[bp] handle account details page not being available

Sometimes the details page is not yet accessible for some accounts.
Prevent crashing in this case.

modules/bp/browser.py

@@ -564,9 +564,15 @@ class BPBrowser(LoginBrowser, StatesMixin):
                     elif account.type == Account.TYPE_PERP:
                         # PERP balances must be fetched from the details page,
                         # otherwise we just scrape the "Rente annuelle estimée":
-                        balance = self.open(account.url).page.get_balance()
-                        if balance is not None:
-                            account.balance = balance
+                        balance_page = self.open(account.url).page
+                        # Sometimes the balance page is not available (with the site saying the account detail will
+                        # soon be available). In the meantime, just skip it and use the balance we already have.
+                        if balance_page and isinstance(balance_page, SavingAccountSummary):
+                            balance = balance_page.get_balance()
+                            if balance is not None:
+                                account.balance = balance
+                        else:
+                            self.logger.debug('Balance page not yet available for account %r', account)
                         accounts.append(account)
 
                     else:

modules/bp/pages/accountlist.py

@@ -116,6 +116,10 @@ class item_account_generic(ItemElement):
             coming = 0
 
             details_page = self.page.browser.open(Field('url')(self))
+            if not details_page.page:
+                # Details page might not always be available
+                return NotAvailable
+
             coming_op_link = Link(
                 '//a[contains(text(), "Opérations à venir")]',
                 default=NotAvailable
@@ -157,6 +161,10 @@ class item_account_generic(ItemElement):
             return NotAvailable
 
         details_page = self.page.browser.open(Field('url')(self)).page
+        if not details_page:
+            # Details page might not always be available
+            return NotAvailable
+
         rib_link = Link('//a[abbr[contains(text(), "RIB")]]', default=NotAvailable)(details_page.doc)
         if rib_link:
             response = self.page.browser.open(rib_link)
@@ -516,7 +524,12 @@ class AccountRIB(LoggedPage, RawPage):
     iban_regexp = r'[A-Z]{2}\d{12}[0-9A-Z]{11}\d{2}'
 
     def get_iban(self):
-        m = re.search(self.iban_regexp, extract_text(self.data))
+        content = extract_text(self.data)
+        if not content:
+            # This can happen if there's an error on the site while rendering the PDF (no RIB is shown)
+            return NotAvailable
+
+        m = re.search(self.iban_regexp, content)
         if m:
             return unicode(m.group(0))
         return None