['save_path'],
configs['PROJECT']['name'])):
os.mkdir(
os.path.join(configs['PROJECT']['save_path'],
configs['PROJECT']['name']))
model.train()
train_writer = SummaryWriter(log_dir=os.path.join(
configs['PROJECT']['save_path'], configs['PROJECT']['name']))
if configs['TRAIN']['resume'] == 'None':
start_epoch = 1
else:
start_epoch = torch.load(configs['TRAIN']['resume'])['epoch'] + 1
is_use_gpu = torch.cuda.is_available()
optimizer = eval('torch.optim.' + configs['TRAIN']['opt'])(
model.parameters(), configs['TRAIN']['lr'])
scheduler = torch.optim.lr_scheduler.MultiStepLR(
optimizer,
milestones=configs['TRAIN']['milestones'],
gamma=configs['TRAIN']['gamma'])
train_dataloader = DataLoader(train_datasets,
batch_size=configs['TRAIN']['batch_size'],
shuffle=True)
train_num_iter = len(train_dataloader)
loss_func = [eval(l)() for l in configs['TRAIN']['loss_func']]
all_iter = 0
for epoch in range(start_epoch, configs['TRAIN']['max_epoch'] + 1):
loss_epoch = 0
with tqdm(total=train_num_iter) as train_bar:
for iter, data in enumerate(train_dataloader):
if is_use_gpu:
model = model.cuda()
data = {sensor: data[sensor].cuda() for sensor in data}
fusion_image = model(data)
loss = [
l(data, fusion_image) *
configs['TRAIN']['loss_weights'][loss_func.index(l)]
for l in loss_func
]
loss_batch = sum(loss)
loss_epoch += loss_batch.item()
optimizer.zero_grad()
loss_batch.backward()
optimizer.step()
train_writer.add_scalar('loss',
loss_batch,
global_step=all_iter)
train_bar.set_description(
'Epoch: {}/{}. TRAIN. Iter: {}/{}. All loss: {:.5f}'.
format(epoch, configs['TRAIN']['max_epoch'], iter + 1,
train_num_iter, loss_epoch / train_num_iter))
if configs['TRAIN'][
'debug_interval'] is not None and all_iter % configs[
'TRAIN']['debug_interval'] == 0:
input_imgs, fusion_imgs = debug(configs['MODEL'],
configs['TRAIN_DATASET'],
data, fusion_image)
input_imgs = [
input_imgs[sensor]
for sensor in configs['MODEL']['input_sensors']
]
imgs = input_imgs + [fusion_imgs]
train_writer.add_image('debug',
torch.cat(imgs, dim=2),
all_iter,
dataformats='NCHW')
all_iter += 1
train_bar.update(1)
scheduler.step()
train_writer.add_scalar(
'lr',
optimizer.state_dict()['param_groups'][0]['lr'],
global_step=epoch)
if configs['TRAIN'][
'val_interval'] is not None and all_iter % configs[
'TRAIN']['val_interval'] == 0:
torch.save({
'model': model,
'epoch': epoch
},
os.path.join(configs['PROJECT']['save_path'],
configs['PROJECT']['name'],
f'model_{epoch}.pth'))
def classify(self):
'''Classify some tweets according to the current classifier.'''
