def trainIncoming(name):
from geocoder.deduper import DatabaseGazetteer
import simplejson as json
import dedupe
engine = create_engine(DB_CONN)
deduper = DatabaseGazetteer([{'field': 'complete_address', 'type': 'Address'}],
engine=engine)
sql_table = checkForTable(engine, name)
if sql_table == None:
sys.exit()
primary_key = sql_table.primary_key.columns.keys()[0]
messy_table = '''
SELECT {0}, complete_address
FROM {1}
WHERE address_id IS NULL
AND complete_address IS NOT NULL
'''.format(primary_key, name)
curs = engine.execute(messy_table)
messy_data = ({'complete_address': r.complete_address} for r in curs)
deduper.drawSample(messy_data, sample_size=30000)
if os.path.exists('geocoder/data/training.json'):
print('reading labeled examples from geocoder/data/training.json')
with open('geocoder/data/training.json') as tf :
deduper.readTraining(tf)
dedupe.consoleLabel(deduper)
deduper.train(ppc=0.1, index_predicates=False)
# When finished, save our training away to disk
with open('geocoder/data/training.json', 'w') as tf :
deduper.writeTraining(tf)
# Save our weights and predicates to disk. If the settings file
# exists, we will skip all the training and learning next time we run
# this file.
with open('geocoder/dedupe.settings', 'wb') as sf :
deduper.writeSettings(sf)
deduper.cleanupTraining()
#
# suburbs.mergeTables()
connection.close()
if args.train:
from geocoder.deduper import DatabaseGazetteer
import simplejson as json
import dedupe
from geocoder.app_config import DB_CONN
engine = create_engine(DB_CONN)
deduper = DatabaseGazetteer([{
'field': 'complete_address',
'type': 'Address'
}],
engine=engine)
messy_data = json.load(open('geocoder/data/messy_addresses.json'))
deduper.drawSample(messy_data, sample_size=30000)
if os.path.exists('geocoder/data/training.json'):
print('reading labeled examples from geocoder/data/training.json')
with open('geocoder/data/training.json') as tf:
deduper.readTraining(tf)
dedupe.consoleLabel(deduper)
deduper.train(ppc=0.1, index_predicates=False)
#
# suburbs.run(download_url=download_url)
#
# suburbs.mergeTables()
connection.close()
if args.train:
from geocoder.deduper import DatabaseGazetteer
import simplejson as json
import dedupe
from geocoder.app_config import DB_CONN
engine = create_engine(DB_CONN)
deduper = DatabaseGazetteer([{'field': 'complete_address', 'type': 'Address'}],
engine=engine)
messy_data = json.load(open('geocoder/data/messy_addresses.json'))
deduper.drawSample(messy_data, sample_size=30000)
if os.path.exists('geocoder/data/training.json'):
print('reading labeled examples from geocoder/data/training.json')
with open('geocoder/data/training.json') as tf :
deduper.readTraining(tf)
dedupe.consoleLabel(deduper)
deduper.train(ppc=0.1, index_predicates=False)
# When finished, save our training away to disk
with open('geocoder/data/training.json', 'w') as tf :
def trainIncoming(name):
from geocoder.deduper import DatabaseGazetteer
import simplejson as json
import dedupe
engine = create_engine(DB_CONN)
deduper = DatabaseGazetteer([{
'field': 'complete_address',
'type': 'Address'
}],
engine=engine)
sql_table = checkForTable(engine, name)
if sql_table == None:
sys.exit()
primary_key = sql_table.primary_key.columns.keys()[0]
messy_table = '''
SELECT {0}, complete_address
FROM {1}
WHERE address_id IS NULL
AND complete_address IS NOT NULL
'''.format(primary_key, name)
curs = engine.execute(messy_table)
messy_data = ({'complete_address': r.complete_address} for r in curs)
deduper.drawSample(messy_data, sample_size=30000)
if os.path.exists('geocoder/data/training.json'):
print('reading labeled examples from geocoder/data/training.json')
with open('geocoder/data/training.json') as tf:
deduper.readTraining(tf)
dedupe.consoleLabel(deduper)
deduper.train(ppc=0.1, index_predicates=False)
# When finished, save our training away to disk
with open('geocoder/data/training.json', 'w') as tf:
deduper.writeTraining(tf)
# Save our weights and predicates to disk. If the settings file
# exists, we will skip all the training and learning next time we run
# this file.
with open('geocoder/dedupe.settings', 'wb') as sf:
deduper.writeSettings(sf)
deduper.cleanupTraining()
