parse special cases

weboob/backends/bnporc/pages/account_history.py

@@ -18,16 +18,20 @@
 # along with weboob. If not, see <http://www.gnu.org/licenses/>.
 
 
+import re
 from datetime import date
 
 from weboob.tools.browser import BasePage
 from weboob.capabilities.bank import Operation
+from weboob.capabilities.base import NotAvailable
 
 
 __all__ = ['AccountHistory']
 
 
 class AccountHistory(BasePage):
+    LABEL_PATTERNS = [(u'^CHEQUEN°(?P<no>.*)', u'CHEQUE', u'N°%(no)s')]
+
     def on_loaded(self):
         self.operations = []
 
@@ -44,11 +48,21 @@ class AccountHistory(BasePage):
                     if child.text: label += child.text
                     if child.tail: label += child.tail
                 if tds[1].tail: label += tds[1].tail
+
                 label = label.strip()
+                category = NotAvailable
+                for pattern, _cat, _lab in self.LABEL_PATTERNS:
+                    m = re.match(pattern, label)
+                    if m:
+                        category = _cat % m.groupdict()
+                        label = _lab % m.groupdict()
+                        break
+                else:
+                    if '  ' in label:
+                        category, useless, label = [part.strip() for part in label.partition('  ')]
+
                 amount = tds[2].text.replace('.', '').replace(',', '.')
-                (category, useless, label) = label.partition('  ')
-                category = category.strip()
-                label = label.strip()
+
                 # if we don't have exactly one '.', this is not a floatm try the next
                 operation = Operation(len(self.operations))
                 if amount.count('.') != 1: