import argparse
import DataLoad
import weightComputation as WC
import pytorchEnv as pe
MAXLEN = 500
parser = argparse.ArgumentParser()
parser.add_argument("--bert", help='if selected, model used will be bert')
parser.add_argument("--RNN", help='if selected, model used will be RNN')
args = parser.parse_args()
biblioVidJson = DataLoad.importDataFromJson("bibliovid.json")
Xbib, Ycat, Yspe, label_YCat_dict, label_YSpe_dict = DataLoad.GetDataWithAbstractBibliovid(
biblioVidJson)
TrainBib, Cat_TrainBib, ValBib, Cat_ValBib, TestBib, Cat_TestBib = DataLoad.ProcessSplitBibliovidData(
Xbib, Ycat, catDict=label_YCat_dict)
YCatWeight = WC.getWeight(Cat_TrainBib)
if args.bert:
X_trainBib, YCat_trainBib = pe.prepare_textsBert(TrainBib, Cat_TrainBib,
MAXLEN)
X_validBib, YCat_validBib = pe.prepare_textsBert(ValBib, Cat_ValBib,
MAXLEN)
X_testBib, YCat_testBib = pe.prepare_textsBert(TestBib, Cat_TestBib,
MAXLEN)
train_loader, valid_loader, test_loader = pe.getLoader(
