def post(self):
res = {}
try:
# get all sim words
base_word = request.json['base_word']
print('recv base word:', base_word)
sim_list = []
for item in word_model.most_similar(base_word):
sim_list.append(item[0])
res['sim'] = sim_list
# get all important words
class_name = request.json['class_name']
print('recv classname:', class_name)
important_list = []
for classname in class_name:
remap_name = class_name_remap[classname]
word_df = word_importance_model.ix[remap_name]
important_list += list(flatten(word_df[[2]].head(20).values.tolist()))
res['important'] = important_list
# get all industry words
industry_list = []
for classname in class_name:
value_name = label2value[classname]
industry_df = industry_word_model.loc[
industry_word_model.classname.str.contains(value_name)]
industry_list += list(flatten(industry_df[[1]].head(20).values.tolist()))
res['industry'] = industry_list
except Exception as e:
print(e)
print("res words:", res)
return res
def predict(self, sent, pipeline=False):
self.encode(sent)
if self.full:
constraints = [] if self.args.no_lin_constraint else self.get_tree_constraints(
sent)
res = self.decode(sent[self.pred_input_key], constraints)
sent['linearized_tokens'] = res['seq']
else:
for token in traverse_bottomup(sent.root):
domain = (
[token] +
token['pdeps']) if self.args.pred_tree else token['domain']
if len(domain) > 1:
constraints = [] if self.args.no_lin_constraint else self.get_subtree_constraints(
token)
res = self.decode(domain, constraints)
token['linearized_domain'] = res['seq']
# add predicted sequential information
f_vec = self.f_lstm.initial_state().transduce(
[t.vecs[self.vec_key] for t in res['seq']])[-1]
b_vec = self.b_lstm.initial_state().transduce(
[t.vecs[self.vec_key] for t in res['seq'][::-1]])[-1]
token.vecs[self.vec_key] += (f_vec + b_vec)
else:
token['linearized_domain'] = [token]
sent['linearized_tokens'] = flatten(token, 'linearized_domain')
def post(self):
class_name = request.json['classname']
print('recv class_name:', class_name)
class_name = class_name_remap[class_name]
word_df = word_importance_model.ix[class_name]
word_list = list(flatten(word_df[[2]].head(20).values.tolist()))
return word_list
def post(self):
class_name = request.json['classname']
print('recv class_name:', class_name)
class_name = class_name_remap[class_name]
word_df = word_importance_model.ix[class_name]
word_list = list(flatten(word_df[[2]].head(20).values.tolist()))
return word_list
def post(self):
class_name = request.json['classname']
print('recv class_name:', class_name)
class_name = label2value[class_name]
#industy_word_path = "../data/words/word_{name}".format(name=class_name)
industry_df = industry_word_model.loc[
industry_word_model.classname.str.contains(class_name)]
word_list = list(flatten(industry_df[[1]].head(20).values.tolist()))
return word_list
def post(self):
class_name = request.json['classname']
print('recv class_name:', class_name)
class_name = label2value[class_name]
#industy_word_path = "../data/words/word_{name}".format(name=class_name)
industry_df = industry_word_model.loc[
industry_word_model.classname.str.contains(class_name)]
word_list = list(flatten(industry_df[[1]].head(20).values.tolist()))
return word_list
def post(self):
res = {}
try:
# get all sim words
base_word = request.json['base_word']
print('recv base word:', base_word)
sim_list = []
for item in word_model.most_similar(base_word):
sim_list.append(item[0])
res['sim'] = sim_list
# get all important words
class_name = request.json['class_name']
print('recv classname:', class_name)
important_list = []
for classname in class_name:
remap_name = class_name_remap[classname]
word_df = word_importance_model.ix[remap_name]
important_list += list(
flatten(word_df[[2]].head(20).values.tolist()))
res['important'] = important_list
# get all industry words
industry_list = []
for classname in class_name:
value_name = label2value[classname]
industry_df = industry_word_model.loc[
industry_word_model.classname.str.contains(value_name)]
industry_list += list(
flatten(industry_df[[1]].head(20).values.tolist()))
res['industry'] = industry_list
except Exception as e:
print(e)
print("res words:", res)
return res
def train_one_step(self, sent):
total = correct = loss = 0
t0 = time()
self.encode(sent)
if self.full:
constraints = [] if self.args.no_lin_constraint else self.get_tree_constraints(
sent)
res = self.decode(sent[self.train_input_key], constraints, True)
loss = res['loss']
total += 1
sent['linearized_tokens'] = res['seq']
