GeoText extracts countries, nationalities, states and cities mentions from text.
It gets a block of text as input and produces a tuple of Place objects as a result respresenting detected countries, nationalities, states and cities.
Each Place object has the following fields:
- name: name of the palce, e.g. 'London', 'New York' for cities; 'France', 'Germany' for countries, etc.
- population: number of people living in this place, available only for cities and countries
Also there're additional place-specific fields.
City has:
- state: (optional, None by default) a State object representing region of the city, e.g. "State: California, United States"
- country: a Country (Place) object of this city
State has:
- country: a Country (Place) object of this state/region
Nationality object is the same as Country object ans represents countries mentioned by nationality.
See usage below for details.
- Free software: MIT license
from geotext import GeoText
geo_text = GeoText()
geo_text.read(
"I'm French, but live in NY. "
"I like to visit my friends in France from time to time."
)
geo_text.results
# Results(
# countries=(Country: France,),
# nationalities=(Country: France,),
# states=(),
# cities=(City: New York, New York, United States,)
# )
[city.name for city in geo_text.results.cities]
# ['New York']
city = geo_text.results.cities[0]
city.__dict__
# {'_key': 'London',
# '_search_field': 'london',
# 'country': Country: United Kingdom,
# 'name': 'London',
# 'population': 7556900,
# 'state': State: England, United Kingdom}
[country.name for country in geo_text.results.countries]
# ['France']
geo_text.get_country_mentions()
# OrderedDict([(Country: France, 2), (Country: United States, 1)])
GeoText('Voronezh and NY').get_country_mentions()
# OrderedDict([(Country: Russia, 1), (Country: United States, 1)])
GeoText('I live in Izumiōtsu').results.cities
# (City: Izumiotsu, Osaka, Japan,)
# Take only large cities into account
GeoText().read(
'Voronezh and New York', min_population=1000000
).get_country_mentions()
# OrderedDict([(Country: United States, 1)])- Fast
- Data from http://www.geonames.org licensed under the Creative Commons Attribution 3.0 License.
geography: geography is more advanced and bigger in scope compared to geotext and can do everything geotext does. On the other hand geotext is leaner: has no external dependencies, is faster (re vs nltk) and also depends on libraries and data covered with more permissive licenses.