self.text_output.delete('1.0', END)
# checking whether the classifier has been trained or not
if not self.FFF._classifier.trained:
self.text_output.insert(END, 'Gagal mengklasifikasikan, belum ditraining' + '\n')
else:
try:
num_tweet = int(self.ent_num_tweet_classify.get())
random_seed = int(self.ent_random_seed_classify.get())
keyword = self.ent_keyword_classify.get()
# date
start_time = datetime.strptime(self.ent_start_date.get() + ' ' + self.ent_start_time.get(), '%d-%m-%Y %H:%M:%S')
end_time = datetime.strptime(self.ent_end_date.get() + ' ' + self.ent_end_time.get(), '%d-%m-%Y %H:%M:%S')
except Exception, e:
raise e
debug(str(e))
test_data = tm.get_test_data(keyword, start_time, end_time)
if random_seed != 0:
random.seed(random_seed)
random.shuffle(test_data)
list_tweet = self.FFF.classify_tweets(test_data, keyword, num_tweet)
# print to the output text
self.text_output.delete('1.0', END)
self.text_output.insert(END, 'Hasil Klasifikasi' + '\n\n')
self.text_output.insert(END, 'Preprocess Parameter : ' + '\n\n')
self.text_output.insert(END, 'Fold Case : ' + str(bool(self.FFF._classifier.dict_param['fold_case'])) + '\n')
self.text_output.insert(END, 'Remove RT : ' + str(bool(self.FFF._classifier.dict_param['remove_RT'])) +'\n')
self.text_output.insert(END, 'Remove Hashtag : ' + str(bool(self.FFF._classifier.dict_param['remove_hashtag'])) + '\n')
self.text_output.insert(END, 'Remove Username : '******'remove_username'])) + '\n')
self.text_output.insert(END, 'Convert Number : ' + str(bool(self.FFF._classifier.dict_param['convert_number'])) + '\n')
self.text_output.insert(END, 'Clean Number : ' + str(bool(self.FFF._classifier.dict_param['clean_number'])) + '\n')
self.text_output.insert(END, 'Convert Emoticon : ' + str(bool(self.FFF._classifier.dict_param['convert_emoticon'])) + '\n')
self.text_output.insert(END, 'Remove Punctuation : ' + str(bool(self.FFF._classifier.dict_param['remove_punctuation_string'])) + '\n')
self.text_output.insert(END, 'Convert Word : ' + str(bool(self.FFF._classifier.dict_param['convert_word'])) + '\n')
self.text_output.insert(END, 'Remove Stop Word : ' + str(bool(self.FFF._classifier.dict_param['remove_stop_words'])) + '\n')
self.text_output.insert(END, 'Convert Negation : ' + str(bool(self.FFF._classifier.dict_param['convert_negation'])) + '\n\n')
self.text_output.insert(END, 'Minimal Kemunculan : ' + str(self.FFF._classifier.min_occur) + '\n\n')
self.text_output.insert(END, 'Keyword : ' + keyword + '\n')
self.text_output.insert(END, 'Banyak Tweet : ' + str(self.FFF._classifier.num_tweet_classified) + '\n')
self.text_output.insert(END, 'Random Seed : ' + str(random_seed) + '\n')
# show result
self.show_tweet = self.show_tweet_val.get()
list_sentiment = []
if self.show_tweet == 1:
for t in list_tweet:
self.text_output.insert(END, 'Tweet : ' + str(t.get_normal_text()) + '\n')
self.text_output.insert(END, 'Sentiment : ' + str(t.sentiment) + '\n\n')
list_sentiment.append(t.sentiment)
else:
for t in list_tweet:
list_sentiment.append(t.sentiment)
self.text_output.insert(END, '\n')
self.text_output.insert(END, 'Result : \n')
self.text_output.insert(END, 'Positif :'+ str(list_sentiment.count(1)) + ' \n')
self.text_output.insert(END, 'Netral :'+ str(list_sentiment.count(0)) + ' \n')
self.text_output.insert(END, 'Negatif :'+ str(list_sentiment.count(-1)) + ' \n')
# this lontong function
from xlwt import Workbook
from tempfile import TemporaryFile
book = Workbook()
try:
activeSheet = book.add_sheet(str('fuuu'))
i = 1
activeSheet.write(i, 0, 'No')
activeSheet.write(i, 1, 'Created')
activeSheet.write(i, 2, 'Text')
activeSheet.write(i, 3, 'Sentiment')
i += 1
for tweet in list_tweet:
activeSheet.write(i, 0, str(i - 1))
activeSheet.write(i, 1, tweet.time.__str__())
activeSheet.write(i, 2, str(tweet.get_normal_text()))
activeSheet.write(i, 3, tweet.sentiment)
i += 1
pret = str(start_time).replace(':', '-')
book.save('lontong' + pret+ '.xls')
book.save(TemporaryFile())
except Exception, e:
util.debug(str(e))
def load_classifier(self):
'''Load classifier.'''
