lambda x: len(x) <= second_layer_screen or len(x) >= 10000)
not_blds.reset_index(inplace=True)
# Recluster buildings in only x-y to negate the effects of funny-shaped rooves.
db = DBSCAN(eps=eps, min_samples=min_samples).fit(blds[['x',
'y']].as_matrix())
blds = pd.DataFrame([blds.id, db.labels_]).T
blds.columns = ['id', 'label']
blds['id'] = blds['id'].astype(int)
# Is as building: id, label
# Not a building: id
return blds, not_blds['id']
if __name__ == '__main__':
import os
from pgMethods import Methods
m = Methods(projection=32610)
df = m.pg_df('SELECT x, y, z, id FROM lidar_nyc_small')
print 'df loaded'
#df = pd.read_csv('../tests/data/CrossValidationData.csv')
blds, not_blds = dbscan(df)
m.df_pg(blds, 'temp12345678')
m.pg_post(
'DROP TABLE IF EXISTS checkcheck; CREATE TABLE checkcheck AS SELECT a.* FROM z_20130805_usgsnyc14_18TWL835075 a RIGHT JOIN temp12345678 b ON a.id=b.id;'
)
m.pg_post('DROP TABLE temp12345678;')
# Recluster buildings in only x-y to negate the effects of funny-shaped rooves.
db = DBSCAN(eps=eps, min_samples=min_samples).fit(blds[['x', 'y']].as_matrix())
blds = pd.DataFrame([blds.id, db.labels_]).T
blds.columns = ['id', 'label'];
blds['id'] = blds['id'].astype(int)
# Is as building: id, label
# Not a building: id
return blds, not_blds['id']
if __name__ == '__main__':
import os
from pgMethods import Methods
m = Methods(projection=32610);
df = m.pg_df('SELECT x, y, z, id FROM lidar_nyc_small')
print 'df loaded'
#df = pd.read_csv('../tests/data/CrossValidationData.csv')
blds, not_blds = dbscan(df)
m.df_pg(blds, 'temp12345678');
m.pg_post('DROP TABLE IF EXISTS checkcheck; CREATE TABLE checkcheck AS SELECT a.* FROM z_20130805_usgsnyc14_18TWL835075 a RIGHT JOIN temp12345678 b ON a.id=b.id;')
m.pg_post('DROP TABLE temp12345678;')
