Dissociator to generate names by example

namegen.py

#!/usr/bin/env python
#
# Based on https://gist.github.com/plaidfluff/aa9983161b2b56f8f2750d661279cb9e
#
# Silly thing to generate random names from examples. Uses the CSV files obtained from
# http://catalog.data.gov/dataset/baby-names-from-social-security-card-applications-national-level-data
#
# Any corpus will work if it's formatted like:
#   Name,[ignored],weight

import csv
import sys
import collections
import random
import argparse

parser = argparse.ArgumentParser(description='Generate one or more random names. Takes data such as http://catalog.data.gov/dataset/baby-names-from-social-security-card-applications-national-level-data as input.')
parser.add_argument('-n', '--count', type=int, default=1, help='Number of names to generate.')
options, filenames = parser.parse_known_args()
if len(filenames) == 0:
    parser.error('Must pass at least one file containing raw name data like "Sam,G,9999".')

class Node:
    def __init__(self):
        self.total = 0
        self.next_nodes = collections.defaultdict(int)

nodes = collections.defaultdict(Node)

for arg in filenames:
    with open(arg, 'r') as file:
        reader = csv.reader(file)
        for row in reader:
            weight = int(row[2])
            node = nodes[None]
            pos = 0
            for c in row[0]:
                node.total += weight
                node.next_nodes[c] += weight
                node = nodes[(c,pos)]
                pos += 1
            node.total += weight
            node.next_nodes[None] += weight

# for letter,weights in nodes.items():
#     print "{} = {}".format(letter, weights.total)
#     for nn,wt in weights.next_nodes.items():
#         print "    -> {} = {}".format(nn, wt)
# sys.exit(0)

def pick_weighted(node):
    rnd = random.randint(0, node.total)
    # print "{} {}".format(node.total, rnd)
    for k,v in node.next_nodes.items():
        rnd -= v
        if rnd <= 0:
            return k

def make_up_one_name():
    out = ''
    node = nodes[None]
    pos = 0
    while True:
        letter = pick_weighted(node)
        if not letter:
            break
        out += letter
        node = nodes[(letter, pos)]
        pos += 1

    return out

def make_up_whole_name():
    given_name = None
    middle_name = None
    family_name = None
    cutoff = len('Joseana')
    while given_name is None or family_name is None:
        new_name = make_up_one_name()
        if len(new_name) <= cutoff:
            if given_name is not None:
                middle_name = new_name
            else:
                given_name = new_name
        else:
            family_name = new_name
    return given_name, middle_name, family_name

def generate_names(count):
    while count > 0:
        yield make_up_whole_name()
        count = count - 1

for given_name, middle_name, family_name in generate_names(options.count):
    print(' '.join([given_name] + [middle_name] * bool(middle_name) + [family_name]))

Example output:

• Gelviss Mesele Maizaeler
• Ellon Frysthas
• Midob Alil Dariqung
• Jamera Ischua Degareus
• Jesene Dohn Eriandas
• Gatold Anialice
• Jonney Elvie Rynneland
• Moharam Anily Mandicin
• Janan Phyc Milianera
• Cydnze Tachon Degartto
• Jolly Brishas Deonolla
• Beles Chrrie Flizamiater
• Whede Dosstn Ermenigh
• Kichas Jasherdra
• Saman Chexa Pebrinether
• Hary Kesse Carrryne
• Aammy Jouirtina
• Genort Lathara Jarobyny
• Rimmen Angenet Mishabene
• Larkabe Genien Elexarder