correct += int(
sent['linearized_tokens'] == sent['gold_linearized_tokens'])
else:
for token in traverse_bottomup(sent.root):
domain = (
[token] +
token['pdeps']) if self.args.pred_tree else token['domain']
if len(domain) > 1:
constraints = [] if self.args.no_lin_constraint else self.get_subtree_constraints(
token)
res = self.decode(domain, constraints, True)
token['linearized_domain'] = res['seq']
loss += res['loss']
total += 1
correct += int(token['linearized_domain'] ==
token['gold_linearized_domain'])
# add predicted sequential information
f_vec = self.f_lstm.initial_state().transduce(
[t.vecs[self.vec_key] for t in res['seq']])[-1]
b_vec = self.b_lstm.initial_state().transduce(
[t.vecs[self.vec_key] for t in res['seq'][::-1]])[-1]
token.vecs[self.vec_key] += (f_vec + b_vec)
else:
token['linearized_domain'] = [token]
sent['linearized_tokens'] = flatten(token, 'linearized_domain')
loss_value = loss.value() if loss else 0
return {
'time': time() - t0,
'loss': loss_value,
'loss_expr': loss,
'total': total,
'correct': correct
}
def post(self):
res = {}
try:
# get all important words
base_word = request.json['base_word']
class_name = request.json['class_name']
print('recv classname:', class_name)
important_list = []
for classname in class_name:
remap_name = class_name_remap[classname]
word_df = word_importance_model.ix[remap_name]
important_list += list(flatten(word_df[[2]].head(20).values.tolist()))
res['important'] = important_list
except Exception as e:
print(e)
print("Important words:", res)
return res
def post(self):
res = {}
try:
# get all industry words
base_word = request.json['base_word']
class_name = request.json['class_name']
industry_list = []
for classname in class_name:
value_name = label2value[classname]
industry_df = industry_word_model.loc[
industry_word_model.classname.str.contains(value_name)]
industry_list += list(flatten(industry_df[[1]].head(20).values.tolist()))
res['industry'] = industry_list
except Exception as e:
print(e)
print("Industry words:", res)
return res
def post(self):
res = {}
try:
# get all important words
base_word = request.json['base_word']
class_name = request.json['class_name']
print('recv classname:', class_name)
important_list = []
for classname in class_name:
remap_name = class_name_remap[classname]
word_df = word_importance_model.ix[remap_name]
important_list += list(
flatten(word_df[[2]].head(20).values.tolist()))
res['important'] = important_list
except Exception as e:
print(e)
print("Important words:", res)
return res
def post(self):
res = {}
try:
# get all industry words
base_word = request.json['base_word']
class_name = request.json['class_name']
industry_list = []
for classname in class_name:
value_name = label2value[classname]
industry_df = industry_word_model.loc[
industry_word_model.classname.str.contains(value_name)]
industry_list += list(
flatten(industry_df[[1]].head(20).values.tolist()))
res['industry'] = industry_list
except Exception as e:
print(e)
print("Industry words:", res)
return res
def get_tx_events(nodes):
run('mkdir ./logs/')
run('rm -rf ./logs/*_txs')
all_events = pmap(get_tx_events_single_node, nodes)
return sorted(data.flatten(all_events))
#channels = range(0, 5) # DO NOT CHANGE
surrounding = 250
labels = [
0, 1, 1, 0, 2, 1, 2, 0, 0, 2, 1, 0, 2, 1, 1, 0, 1, 1, 0, 0, 1, 1, 1, 0, 2,
1, 0, 0, 1, 2, 1, 2, 1, 2, 2, 0, 1, 0, 1, 2, 2, 0, 2, 2, 1, 2, 0, 0, 0, 1,
0, 0, 2, 2, 2, 2, 2, 1, 2, 1, 0, 2, 2, 0, 0, 2, 0, 2, 2, 1, 1, 2, 2, 0, 1,
1, 2, 1, 2, 1, 0, 0, 0, 2, 0, 1, 2, 2, 0, 0, 1, 0, 2, 1, 2, 2, 1, 2, 2, 1,
0, 1, 0, 1, 1, 0, 1, 0, 0, 2, 2, 2, 0, 0, 1, 0, 2, 0, 2, 2, 0, 2, 0, 1, 0,
1, 1, 0, 0, 1, 0, 1, 1, 0, 1, 1, 1, 1, 2, 0, 0, 2, 1, 2, 1, 2, 2, 1, 2, 0
]
sequence_groups = transform_data([
data.flatten(
data.process_scrambled(labels, ['colors2.txt'],
channels=channels,
sample_rate=250,
surrounding=surrounding,
exclude=set([71, 155, 317])))
])
#silence_group = transform_data(data.process(1, ['math6_silence.txt'], channels=channels, sample_rate=250, surrounding=surrounding))
silence_group = transform_data(
data.process_silence_between(['colors2.txt'],
channels=channels,
sample_rate=250,
surrounding=surrounding))
sequence_groups = np.array(list(sequence_groups) + list(silence_group))
print len(sequence_groups)
print map(len, sequence_groups)
def get_tx_events(nodes, tx_filename):
run('mkdir ./logs/')
run('rm -rf ./logs/*_txs')
all_events = pmap(
lambda node: get_tx_events_single_node(node, tx_filename), nodes)
return sorted(data.flatten(all_events))
