From 9619ddcaa28a787d0ba261f0708d310019be5de4 Mon Sep 17 00:00:00 2001
From: Laurent Bachelier <laurent@bachelier.name>
Date: Wed, 4 Jun 2014 00:49:00 +0200
Subject: [PATCH] browser2: Add RawText filter

Allows getting .text of elements without any alteration.
This is useful for at least textarea and pre tags.

Maybe the .join character should be configurable.
---
 weboob/tools/browser2/filters.py | 12 ++++++++++++
 1 file changed, 12 insertions(+)

diff --git a/weboob/tools/browser2/filters.py b/weboob/tools/browser2/filters.py
index 6be44419..c87ef0a1 100644
--- a/weboob/tools/browser2/filters.py
+++ b/weboob/tools/browser2/filters.py
@@ -191,6 +191,18 @@ class CleanHTML(Filter):
     def clean(cls, txt):
         return html2text(html.tostring(txt, encoding=unicode))
 
+
+class RawText(Filter):
+    def filter(self, el):
+        if isinstance(el, (tuple,list)):
+            return u' '.join([self.filter(e) for e in el])
+
+        if el.text is None:
+            return self.default
+        else:
+            return unicode(el.text)
+
+
 class CleanText(Filter):
     """
     Get a cleaned text from an element.