gca-filter

#!/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 simplify_str

import sys
import argparse
import codecs


def make_filter(string, exact=False):
    x = string.split('=')
    assert(len(x) == 2)
    field = x[0].strip()
    fl = x[1].strip()

    if exact:
        return lambda abstract: abstract.select_field(field, fold=True) == fl
    else:
        fl = simplify_str(fl)
        return lambda abstract: simplify_str(abstract.select_field(field, fold=True)) == fl


if __name__ == '__main__':
    parser = argparse.ArgumentParser(description='GCA Filter - filter list of abstract by files')
    parser.add_argument('file', type=str, default='-')
    parser.add_argument('filter', type=str, nargs='+', help='filter to apply to each abstract')
    parser.add_argument('-e', '--exact', default=False, action='store_true', help='Do use non-modified strings for comparisons. Default is to compare lower case and simplified strings.')
    parser.add_argument('--start-at', dest='P', default=None, type=int, help='start at abstract number')
    parser.add_argument('--max-num', dest='N', default=None, type=int, help='maximum number of abstracts to take')
    parser.add_argument('--or', dest='ft_or', default=False, action='store_true', help='use OR instead of AND')
    args = parser.parse_args()

    fd = codecs.open(args.file, 'r', encoding='utf-8') if args.file != '-' else sys.stdin
    abstracts = Abstract.from_data(fd.read())
    fd.close()

    filter_list = [make_filter(f, args.exact) for f in args.filter]

    if args.ft_or:
        result = []
        for f in filter_list:
            result += filter(f, abstracts)
        de_dup = {k.uuid: k for k in result}
        abstracts = [a for k, a in de_dup.items()]
    else:
        for f in filter_list:
            abstracts = filter(f, abstracts)

    if args.P is not None:
        abstracts = abstracts[args.P-1:]

    if args.N is not None:
        abstracts = abstracts[:args.N]

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