Fixed encoding.

[bookmarks_db.git] / Robots / parse_html_beautifulsoup.py

diff --git a/Robots/parse_html_beautifulsoup.py b/Robots/parse_html_beautifulsoup.py
index 209486c4054f015a179d4469ccaa5adc3f8bbe8f..e03dfce99faa55b980e21a75e38249790d50c0a7 100644 (file)
--- a/Robots/parse_html_beautifulsoup.py
+++ b/Robots/parse_html_beautifulsoup.py
@@ -1,23 +1,13 @@
 """
    HTML Parser using BeautifulSoup
 
-   Written by BroytMann. Copyright (C) 2007 PhiloSoft Design
+   Written by Broytman. Copyright (C) 2007-2010 PhiloSoft Design
 """
 
 import re
 from sgmllib import SGMLParser, SGMLParseError
-from HTMLParser import HTMLParser
 from BeautifulSoup import BeautifulSoup, CData
-
-
-class BSoupParser(HTMLParser):
-   def __init__(self, charset, meta, title, refresh, icon):
-      object.__init__(self)
-      self.charset = charset
-      self.meta_charset = meta
-      self.title = title
-      self.refresh = refresh
-      self.icon = icon
+from parse_html_util import HTMLParser
 
 
 # http://groups.google.com/group/beautifulsoup/browse_thread/thread/69093cb0d3a3cf63
@@ -39,7 +29,7 @@ class BadDeclParser(BeautifulSoup):
              except SGMLParseError:
                  # Could not parse the DOCTYPE declaration
                  # Try to just skip the actual declaration
-                 match = re.search(r'<!DOCTYPE([^>]*?)>', self.rawdata[i:], re.MULTILINE)
+                 match = re.search(r'<!DOCTYPE([^>]*?)>', self.rawdata[i:], re.MULTILINE|re.IGNORECASE)
                  if match:
                      toHandle = self.rawdata[i:match.end()]
                  else:
@@ -49,28 +39,74 @@ class BadDeclParser(BeautifulSoup):
          return j
 
 
-def parse_html(filename, charset=None):
+def _parse_html(filename, charset):
    infile = open(filename, 'r')
    try:
-      root = BadDeclParser(infile, fromEncoding=charset)
+      return BadDeclParser(infile, fromEncoding=charset)
    except TypeError:
       return None
    finally:
       infile.close()
 
-   try:
-      head = root.html.head
-   except AttributeError:
+def parse_html(filename, charset=None, log=None):
+   root = _parse_html(filename, charset)
+   if root is None:
       return None
 
+   _charset = root.originalEncoding
+   if _charset in ("ISO-8859-2", "windows-1252", "MacCyrillic"): # Replace default
+      _charset = DEFAULT_CHARSET
+      root = _parse_html(filename, _charset)
+      if root is None:
+         return None
+
+   html = root.html
+   if html is None:
+      html = root
+
+   head = html.head
    if head is None:
-      head = root.html # Some sites put TITLE in HTML without HEAD
+      head = html # Some sites put TITLE in HTML without HEAD
 
-   _charset = root.originalEncoding
-   try:
-      title = head.title.string.encode(_charset)
-   except AttributeError:
-      title = '' # HEAD but no TITLE
+   title = head.title
+   if (title is None) and (html is not head):
+      # Some sites put TITLE in HTML outside of HEAD
+      title = html.title
+
+   if title is None:
+      # Lookup TITLE in the root
+      title = root.title
+
+   if title is None:
+      return None
+
+   if title.string:
+      title = title.string
+   else:
+      parts = []
+      for part in title:
+         if not isinstance(part, basestring):
+            part = unicode(part)
+         parts.append(part.strip())
+      title = ''.join(parts)
+
+   meta = head.find(_find_contenttype, recursive=False)
+   if meta:
+      try:
+         meta_content = meta.get("content")
+         if meta_content:
+             __charset = meta_content.lower().split('charset=')[1].split(';')[0]
+         else:
+             __charset = False
+      except IndexError: # No charset in the META Content-Type
+         meta_charset = False
+      else:
+         meta_charset = _charset == __charset
+   else:
+      meta_charset = False
+
+   if _charset or meta_charset:
+      title = title.encode(_charset or meta_charset)
 
    meta = head.find(_find_refresh, recursive=False)
    if meta:
@@ -84,7 +120,11 @@ def parse_html(filename, charset=None):
    else:
       icon = None
 
-   return BSoupParser(_charset, _charset != charset, title, refresh, icon)
+   return HTMLParser(_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 \