adding zipcode cleaner

data_migrator/contrib/dutch.py

@@ -17,13 +17,35 @@ def clean_phone(v):
         ['+31620202020','+31620202020','+31201233456','+4402030203','+4402030203']
 
     Args:
-        v: value to clean
+        v (str): value to clean
 
     Returns:
-        cleaned phone number
+        str: cleaned phone number
     '''
 
     v = _PHONE_CHARS.sub('', v)
     v = _INTERNATIONAL_ZERO_START.sub('+', v)
     v = _MUNICIPALY_ZERO_START.sub('+31', v)
     return v
+
+
+ZIP_CODE=re.compile('^([0-9]{4})[\t ]*([a-zA-Z]{2})$')
+def clean_zip_code(v):
+    '''Cleans a dutch zipcode
+
+        >>> [clean_zip_code(x) for x in ['1234 aa', '1234AB', '1234   Ba']]
+        ['1234AA', '1234AB', '1234BA']
+
+    Args:
+        v (str): zipcode to clean
+
+    Returns:
+        str: cleaned zip code
+    '''
+    v = v.strip()
+    try:
+        z = ZIP_CODE.match(v).groups()
+        r = "%s%s" % (z[0], z[1].upper())  # Dutch zip code is 1234AB
+    except AttributeError:
+        r = v
+    return r

data_migrator/models/fields.py

@@ -174,6 +174,7 @@ def __init__(self, data_map, as_json=False, **kwargs):
         self.as_json = as_json
 
     def _value(self, v):
+        print self.name, v
         if v is None:
             return v
         else:

docs/ref/contrib.txt

@@ -14,6 +14,13 @@ Contrib reference
 .. autofunction:: data_migrator.contrib.read.read_map_from_csv
 
 
+contrib.dutch
+=============
+
 ``clean_phone``
 ---------------
 .. autofunction:: data_migrator.contrib.dutch.clean_phone
+
+``clean_zip_code``
+---------------
+.. autofunction:: data_migrator.contrib.dutch.clean_zip_code

test/test_contrib.py

@@ -3,7 +3,7 @@
 
 import unittest
 
-from data_migrator.contrib.dutch import clean_phone
+from data_migrator.contrib.dutch import clean_phone, clean_zip_code
 
 class TestDutch(unittest.TestCase):
     def test_phone(self):
@@ -17,3 +17,15 @@ def test_phone(self):
         ]
         for i, o in l:
             self.assertEquals(o, clean_phone(i))
+
+    def test_zip_code(self):
+        '''test zip code'''
+        l = [
+            ('1234 AB','1234AB'),
+            ('1234ba','1234BA'),
+            ('1234    ba','1234BA'),
+            ('1 2 3 4 A B','1 2 3 4 A B'),
+            ('blabla', 'blabla')
+        ]
+        for i, o in l:
+            self.assertEquals(o, clean_zip_code(i))