A script to generate alliterative nicknames

nicknamer.py

#!/usr/bin/env python
from nltk.corpus import brown
import argparse


def is_valid(name, word, tag=None, allowed_parts_of_speech=None):
    """
    Tests if allowed part of speech and is alliterative
    """

    # Must be allowed part of speech
    if tag and tag not in allowed_parts_of_speech:
        return False

    # Must be alliterative
    if word.lower()[0] != name.lower()[0]:
        return False

    return True


def get_brown_candidates(name, allowed_parts_of_speech, categories=None):
    """
    Uses some tagged data set (part of speech and text category) to generate
    candidates.
    """

    # Use some tagged data set for (word, part_of_speech) pairs
    # See: http://www.nltk.org/book/ch05.html
    tagged_words = brown.tagged_words(categories=categories, tagset='universal')

    # Generate candidates / deduplicate
    return list(set([ word.lower() for (word, tag) in tagged_words \
        if is_valid(name, word, tag=tag,
        allowed_parts_of_speech=allowed_parts_of_speech) ]))


if __name__ == '__main__':
    parser = argparse.ArgumentParser()
    parser.add_argument('name', help='The root nickname (suffix)')
    parser.add_argument('--skip_nouns', action='store_true', help='Skip nouns for prefixes')
    parser.add_argument('--skip_adjectives', action='store_true', help='Skip adjectives for prefixes')
    parser.add_argument('--categories', default=None, help=str(brown.categories()))

    args = parser.parse_args()

    # Determine allowed parts of speech
    allowed_parts_of_speech = set(['NOUN', 'ADJ'])
    if args.skip_nouns and args.skip_adjectives:
        raise Exception('Cannot specify both --skip_nouns and --skip_adjectives')
    elif args.skip_nouns:
        allowed_parts_of_speech.remove('NOUN')
    elif args.skip_adjectives:
        allowed_parts_of_speech.remove('ADJ')

    # Generate candidates
    categories = args.categories.split(',') if args.categories else None
    candidates = get_brown_candidates(args.name, allowed_parts_of_speech, categories=categories)

    # Show results
    for candidate in candidates:
        print '{} {}'.format(candidate, args.name)