Update getiban script

arthurdejong · arthurdejong · commit 6c49ca83e2f1 · 2013-12-31T12:41:58.000+01:00
This switches to use the csv module to support multi-line column values.
This also handles some problems in the BBAN structure column that would
contain an IBAN structure.
diff --git a/getiban.py b/getiban.py
@@ -23,22 +23,20 @@
 Financial Telecommunication which is the official IBAN registrar) to get
 the data needed to correctly parse and validate IBANs."""
 
+import csv
 import urllib
 
 
 # The place where the current version of IBAN_Registry.txt can be downloaded.
 download_url = 'http://www.swift.com/dsp/resources/documents/IBAN_Registry.txt'
 
 
-def splitlines(f):
-    """Read lines from the TAB-delimited IBAN_Registry.txt file and return
-    a dictionary per read line. We clean up the values a bit because it
-    contains some junk."""
+def clean_row(row):
+    """Clean up a read row from the CSV file."""
     stripit = ' \t\n\r;:\'"'
-    firstline = [x.strip(stripit) for x in f.readline().lower().split('\t')]
-    for line in f:
-        yield dict(zip(firstline, [x.strip(stripit)
-                                   for x in line.split('\t')]))
+    return dict(
+        (k.strip(stripit).lower(), v.strip(stripit))
+        for k, v in row.items())
 
 
 def get_country_codes(line):
@@ -54,11 +52,17 @@ def parse(f):
     """Parse the specified file."""
     print '# generated from IBAN_Registry.txt, downloaded from'
     print '# %s' % download_url
-    for line in splitlines(f):
-        for cc in get_country_codes(line):
+    for row in csv.DictReader(f, delimiter='\t', quotechar='"'):
+        row = clean_row(row)
+        bban = row['bban structure']
+        if not(bban) or bban.lower() == 'not in use':
+            bban = row['iban structure']
+        for cc in get_country_codes(row):
+            if bban.startswith(cc + '2!n'):
+                bban = bban[5:]
             # print country line
             print '%s country="%s" bban="%s"' % (
-                cc, line['name of country'], line['bban structure'])
+                cc, row['name of country'], bban)
             # TODO: some countries have a fixed check digit value
             # TODO: some countries have extra check digits
             # TODO: use "Bank identifier position within the BBAN" field
