Get Timezone from City in Python/Django
pytz
is a wrapper around IANA Time Zone Database (Olson database). It does not contain data to map an arbitrary city in the world to the timezone it is in.
You might need a geocoder such as geopy
that can translate a place (e.g., a city name) to its coordinates (latitude, longitude) using various web-services:
from geopy import geocoders # pip install geopyg = geocoders.GoogleV3()place, (lat, lng) = g.geocode('Singapore')# -> (u'Singapore', (1.352083, 103.819836))
Given city's latitude, longitude, it is possible to find its timezone using tz_world, an efele.net/tz map / a shapefile of the TZ timezones of the world e.g., via postgis timezone db or pytzwhere
:
import tzwherew = tzwhere()print w.tzNameAt(1.352083, 103.819836)# -> Asia/Singapore
There are also web-services that allow to convert (latitude, longitude) into a timezone e.g., askgeo, geonames, see Timezone lookup from latitude longitude.
As @dashesy pointed out in the comment, geopy
also can find timezone (since 1.2):
timezone = g.timezone((lat, lng)) # return pytz timezone object# -> <DstTzInfo 'Asia/Singapore' LMT+6:55:00 STD>
GeoNames also provides offline data that allows to get city's timezone directly from its name e.g.:
#!/usr/bin/env pythonimport osfrom collections import defaultdictfrom datetime import datetimefrom urllib import urlretrievefrom urlparse import urljoinfrom zipfile import ZipFileimport pytz # pip install pytzgeonames_url = 'http://download.geonames.org/export/dump/'basename = 'cities15000' # all cities with a population > 15000 or capitalsfilename = basename + '.zip'# get fileif not os.path.exists(filename): urlretrieve(urljoin(geonames_url, filename), filename)# parse itcity2tz = defaultdict(set)with ZipFile(filename) as zf, zf.open(basename + '.txt') as file: for line in file: fields = line.split(b'\t') if fields: # geoname table http://download.geonames.org/export/dump/ name, asciiname, alternatenames = fields[1:4] timezone = fields[-2].decode('utf-8').strip() if timezone: for city in [name, asciiname] + alternatenames.split(b','): city = city.decode('utf-8').strip() if city: city2tz[city].add(timezone)print("Number of available city names (with aliases): %d" % len(city2tz))#n = sum((len(timezones) > 1) for city, timezones in city2tz.iteritems())print("")print("Find number of ambigious city names\n " "(that have more than one associated timezone): %d" % n)#fmt = '%Y-%m-%d %H:%M:%S %Z%z'city = "Zurich"for tzname in city2tz[city]: now = datetime.now(pytz.timezone(tzname)) print("") print("%s is in %s timezone" % (city, tzname)) print("Current time in %s is %s" % (city, now.strftime(fmt)))
Output
Number of available city names (with aliases): 112682Find number of ambigious city names (that have more than one associated timezone): 2318Zurich is in Europe/Zurich timezoneCurrent time in Zurich is 2013-05-13 11:36:33 CEST+0200
I think you're going to need to manually search the timezone database for the city you're looking for:
from pytz import country_timezones, timezonedef find_city(query): for country, cities in country_timezones.items(): for city in cities: if query in city: yield timezone(city)for tz in find_city('Zurich'): print(tz)
(that's just a quick-and-dirty solution, it for instance doesn't try to match only the city-part of a timezone – try searching for Europe
, it does substring matches, doesn't search case-insensitive, etc.)
there have been a lot of possible solutions proposed here and they're all a bit tedious to set up.
To make things quicker for the next person with this problem, I took the one from Will Charlton and made a quick python library out of it: https://pypi.python.org/pypi/whenareyou
from whenareyou import whenareyoutz = whenareyou('Hamburg')tz.localize(datetime(2002, 10, 27, 6, 0, 0))
Gets you datetime.datetime(2002, 10, 27, 6, 0, tzinfo=<DstTzInfo 'Europe/Berlin' CET+1:00:00 STD>)
.
This gets you a pytz object (tz
in the example) so you can use it pythonicly.
- It uses the google API
- Leaves daylight savings calculation to pytz, only one call per city, rest happens offline
- LRU caches the requests so you shouldn't hit the API limit easily
- Should also work with any address or anything google maps understands