file_opt = options = {}
options['defaultextension'] = ''
options['filetypes'] = [('pickle file', '.pickle')]
options['initialdir'] = getcwd() + '\\data'
options['title'] = 'Load classifier'
options['parent'] = self.parent
# create dir if not exist
if not path.exists(options['initialdir']):
makedirs(options['initialdir'])
file_name = tkFD.askopenfilename(**file_opt)
if file_name:
print file_name
flag, self.FFF._classifier = self.FFF._classifier.load(file_name)
self.text_output.delete('1.0', END)
# print to the output text
if flag:
util.debug('Loaded')
self.text_output.insert(END, 'Classifier loaded...' + '\n\n')
self.text_output.insert(END, 'Preprocess Parameter : ' + '\n')
self.text_output.insert(
END, 'Fold Case : ' +
str(bool(self.FFF._classifier.dict_param['fold_case'])) +
'\n')
self.text_output.insert(
END, 'Remove RT : ' +
str(bool(self.FFF._classifier.dict_param['remove_RT'])) +
'\n')
self.text_output.insert(
END, 'Remove Hashtag : ' +
str(bool(
self.FFF._classifier.dict_param['remove_hashtag'])) +
'\n')
self.text_output.insert(
END, 'Remove Username : '******'remove_username']
)) + '\n')
self.text_output.insert(
END, 'Convert Number : ' +
str(bool(
self.FFF._classifier.dict_param['convert_number'])) +
'\n')
self.text_output.insert(
END, 'Clean Number : ' +
str(bool(self.FFF._classifier.dict_param['clean_number']))
+ '\n')
self.text_output.insert(
END, 'Convert Emoticon : ' + str(
bool(self.FFF._classifier.
dict_param['convert_emoticon'])) + '\n')
self.text_output.insert(
END, 'Remove Punctuation : ' + str(
bool(self.FFF._classifier.
dict_param['remove_punctuation_string'])) + '\n')
self.text_output.insert(
END, 'Convert Word : ' +
str(bool(self.FFF._classifier.dict_param['convert_word']))
+ '\n')
self.text_output.insert(
END, 'Remove Stop Word : ' + str(
bool(self.FFF._classifier.
dict_param['remove_stop_words'])) + '\n')
self.text_output.insert(
END, 'Convert Negation : ' + str(
bool(self.FFF._classifier.
dict_param['convert_negation'])) + '\n\n')
self.text_output.insert(
END, 'Minimal Occur : ' +
str(self.FFF._classifier.min_occur) + '\n')
else:
self.text_output.insert(END, 'Failed to load classifier...')
util.debug('Failed to load')
def extract_topic(self):
'''Get when the breakpoints happened. Show the graph if it's selected.'''
self.text_output.delete('1.0', END)
# check if the data is not empty
if self.FFF._factor_finder.list_tweet == None:
self.text_output.insert(
END, 'Failed to extract topic, list tweet is empty' + '\n')
else:
# retrieve parameter
try:
# date
start_time = datetime.strptime(
self.ent_start_date.get() + ' ' +
self.ent_start_time.get(), '%d-%m-%Y %H:%M:%S')
end_time = datetime.strptime(
self.ent_end_date.get() + ' ' + self.ent_end_time.get(),
'%d-%m-%Y %H:%M:%S')
duration_hour = int(self.ent_duration.get())
decay_factor = float(self.ent_decay_factor.get())
show_graph = bool(self.show_graph_val.get())
except Exception, e:
util.debug('retrieve parameter error')
self.text_output.insert(END, 'Parameter error' + '\n')
