import parser import Pyscandl from fetchers import fetcher_enum if __name__ == "__main__": args = parser.parse_arg() fetcher = fetcher_enum.Fetcher.get(args.fetcher).value pyscandl = Pyscandl.Pyscandl(fetcher, chapstart=args.chapter_start, output=args.output, keepimage=args.keep_images, all=args.all, link=args.link, manga=args.manga, download_number=args.download_number) pyscandl.full_download()
"question_weight": 1,
"concepts_weight": 1,
"narrative_weight": 1,
"query_path": "../queries/query-train.xml",
"cdn": 1,
"ctc": 1,
"cte": 1,
"cts": 1,
"unigram_weight": 1,
"bigram_weight": 1,
"rocchio_iters": 1,
"use_cosine": False
}
if __name__ == '__main__':
parser.parse_arg(configs)
print("Rocchio Mode: ", configs["use_rocchio"])
fname_to_id, id_to_fname = parser.parse_file_list(configs)
vocab_to_id, id_to_vocab = parser.parse_vocab_list(configs)
doc_count = len(fname_to_id)
inverted_files, gram_to_id, gram_count, id_to_doclen = parser.parse_inverted_file(
configs, doc_count)
configs["gram_count"] = gram_count
configs["doc_count"] = doc_count
# Save checkpoint for notebook
avdl = sum(id_to_doclen.values()) / len(id_to_doclen)
corpus = {
"fname_to_id": fname_to_id,
"id_to_doclen": id_to_doclen,
"id_to_fname": id_to_fname,
"vocab_to_id": vocab_to_id,
from train import TrainNet
from parser import parse_arg, parse_shape
class Predict(TrainNet):
def valid(self):
for i in range(self.data.len):
y, x = self.data.get_for_test(i)
x = x.reshape(x.shape[0], 1)
pred = self.forward(x)
if pred > 0.5:
p = "B"
else:
p = "M"
print("ID: ", self.data.ID[i], " Predict: ", pred[0, 0],
" Class: ", p)
if __name__ == "__main__":
args = parse_arg()
if not args.weight:
print("Give me weights")
exit(0)
else:
shape = parse_shape(args.weight)
print("Shape: ", shape)
net = Predict(shape, args.train)
net.read(args.weight)
net.valid()