deepcom_insert_seq.py

from argparse import ArgumentParser
import nltk
nltk.download('punkt')

def process_seq(seq):
    tokens = nltk.word_tokenize(seq.lower())
    return tokens

def load_align(align_file):
    result = {}
    with open(align_file, 'r') as file:
        for line in file:
            line = line.strip()
            parts = line.split(' ', 1)
            example_id = parts[0]
            target_seq = parts[1]
            result[example_id] = process_seq(target_seq)
    return result

if __name__ == '__main__':
    parser = ArgumentParser()
    parser.add_argument("-in", "--input", dest="input_file",
                        help="path to input file", required=True)
    parser.add_argument("-out", "--output", dest="output_file", required=True)
    parser.add_argument("-al", "--align", dest="align_file", required=True)
    args = parser.parse_args()
    
    example_to_target = load_align(args.align_file)
    
    #os.mkdir(args.output_dir)
    with open(args.input_file, 'r') as infile:
        with open(args.output_file, 'w') as outfile:
            for i,line in enumerate(infile.readlines()):
                line = line.strip()
                parts = line.split(' ', 1)
                exampled_id = parts[0]
                contexts = parts[1]
                target_seq = example_to_target[exampled_id]
                target_seq_string = '|'.join(target_seq)
                outfile.write('%s %s\n' % (target_seq_string, contexts))