gca-sort

#!/usr/bin/env python3
# -*- coding: utf-8 -*-
from __future__ import print_function
from __future__ import unicode_literals

from gca.core import Abstract
from gca.util import make_fields, simplify_str

import argparse
import codecs
import sys

from six import string_types


def sanitize_field(x):
    if type(x) == bytes:
        return x.strip()
    elif isinstance(x, string_types):
        """ Why only the first character???
        if x[0] in s_conv:
            x = s_conv[x[0]] + x[1:]
        """
        return simplify_str(x)
    elif isinstance(x, (tuple, list)):
        x = [sanitize_field(xx) for xx in x]
    return x


def get_sorter(field):

    if field == '=':
        return None

    reverse = True
    if field.startswith('+'):
        reverse = False
        field = field[1:]

    fields = field

    def sort_fn(abstract):
        val = abstract.select_field(fields, fold=True)
        return sanitize_field(val)

    return {'fn': sort_fn, 'reverse': reverse}


def set_sort_id(abstract, sort_id, base, use_altid):

    old_sort_id = abstract.sort_id
    abstract.sort_id = sort_id + base

    if use_altid:
        abstract.alt_id = old_sort_id

    return abstract

if __name__ == '__main__':
    parser = argparse.ArgumentParser(description='GCA Sort - sort list of abstract by fields')
    parser.add_argument('file', type=str, default='-')
    parser.add_argument('fields', type=str, nargs='+', help='fields to sort for')
    parser.add_argument('--assign', type=int, default=None, help='assign a sortID')
    parser.add_argument('--altid', default=False, dest='altid', action='store_true')
    args = parser.parse_args()

    sorter = filter(lambda x: x is not None, [get_sorter(f) for f in args.fields])
    fd = codecs.open(args.file, 'r', encoding='utf-8') if args.file != '-' else sys.stdin
    abstracts = Abstract.from_data(fd.read())
    fd.close()

    for s in sorter:
        abstracts = sorted(abstracts, key=s['fn'], reverse=s['reverse'])

    if args.assign is not None:
        abstracts = [set_sort_id(a, idx, args.assign, args.altid) for idx, a in enumerate(abstracts)]

    data = Abstract.to_json(abstracts)
    sys.stdout.write(data)