def to_bow(train_path,test_path):
train=seq.create_dataset(train_path)
test=seq.create_dataset(test_path)
bow=create_dict(train.instances +test.instances)
train_feats=create_vectors(train,bow)
test_feats=create_vectors(test,bow)
train_data=(train_feats,train.get_labels())
test_data=(test_feats,test.get_labels())
return train_data,test_data
def unify(a_path,b_path,out_path):
a_dataset=seq.create_dataset(a_path)
b_dataset=seq.create_dataset(b_path)
a_vec=bow.compute_bow(a_dataset)
b_vec=bow.compute_bow(b_dataset)
united_vec=[av +bv for av,bv in zip(a_vec,b_vec)]
labels=a_dataset.get_labels()
persons=a_dataset.get_persons()
utils.to_labeled_file(out_path,united_vec,labels)
return united_vec,persons,labels
def basic_bow(in_path,_bow=None):
dataset=seq.create_dataset(in_path)
labels=dataset.get_labels()
if(_bow==None):
vectors,_bow=bow.compute_bow(dataset,True)
vectors=np.array(vectors)
return _bow,(vectors,labels)
else:
vectors=bow.apply_bow(dataset,_bow)
vectors=np.array(vectors)
return vectors,labels
def get_unlabeled_vectors(in_path,compute_features,suffix=".csv"):
out_path=in_path.replace(".seq",suffix)
dataset=seq.create_dataset(in_path)
vectors=compute_features(dataset)
utils.to_csv_file(out_path,vectors,labels)
def get_labeled_vectors(in_path,compute_features,suffix=".lb"):
out_path=in_path.replace(".seq",suffix)
dataset=seq.create_dataset(in_path)
labels=dataset.get_labels()
vectors=compute_features(dataset)
utils.to_labeled_file(out_path,vectors,labels)
