gca-lint

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

from collections import defaultdict

from gca.core import Abstract

import argparse
import codecs
import sys


class Issue(object):
    def __init__(self, level, message, abstract):
        self.level = level
        self.message = message
        self.abstract = abstract

    def __str__(self):
        return '[%s] %s {%s}' % (self.level, self.message, self.abstract.uuid)

    def __repr__(self):
        return str(self)


def check_abstract(abstract, args):
    if abstract.title is None or not len(abstract.title):
        yield Issue('Error', 'Abstract has no title', abstract)

    if not len(abstract.authors):
        yield Issue('Error', 'Abstract has no authors', abstract)

    if not len(abstract.affiliations):
        yield Issue('Error', 'Abstract has no affiliations', abstract)

    for author in abstract.authors:
        if len(author.affiliations) < 1:
            aid = author.last_name if author.last_name else author.uuid
            yield Issue('Error', 'Author "%s" has no affiliation' % aid, abstract)

    nrefs = len(abstract.references)
    if nrefs > 10:
        yield Issue('Warning', 'Abstract has %d references' % nrefs, abstract)

    nfigures = len(abstract.figures)
    if nfigures > 1:
        yield Issue('Warning', 'Abstract has %d figures' % nfigures, abstract)

    check_log = not args.ignore_log
    if check_log and len(abstract.log) < 1:
        yield Issue('Warning', 'Abstract has no state log', abstract)

    check_owners = not args.ignore_owners
    if check_owners and len(abstract.owners) < 1:
        yield Issue('Error', 'Abstract has no owner', abstract)

    if abstract.sort_id < 1:
        yield Issue('Error', 'Abstract has no sort id', abstract)

if __name__ == '__main__':
    parser = argparse.ArgumentParser(description='GCA Linter')
    parser.add_argument('file', type=str, default='-')
    parser.add_argument('--ignore-log', action='store_true', default=False)
    parser.add_argument('--ignore-owners', action='store_true', default=False)
    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())
    abstract_map = {abstract.uuid: abstract for abstract in abstracts}

    issues = [(abstract.uuid, list(check_abstract(abstract, args))) for abstract in abstracts]
    issues = filter(lambda x: len(x[1]), issues)

    d = defaultdict(list)
    for i in issues:
        d[i[0]] += i[1]

    for k, v in d.items():
        abstract = abstract_map[k]
        author = abstract.authors[0].last_name if len(abstract.authors) else ''
        title = abstract.title if abstract.title is not None else ''
        print('%s' % k)
        print('  %.10s - %.60s...' % (author, title))
        for i in v:
            print('    [%s] %s' % (i.level, i.message))
        print('')

    print('%d abstract with issues.' % len(d))