def split_dataset(seqs_info, split_options):
"""splits dataset for learning and testing
"""
if split_options["method"] == "wsample":
# try weighted sample
# first group seqs based on length
grouped_seqs = group_seqs_by_length(seqs_info)
# second get a weighted sample based on seqs length
w_sample = weighted_sample(
grouped_seqs, trainset_size=split_options["trainset_size"]
)
# print("w_sample ", w_sample)
# third aggregate the seqs in training category and testing category
data_split = aggregate_weightedsample(w_sample)
elif split_options["method"] == "cross_validation":
# try cross validation
seqs_id = list(seqs_info.keys())
data_split = split_data(seqs_id, split_options)
elif split_options["method"] == "random":
seqs_id = list(seqs_info.keys())
data_split = split_data(seqs_id, split_options)
elif split_options["method"] == "none":
seqs_id = list(seqs_info.keys())
data_split = {0: {"train": seqs_id}}
return data_split
def split_dataset(self, seqs_info, split_options):
if(split_options['method'] == "wsample"):
# try weighted sample
# first group seqs based on length
grouped_seqs = group_seqs_by_length(seqs_info)
# second get a weighted sample based on seqs length
w_sample = weighted_sample(grouped_seqs, trainset_size=split_options['trainset_size'])
print("w_sample ", w_sample)
# third aggregate the seqs in training category and testing category
data_split = aggregate_weightedsample(w_sample)
elif(split_options['method'] == "cross_validation"):
# try cross validation
seqs_id = list(seqs_info.keys())
data_split = split_data(seqs_id, split_options)
elif(split_options['method'] == 'random'):
seqs_id = list(seqs_info.keys())
data_split = split_data(seqs_id, split_options)
elif(split_options['method'] == 'none'):
seqs_id = list(seqs_info.keys())
data_split = {0:{'train':seqs_id}}
return(data_split)