# to avoid long variable name, just call it div_sent short of divide_sentiment, pffttt...
div_sent = self.FFF._factor_finder.divide_sentiment_time(
start_time, end_time, duration_hour, decay_factor)
break_point = self.FFF._factor_finder.get_break_points()
# print retval_divide_sentiment and retval_break_point to output text
self.text_output.insert(END, 'Topic Extraction : ' + '\n\n')
# topic extraction
# topics = self.FFF._factor_finder.get_break_point_topics()
topics_pos = self.FFF._factor_finder.get_all_topics(
5, decay_factor, 1)
topics_neg = self.FFF._factor_finder.get_all_topics(
5, decay_factor, -1)
self.text_output.insert(END,
'Topic Extraction parameters : ' + '\n')
datetime.now().strftime('%d-%m-%Y')
self.text_output.insert(END,
'Start time : ' + str(start_time) + '\n')
self.text_output.insert(END, 'End time : ' + str(end_time) + '\n')
self.text_output.insert(END,
'Duration : ' + str(duration_hour) + '\n')
self.text_output.insert(
END,
'Discounted Cumulative Factor : ' + str(decay_factor) + '\n\n')
self.text_output.insert(
END, 'No\tStart time\t\t Num\t\tSentiment\tKumulatif\n')
i = 0
for idx in div_sent:
#self.text_output.insert(END, str(i + 1) + '\t' + str(div_sent[idx]['start_time']) + '\t' + str(div_sent[idx]['end_time'])[:13] + '\t' + str(len(div_sent[idx]['list_tweet'])) + '\t\t' + str(div_sent[idx]['sentiment'])[:6] + '\t' + str(div_sent[idx]['cum_sentiment'])[:6] +'\n')
self.text_output.insert(
END,
str(i + 1) + '\t' + str(div_sent[idx]['start_time']) +
'\t\t' + str(len(div_sent[idx]['list_tweet'])) + '\t\t' +
str(div_sent[idx]['sentiment'])[:6] + '\t' +
str(div_sent[idx]['cum_sentiment'])[:6] + '\n')
self.text_output.insert(
END, '\tPositif Topics :\t' + ', '.join(topics_pos[idx]) +
'\n')
self.text_output.insert(
END, '\tNegatif Topics :\t' + ', '.join(topics_neg[idx]) +
'\n')
i += 1
# show graph or not
if show_graph:
self.text_output.insert(END, '\nShowing graph' + '\n')
self.text_output.insert(END, 'Close pop up to continue' + '\n')
self.FFF._factor_finder.plot_graph()
def classify(self):
'''Classify some tweets according to the current classifier.'''
