Test meta charset by looking in META HTTP-Equiv.

[bookmarks_db.git] / Robots / parse_html_beautifulsoup.py

diff --git a/Robots/parse_html_beautifulsoup.py b/Robots/parse_html_beautifulsoup.py
index 209486c4054f015a179d4469ccaa5adc3f8bbe8f..c45af2b0221a61f441cf5165fc2a25f6fe978bb5 100644 (file)
--- a/Robots/parse_html_beautifulsoup.py
+++ b/Robots/parse_html_beautifulsoup.py
@@ -1,7 +1,7 @@
 """
    HTML Parser using BeautifulSoup
 
-   Written by BroytMann. Copyright (C) 2007 PhiloSoft Design
+   Written by BroytMann. Copyright (C) 2007, 2008 PhiloSoft Design
 """
 
 import re
@@ -67,11 +67,29 @@ def parse_html(filename, charset=None):
       head = root.html # Some sites put TITLE in HTML without HEAD
 
    _charset = root.originalEncoding
+   if _charset == "windows-1252": # Replace default
+      _charset = DEFAULT_CHARSET
+
    try:
       title = head.title.string.encode(_charset)
    except AttributeError:
       title = '' # HEAD but no TITLE
 
+   if (not title) and (head is not root.html):
+      # Some sites put TITLE in HTML outside of HEAD
+
+      try:
+         title = root.html.title.string.encode(_charset)
+      except AttributeError:
+         title = '' # no TITLE in HTML too
+
+   meta = head.find(_find_contenttype, recursive=False)
+   if meta:
+      __charset = meta.get("content").lower().split('charset=')[1].split(';')[0]
+      meta_charset = _charset == __charset
+   else:
+      meta_charset = False
+
    meta = head.find(_find_refresh, recursive=False)
    if meta:
       refresh = meta.get("content")
@@ -84,7 +102,11 @@ def parse_html(filename, charset=None):
    else:
       icon = None
 
-   return BSoupParser(_charset, _charset != charset, title, refresh, icon)
+   return BSoupParser(_charset, meta_charset, title, refresh, icon)
+
+def _find_contenttype(Tag):
+   return (Tag.name == "meta") and \
+      (Tag.get("http-equiv", '').lower() == "content-type")
 
 def _find_refresh(Tag):
    return (Tag.name == "meta") and \