Update add_ons_to_gss command to support Python 2.

The command would error when running with Python 2.7, failing to
recognise the Welsh characters `ô` and `â` with "UnicodeEncodeError:
'ascii' codec can't encode character u'\xf4' in position 16: ordinal not
in range(128)". This is due to `csv.reader` changing between Python 2
and 3 - in 2 it expected a file-like object on which next() would return
raw bytes, and in 3 one which would return unicode strings. We opt to
explicitly check which version of Python is running and run the old code
for < Python 3.

We also stop assigning `name` for the 'code-change.csv' as it is not
used anywhere else in the script and the encoding probably won't be
handled correctly with Python 2. We stop assigning `type` and `name` for
'missing-codes.csv' for the same reasons.

mapit_gb/management/commands/mapit_UK_add_ons_to_gss.py

@@ -4,10 +4,23 @@
 import codecs
 import csv
 import os.path
+import sys
 from django.core.management.base import BaseCommand
 from mapit.models import Area, CodeType
 from psycopg2 import IntegrityError
 
+python_version = sys.version_info[0]
+
+
+def open_csv(filename):
+    if python_version < 3:
+        o = open(filename)
+    else:
+        o = codecs.open(filename, 'r', 'latin-1')
+    mapping = csv.reader(o)
+    next(mapping)
+    return mapping
+
 
 def process(new_code, old_code):
     try:
@@ -30,22 +43,10 @@ class Command(BaseCommand):
     help = 'Inserts the old ONS codes into mapit'
 
     def handle(self, **options):
-        code_changes = codecs.open(os.path.join(
-            os.path.dirname(__file__),
-            '../../data/BL-2010-10-code-change.csv'
-            ), 'r', 'latin-1')
-        mapping = csv.reader(code_changes)
-        next(mapping)
-        for row in mapping:
-            new_code, name, old_code = row[0], row[1], row[3]
+        for row in open_csv(os.path.join(os.path.dirname(__file__), '../../data/BL-2010-10-code-change.csv')):
+            new_code, old_code = row[0], row[3]
             process(new_code, old_code)
 
-        missing_codes = codecs.open(os.path.join(
-            os.path.dirname(__file__),
-            '../../data/BL-2010-10-missing-codes.csv'
-            ), 'r', 'latin-1')
-        mapping = csv.reader(missing_codes)
-        next(mapping)
-        for row in mapping:
-            type, new_code, old_code, name = row
+        for row in open_csv(os.path.join(os.path.dirname(__file__), '../../data/BL-2010-10-missing-codes.csv')):
+            new_code, old_code = row[1], row[2]
             process(new_code, old_code)