self.text_output.delete('1.0', END)
# checking whether the classifier has been trained or not
if not self.FFF._classifier.trained:
self.text_output.insert(
END, 'Gagal mengklasifikasikan, belum ditraining' + '\n')
else:
try:
num_tweet = int(self.ent_num_tweet_classify.get())
random_seed = int(self.ent_random_seed_classify.get())
keyword = self.ent_keyword_classify.get()
# date
start_time = datetime.strptime(
self.ent_start_date.get() + ' ' +
self.ent_start_time.get(), '%d-%m-%Y %H:%M:%S')
end_time = datetime.strptime(
self.ent_end_date.get() + ' ' + self.ent_end_time.get(),
'%d-%m-%Y %H:%M:%S')
except Exception, e:
raise e
debug(str(e))
test_data = tm.get_test_data(keyword, start_time, end_time)
if random_seed != 0:
random.seed(random_seed)
random.shuffle(test_data)
list_tweet = self.FFF.classify_tweets(test_data, keyword,
num_tweet)
# print to the output text
self.text_output.delete('1.0', END)
self.text_output.insert(END, 'Hasil Klasifikasi' + '\n\n')
self.text_output.insert(END, 'Preprocess Parameter : ' + '\n\n')
self.text_output.insert(
END, 'Fold Case : ' +
str(bool(self.FFF._classifier.dict_param['fold_case'])) + '\n')
self.text_output.insert(
END, 'Remove RT : ' +
str(bool(self.FFF._classifier.dict_param['remove_RT'])) + '\n')
self.text_output.insert(
END, 'Remove Hashtag : ' +
str(bool(self.FFF._classifier.dict_param['remove_hashtag'])) +
'\n')
self.text_output.insert(
END, 'Remove Username : '******'remove_username'])) +
'\n')
self.text_output.insert(
END, 'Convert Number : ' +
str(bool(self.FFF._classifier.dict_param['convert_number'])) +
'\n')
self.text_output.insert(
END, 'Clean Number : ' +
str(bool(self.FFF._classifier.dict_param['clean_number'])) +
'\n')
self.text_output.insert(
END, 'Convert Emoticon : ' +
str(bool(self.FFF._classifier.dict_param['convert_emoticon']))
+ '\n')
self.text_output.insert(
END, 'Remove Punctuation : ' + str(
bool(self.FFF._classifier.
dict_param['remove_punctuation_string'])) + '\n')
self.text_output.insert(
END, 'Convert Word : ' +
str(bool(self.FFF._classifier.dict_param['convert_word'])) +
'\n')
self.text_output.insert(
END, 'Remove Stop Word : ' +
str(bool(self.FFF._classifier.dict_param['remove_stop_words']))
+ '\n')
self.text_output.insert(
END, 'Convert Negation : ' +
str(bool(self.FFF._classifier.dict_param['convert_negation']))
+ '\n\n')
self.text_output.insert(
END, 'Minimal Kemunculan : ' +
str(self.FFF._classifier.min_occur) + '\n\n')
self.text_output.insert(END, 'Keyword : ' + keyword + '\n')
self.text_output.insert(
END, 'Banyak Tweet : ' +
str(self.FFF._classifier.num_tweet_classified) + '\n')
self.text_output.insert(END,
'Random Seed : ' + str(random_seed) + '\n')
# show result
self.show_tweet = self.show_tweet_val.get()
list_sentiment = []
if self.show_tweet == 1:
for t in list_tweet:
self.text_output.insert(
END, 'Tweet : ' + str(t.get_normal_text()) + '\n')
self.text_output.insert(
END, 'Sentiment : ' + str(t.sentiment) + '\n\n')
list_sentiment.append(t.sentiment)
else:
for t in list_tweet:
list_sentiment.append(t.sentiment)
self.text_output.insert(END, '\n')
self.text_output.insert(END, 'Result : \n')
self.text_output.insert(
END, 'Positif :' + str(list_sentiment.count(1)) + ' \n')
self.text_output.insert(
END, 'Netral :' + str(list_sentiment.count(0)) + ' \n')
self.text_output.insert(
END, 'Negatif :' + str(list_sentiment.count(-1)) + ' \n')
# this lontong function
from xlwt import Workbook
from tempfile import TemporaryFile
book = Workbook()
try:
activeSheet = book.add_sheet(str('fuuu'))
i = 1
activeSheet.write(i, 0, 'No')
activeSheet.write(i, 1, 'Created')
activeSheet.write(i, 2, 'Text')
activeSheet.write(i, 3, 'Sentiment')
i += 1
for tweet in list_tweet:
activeSheet.write(i, 0, str(i - 1))
activeSheet.write(i, 1, tweet.time.__str__())
activeSheet.write(i, 2, str(tweet.get_normal_text()))
activeSheet.write(i, 3, tweet.sentiment)
i += 1
pret = str(start_time).replace(':', '-')
book.save('lontong' + pret + '.xls')
book.save(TemporaryFile())
except Exception, e:
util.debug(str(e))
