USA is getting labeled with a feature code 'BLDG'

[shahvaiz]

Hello - If "USA" is in a sentence, it's getting tagged as a building. See below. Thanks so much.

geo.geoparse("We traveled to the USA")

[{'word': 'USA',
  'spans': [{'start': 19, 'end': 22}],
  'country_predicted': 'USA',
  'country_conf': 0.9998105,
  'geo': {'admin1': 'California',
   'lat': '34.00474',
   'lon': '-117.33588',
   'country_code3': 'USA',
   'geonameid': '7195491',
   'place_name': 'Sgi-Usa Riverside Community Center',
   'feature_class': 'S',
   'feature_code': 'BLDG'}}]

[ahalterman]

Thanks for the report. Mordecai wasn't really intended to geolocate country names, rather than subnational place names, but that's definitely not a good result to return.

[ylieder]

Same behavior occurs with "France": geo.geoparse("We traveled to France")

[{
  "word": "France",
  "spans": [
    {
      "start": 15,
      "end": 21
    }
  ],
  "country_predicted": "FRA",
  "country_conf": 0.9998105,
  "geo": {
    "admin1": "\u00cele-de-France",
    "lat": 48.94956,
    "lon": 2.5684,
    "country_code3": "FRA",
    "geonameid": "2971874",
    "place_name": "Tremblay-en-France",
    "feature_class": "P",
    "feature_code": "PPL",
    "country_code2": "FR"
  }
}]

Maybe it would be possible to weight records with feature class A.PCLI (independent political entity) higher, which are currently 192 countries.
I don't know, how the search is implemented, but both, "France" and "USA" are listed as aliases for the respective country in the GeoNames Gazette.

Reference: http://download.geonames.org/export/dump/featureCodes_en.txt

[akankshanb]

1. The same behavior is seen for China as well.
   geo.geoparse('We traveled to China')

[{'country_conf': 0.68758196,
  'country_predicted': 'CHN',
  'geo': {'admin1': 'Hubei',
          'country_code3': 'CHN',
          'feature_class': 'S',
          'feature_code': 'SCHC',
          'geonameid': '6620465',
          'lat': '30.52047',
          'lon': '114.39637',
          'place_name': 'China University of Geosciences'},
  'spans': [{'end': 20, 'start': 15}],
  'word': 'China'}]

2. It predicts Canada as a country if U.S in the sentence is used
   geo.geoparse('We traveled to the U.S')

[{'country_conf': 0.28868943,
  'country_predicted': 'CAN',
  'spans': [{'end': 22, 'start': 19}],
  'word': 'U.S'}]

Kindly look in this. Thanks!