dataset, tasktype))
parser.add_argument('--model-dir',
default='check_{}_{}_model/'.format(dataset, tasktype))
parser.add_argument(
'--mode',
choices=['classification', 'regression'],
default='regression' if tasktype == 'r' else 'classification')
args = parser.parse_args()
if not os.path.exists(args.model_dir):
os.mkdir(args.model_dir)
start_time = time.time()
print('Dataset:', args.train_csv)
df_X, df_y, model_config, _ = load_data(args.train_csv)
model_config['mode'] = args.mode
model_config_filename = os.path.join(args.model_dir, 'model_config.pkl')
with open(model_config_filename, 'wb') as fout:
pickle.dump(model_config, fout, protocol=pickle.HIGHEST_PROTOCOL)
# eval dataset during the training
size = min(df_X.shape[0] // 10, 1000)
X_train, X_eval, y_train, y_eval = train_test_split(df_X,
df_y,
test_size=size)
train_dir = 'catboost_info/'
if os.path.exists(train_dir):
parser.add_argument('--test-csv', required=True)
parser.add_argument('--prediction-csv',
type=argparse.FileType('w'),
required=True)
parser.add_argument('--model-dir', required=True)
args = parser.parse_args()
start_time = time.time()
# load model
model_config_filename = os.path.join(args.model_dir, 'model_config.pkl')
with open(model_config_filename, 'rb') as fin:
model_config = pickle.load(fin)
X_scaled, _, _, df = load_data(args.test_csv,
datatype='test',
cfg=model_config)
if model_config['time_leakage']['is_leakage']:
df[model_config['time_leakage']
['id_col']] = model_config['time_leakage']['id_series']
df[model_config['time_leakage']
['dt_col']] = model_config['time_leakage']['dt_series']
df[model_config['time_leakage']
['num_col']] = model_config['time_leakage']['num_series']
df = use_time_leakage(df, model_config['time_leakage'])
df['prediction'] = df['prediction'].fillna(0)
else:
model = model_config['model']
#df = pd.read_csv(args.test_csv, usecols=['line_id',])
#print(args.test_csv)