Always log guessed charset even if it's utf-8.

[bookmarks_db.git] / Robots / parse_html_beautifulsoup.py

diff --git a/Robots/parse_html_beautifulsoup.py b/Robots/parse_html_beautifulsoup.py
index 5f010ad698e894e63d86515fa2a0111ad926da4d..11db56371b0df277dd957484f782e54624721e40 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
@@ -55,27 +55,62 @@ def parse_html(filename, charset=None):
       root = BadDeclParser(infile, fromEncoding=charset)
    except TypeError:
       return None
-   infile.close()
+   finally:
+      infile.close()
 
-   _charset = root.originalEncoding
    try:
-      title = root.html.head.title.string.encode(_charset)
+      head = root.html.head
    except AttributeError:
       return None
 
-   meta = root.html.head.find(_find_refresh, recursive=False)
+   if head is 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:
+      try:
+         __charset = meta.get("content").lower().split('charset=')[1].split(';')[0]
+      except IndexError: # No charset in the META Content-Type
+         meta_charset = False
+      else:
+         meta_charset = _charset == __charset
+   else:
+      meta_charset = False
+
+   meta = head.find(_find_refresh, recursive=False)
    if meta:
       refresh = meta.get("content")
    else:
       refresh = None
 
-   meta = root.html.head.find(_find_icon, recursive=False)
+   meta = head.find(_find_icon, recursive=False)
    if meta:
       icon = meta.get("href")
    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 \