closeio / tz-trout Goto Github PK

Description

Add a Github action in order to run tests and checks for the following Python versions:

• 3.6
• 3.7
• 3.8
• 3.9
• 3.10

https://github.com/pytries/DAWG seems abandoned 3 years ago. It fails to install on python 3.7. Perhaps we should get rid of it?

Antarctica/Macquarie (+11) should not be in ET (-04)

>>> [tz for tz in tztrout.td.tz_name_to_tz_ids['ET'] if tz.startswith('An')]
[u'Antarctica/Macquarie']

When a single area is extracted from the output of description_for_number, that's interpreted as if it were a state.

Here's an example where a phone from Philadelphia, PA thinks state is philadelphia instead of pennsylvania and fails to find it in normalized_states. Because of this, all the US timezone IDs will be returned for this number.

This behaviour is still present in phonenumbers latest version (8.13.1)

Example:

import phonenumbers
from phonenumbers.geocoder import description_for_number
from tztrout import tz_ids_for_phone

phone_number = "+14455000001"

tz_ids = tz_ids_for_phone(phone_number, "US")
print(tz_ids) # contains all of the US timezones

phone = phonenumbers.parse(phone_number, "US")
area = description_for_number(phone, "en").split(',')
print(area) # ["Philadelphia"]

• regenerate_data.py script should download http://download.geonames.org/export/zip/US.zip and convert it to a new json file in tztrout/data.
  • Keep JSON output consistent and diff-friendly (use this)
  • Only keep zip code, city name, 2-letter state code and coordinates to keep the size down.
  • This should be the first step since other files depend on this data.
  • Commit the data file into repo.
  • Provide attribution in the readme: ZIP code data is provided by www.geonames.org under under Creative Commons Attribution 3.0 License
• Load the data in memory in place of existing pyzipcode data. Make sure to deduplicate city name and state code strings.
  • Get rid of pyzipcode dependency (setup.py + requirements.txt)
• Since old database had timezones and we don't have it from geonames, use https://github.com/MrMinimal64/timezonefinder to look up timezones by zipcode's coordinates.
• Measure memory usage and tz lookup time before and after changes.
• Turn these into test cases:
  • print tztrout.tz_ids_for_address('US', state='NY', zipcode="10065") should return America/New_York but it returns None. This zipcode was introduced in 2006/7.
  • print tztrout.tz_ids_for_address('US', state='IL', zipcode="60642") should return America/Chicago but it returns None. This zipcode was introduced in 2008.
  • print tztrout.tz_ids_for_address('US', state='DC', zipcode="20022") should return America/New_York but it returns None. There's no data on when this zip was introduced at the end of 2004.

We're not going to use pyzipcode3 since @tsx has concerns about maintenance of this library. It only had one database upgrade 2 years ago and no transparent process for future upgrades.

Should be America/Los_Angeles, but instead it defaults to the whole US:

> tztrout.tz_ids_for_phone('+14422425150')
[u'America/New_York',
 u'America/Detroit',
 u'America/Kentucky/Louisville',
 u'America/Kentucky/Monticello',
 u'America/Indiana/Indianapolis',
 u'America/Indiana/Vincennes',
 u'America/Indiana/Winamac',
 u'America/Indiana/Marengo',
 u'America/Indiana/Petersburg',
 u'America/Indiana/Vevay',
 u'America/Chicago',
 u'America/Indiana/Tell_City',
 u'America/Indiana/Knox',
 u'America/Menominee',
 u'America/North_Dakota/Center',
 u'America/North_Dakota/New_Salem',
 u'America/North_Dakota/Beulah',
 u'America/Denver',
 u'America/Boise',
 u'America/Phoenix',
 u'America/Los_Angeles',
 u'America/Metlakatla',
 u'America/Anchorage',
 u'America/Juneau',
 u'America/Sitka',
 u'America/Yakutat',
 u'America/Nome',
 u'America/Adak',
 u'Pacific/Honolulu']

Internal reference: https://help.close.io/agent/case/74267

e.g. 416 area code

The timezone metadata is only loaded on the first use of one of the timezone functions (time_zones_for_number or time_zones_for_geographical_number).

https://github.com/daviddrysdale/python-phonenumbers

I think time-zones are state driven (the eight three letter acronyms below), so that is not too complex: just 8. All postcodes are mapped to a state, so that is easy too, you can easily get a csv of all postcodes in Australia mapped to States from Australia Post (we just bought some recent data that maps suburbs as well)
Australian phone codes are simple too:
Mobiles start with 04 or +614 all over the country, so that is of no use
Landline codes are +612 for NSW & ACT, +613 VIC&TAS, +617 QLD, and +618 for WA, SA and NT which may not be necessarily on same time zone (Source: http://australia.gov.au/about-australia/our-country/telephone-country-and-area-codes)