leon-ai/leon
1
1
Fork 0
mirror of https://github.com/leon-ai/leon.git synced 2024-11-23 09:43:19 +03:00
Code Issues Projects Releases Wiki Activity

feat(python tcp server): improve city entity accuracy

Browse Source
This commit is contained in:
louistiti 2024-11-16 21:13:37 +08:00
parent 6d230498c4
commit f88e26b452
No known key found for this signature in database
GPG Key ID: 92CD6A2E497E1669
1 changed files with 24 additions and 23 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

47
tcp_server/src/lib/nlp.py
View File

@ -27,7 +27,7 @@ spacy_model_mapping = {
}
}
geonamescache = GeonamesCache()
geonamescache = GeonamesCache(min_city_population=5000)
countries = geonamescache.get_countries()
cities = geonamescache.get_cities()
@ -85,30 +85,31 @@ def extract_spacy_entities(utterance: str) -> list[dict]:
delete_unneeded_country_data(resolution['data'])
break
city_population = 0
for city in cities:
alternatenames = [name.casefold() for name in cities[city]['alternatenames']]
if cities[city]['name'].casefold() == ent.text.casefold() or ent.text.casefold() in alternatenames:
if city_population == 0:
entity += ':city'
if ':country' not in entity:
city_population = 0
for city in cities:
alternatenames = [name.casefold() for name in cities[city]['alternatenames']]
if cities[city]['name'].casefold() == ent.text.casefold() or ent.text.casefold() in alternatenames:
if city_population == 0:
entity += ':city'
if cities[city]['population'] > city_population:
resolution['data'] = copy.deepcopy(cities[city])
city_population = cities[city]['population']
if cities[city]['population'] > city_population:
resolution['data'] = copy.deepcopy(cities[city])
city_population = cities[city]['population']
for country in countries:
if countries[country]['iso'] == cities[city]['countrycode']:
resolution['data']['country'] = copy.deepcopy(countries[country])
break
try:
del resolution['data']['geonameid']
del resolution['data']['alternatenames']
del resolution['data']['admin1code']
delete_unneeded_country_data(resolution['data']['country'])
except BaseException:
pass
else:
continue
for country in countries:
if countries[country]['iso'] == cities[city]['countrycode']:
resolution['data']['country'] = copy.deepcopy(countries[country])
break
try:
del resolution['data']['geonameid']
del resolution['data']['alternatenames']
del resolution['data']['admin1code']
delete_unneeded_country_data(resolution['data']['country'])
except BaseException:
pass
else:
continue
entities.append({
'start': ent.start_char,