From 81e3da8bb0e5caa70b2d3a17b8d5af31b97a608b Mon Sep 17 00:00:00 2001
From: Laurent Bachelier <laurent@bachelier.name>
Date: Mon, 4 Feb 2013 15:16:03 +0100
Subject: [PATCH] More robust currency guessing

This allows to parse PayPal balances which includes both symbol and letters, for
instance "$ 42,00 USD".
Also, it does not require adding any new symbols to the regular
expression.
It might break modules, though numerous cases were tested.
---
 setup.cfg                   |  3 ++-
 weboob/capabilities/bank.py | 28 +++++++++++++++++++++++++---
 2 files changed, 27 insertions(+), 4 deletions(-)

diff --git a/setup.cfg b/setup.cfg
index 751135da..ff1ce8c8 100644
--- a/setup.cfg
+++ b/setup.cfg
@@ -1,5 +1,6 @@
 [nosetests]
 verbosity = 2
 detailed-errors = 1
+with-doctest = 1
 where = weboob
-tests = weboob.tools.capabilities.paste,weboob.tools.path
+tests = weboob.tools.capabilities.paste,weboob.tools.path,weboob.capabilities.bank
diff --git a/weboob/capabilities/bank.py b/weboob/capabilities/bank.py
index 7777c67e..29cd9a25 100644
--- a/weboob/capabilities/bank.py
+++ b/weboob/capabilities/bank.py
@@ -41,7 +41,7 @@ class TransferError(UserError):
     A transfer has failed.
     """
 
-class Currency:
+class Currency(object):
     CUR_UNKNOWN        = 0
     CUR_EUR            = 1
     CUR_CHF            = 2
@@ -54,10 +54,32 @@ class Currency:
                u'USD': CUR_USD,
               }
 
+    EXTRACTOR = re.compile(r'[\d\s,\.\-]', re.UNICODE)
+
     @classmethod
     def get_currency(klass, text):
-        text = re.sub(u'[^A-Z€]', '', text.upper())
-        return klass.TXT2CUR.get(text, klass.CUR_UNKNOWN)
+        u"""
+        >>> Currency.get_currency(u'42')
+        0
+        >>> Currency.get_currency(u'42 €')
+        1
+        >>> Currency.get_currency(u'$42')
+        3
+        >>> Currency.get_currency(u'42.000,00€')
+        1
+        >>> Currency.get_currency(u'$42 USD')
+        3
+        >>> Currency.get_currency(u'%42 USD')
+        3
+        >>> Currency.get_currency(u'US1D')
+        0
+        """
+        curtexts = klass.EXTRACTOR.sub(' ', text.upper()).split()
+        for curtext in curtexts:
+            cur = klass.TXT2CUR.get(curtext)
+            if cur is not None:
+                return cur
+        return klass.CUR_UNKNOWN
 
 
 class Recipient(CapBaseObject):