Generate timezone file from optd_por_best_known_so_far.csv

generate_optd_por_tz.py

#!/usr/bin/env python
# -*- coding: utf-8 -*-

import sys
from GeoBases import GeoBase

import pytz

def load_best_known(filename):
    with open(filename) as f:
        f.next() # skipping header
        for row in f:
            _, iata, lat, lng, _, _ = row.rstrip().split('^')
            geocode = float(lat), float(lng)
            yield iata, geocode

def main(filename):
    db_oripor = GeoBase('ori_por', verbose=False)
    db_geonames = GeoBase('cities', verbose=False)

    print 'iata_code^tz'
    for iata, geocode in load_best_known(filename):
        try:
            tz = db_oripor.get(iata, 'timezone')
            pytz.timezone(tz)
        except pytz.exceptions.UnknownTimeZoneError:
            # timezone in ori_por was not valid, we compute it from GeoNames
            _, id_ = db_geonames.findClosestFromPoint(geocode).next()
            tz = db_geonames.get(id_, 'timezone')
        print '{0}^{1}'.format(iata, tz)


if __name__ == '__main__':
    if len(sys.argv) < 2:
        print 'Need one argument: optd_por_best_known_so_far.csv file'
        exit(1)
    else:
        main(sys.argv[1])