if not os.path.exists(data_folder):
os.makedirs(data_folder)
if next(os.walk(data_folder))[1]:
retrain = False
if retrain:
input, target, classes = data.sample(data_folder)
model = classifier.build(input.shape, target.shape)
classifier.train(model, input, target)
classifier.save(models_folder, model, classes)
else:
model, classes = classifier.load(models_folder, sorted(os.listdir(models_folder))[-1])
for root, dirs, files in os.walk(data_folder):
for file in files:
if not file.startswith('.'):
with open(root+'/'+file) as f:
input = data.str2mat(f.read())
output = classifier.run(model, input)
<<<<<<< HEAD
data.backtest(save_loc, classes, input, output)
=======
data.backtest(save_loc+'/'+file, classes, input, output)
>>>>>>> refs/remotes/nathankjer/master
else:
print("""\nNo data found.\nPut subfolders of files by class, within the 'data' folder.""")
if __name__ == "__main__":
main()
