from preprocessing.reader import DatasetReader
from utils.fileprovider import FileProvider
logging.getLogger().setLevel(logging.INFO)
if __name__ == '__main__':
"""##### Parameter parsing"""
parser = argparse.ArgumentParser(
description=
'A baseline based on returning the most common emoji given the user')
parser.add_argument('--workdir', required=True, help='Work path')
args = parser.parse_args()
files = FileProvider(args.workdir)
Y = []
X = []
users = {}
dictionary = {}
with open(files.evalita, 'r', encoding="utf-8") as reader:
for line in reader:
line = line.rstrip()
sample = json.loads(line)
uid = sample["uid"]
label = sample["label"]
if uid not in users:
users[uid] = {}
choices=["train", "userdata"],
help="Use user history to assist prediction")
parser.add_argument("--n-folds",
type=int,
default=10,
help="Use user history to assist prediction")
parser.add_argument("--gpu",
type=int,
default=0,
help="GPU ID to be used [0, 1, -1]")
args = parser.parse_args()
os.environ["CUDA_VISIBLE_DEVICES"] = "{}".format(args.gpu)
files = FileProvider(args.workdir)
logging.info("Starting training with parameters: {0}".format(vars(args)))
assert path.exists(files.evalita), "Unable to find {}".format(
files.evalita)
raw_train = EvalitaDatasetReader(files.evalita)
random_state = 42
raw_train, raw_test = raw_train.split(test_size=0.1,
random_state=random_state)
raw_real_test = EvalitaDatasetReader(files.evalita_real_test)
logging.info("Populating user history")
user_data = None
if args.use_history: