COMPANY_NAME = 9
COUNTRY = 17
print "Opening %s to read..." % args.input
print "Opening %s to output..." % args.output
with open(args.input, "r") as csv_in:
reader = csv.DictReader(csv_in)
with open(args.output, "w") as csv_out:
writer = csv.writer(csv_out)
for in_row in reader:
out_row = [''] * 22
addr = in_row['ADDRESS']
if addr:
print "Cleaning address..."
print "Input: " + addr
addr = clean(addr)
if addr:
(street, city, state, postal, country) = addr
print "---- %s, %s, %s, %s, %s -----" % addr
out_row[STREET] = street if street else ''
out_row[CITY] = city if city else ''
out_row[STATE] = state if state else ''
out_row[POSTAL] = postal if postal else ''
out_row[COUNTRY] = country if country else ''
else:
print "NO MATCHING ADDRESS"
out_row[CUSIP6] = in_row['CUSIP6']
out_row[COMPANY_NAME] = in_row['NAME']
writer.writerow(out_row)
def test_clean_ireland_address(self):
addr = clean("70 SIR JOHN ROGERSON'S QUAY, Dublin 2, L2 2, Ireland")
self.assertEqual((u"70 Sir John Rogerson's Quay", u'Dublin', u'Dublin', None, u'Ireland'), addr)
reload(sys)
sys.setdefaultencoding('utf8')
parser = argparse.ArgumentParser(description='Geocode csv file.')
parser.add_argument("path", help="Path to csv")
parser.add_argument("field", help="Name of address field")
args = parser.parse_args()
def read_fieldnames(path):
with open(path, "r") as path_csv:
reader = csv.reader(path_csv)
row = reader.next()
return row
fieldnames = read_fieldnames(
args.path) + ['STREET', 'CITY', 'STATE', 'POSTAL', 'COUNTRY']
with open(args.path, "r") as path_csv:
reader = csv.DictReader(path_csv)
with open("geocoded.csv", "w") as geocoded_out:
writer = csv.DictWriter(geocoded_out, fieldnames)
for row in reader:
address = row[args.field]
print "Geocoding " + address
geo_address = clean(address)
if geo_address:
(row['STREET'], row['CITY'], row['STATE'], row['POSTAL'],
row['COUNTRY']) = geo_address
writer.writerow(row)
def test_clean_weird_chinese_address(self):
addr = clean('M5 1 Jiuxianqiao East Road, Chaoyang District, Beijing 100016, People s Republic of China')
self.assertEqual(None, addr)
import argparse
import sys
reload(sys)
sys.setdefaultencoding('utf8')
parser = argparse.ArgumentParser(description='Geocode csv file.')
parser.add_argument("path", help="Path to csv")
parser.add_argument("field", help="Name of address field")
args = parser.parse_args()
def read_fieldnames(path):
with open(path, "r") as path_csv:
reader = csv.reader(path_csv)
row = reader.next()
return row
fieldnames = read_fieldnames(args.path) + ['STREET', 'CITY', 'STATE', 'POSTAL', 'COUNTRY']
with open(args.path, "r") as path_csv:
reader = csv.DictReader(path_csv)
with open("geocoded.csv", "w") as geocoded_out:
writer = csv.DictWriter(geocoded_out, fieldnames)
for row in reader:
address = row[args.field]
print "Geocoding " + address
geo_address = clean(address)
if geo_address:
(row['STREET'], row['CITY'], row['STATE'], row['POSTAL'], row['COUNTRY']) = geo_address
writer.writerow(row)