generate-geo-data

#!/usr/bin/env python

from collections import defaultdict
from pprint import pformat
from pycountry import countries, subdivisions
from pycountry_convert import country_alpha2_to_continent_code


subs = defaultdict(dict)
for subdivision in subdivisions:
    # Route53 only supports US states, Dyn (and others) support US states and CA provinces
    if subdivision.country_code not in ('US', 'CA'):
        continue
    subs[subdivision.country_code][subdivision.code[3:]] = {
        'name': subdivision.name
    }
subs = dict(subs)
#pprint(subs)

# These are best guesses at things pycountry_convert doesn't map
continent_backups = {
    'AQ': 'AN',
    'EH': 'AF',
    'PN': 'OC',
    'SX': 'NA',
    'TF': 'AN',
    'TL': 'AS',
    'UM': 'OC',
    'VA': 'EU',
}

geos = defaultdict(dict)
for country in countries:
    try:
        continent_code = country_alpha2_to_continent_code(country.alpha_2)
    except KeyError:
        try:
            continent_code = continent_backups[country.alpha_2]
        except KeyError:
            raise
            print('{} {} {}'.format(country.alpha_2, country.name, getattr(country, 'official_name', '')))

    geos[continent_code][country.alpha_2] = {
        'name': country.name
    }

    try:
        geos[continent_code][country.alpha_2]['provinces'] = subs[country.alpha_2]
    except KeyError:
        pass

geos = dict(geos)
data = pformat(geos).replace('\n', '\n    ')

print('''#
# -*- coding: utf-8 -*-
#
# This file is generated using
#   ./script/generate-geo-data > octodns/record/geo_data.py
# do not modify it directly
#

geo_data = \\
    {}'''.format(data))