rebase.py

#!/usr/bin/env python3

def rebase_parser(rebase_file):
    """
    :param rebase_file: the rebase file to parse
    :type rebase_file: file object
    :return: at each call yields a tuple (str enz name, str binding site)
    :rtype: iterator
    """
    def clean_seq(seq):
        """
        remove each characters which are not a base
        """
        clean_seq = ''
        for char in seq:
            if char in 'ACGT':
                clean_seq += char
        return clean_seq

    for line in rebase_file:
        fields = line.split()
        name = fields[0]
        seq = clean_seq(fields[2])
        yield (name, seq)


def rebase2dict(rebase_path):
    """
    :param rebase_path: the path to rebase file to parse
    :type rebase_path: str
    :return: a dict with items (str enz name, str binding site)
    """
    with open(rebase_path, 'r') as rebase_input:
        # enz_dict = {}
        # for (name, seq) in rebase_parser(rebase_input):
        #     enz_dict[name] = seq
        enz_dict = dict(rebase_parser(rebase_input))
    return enz_dict


if __name__ == '__main__':
    import sys
    import os.path

    if len(sys.argv) != 2:
        sys.exit("Usage: rebase.py rebase_file")
    rebase_path = sys.argv[1]
    if not os.path.exists(rebase_path):
        sys.exit("No such file: {}".format(rebase_path))

    enz_dict = rebase2dict(rebase_path)
    print(enz_dict)