Parse and recode unicode entities.

git-svn-id: file:///home/phd/archive/SVN/bookmarks_db/trunk@8 fdd5c36f-1aea-0310-aeeb-c58d7e2b6c23

diff --git a/Robots/parse_html.py b/Robots/parse_html.py
index e95dcef24e0e2f7dc76c8e3f2908b466e17e83b9..a534e21fa3a13379c9ef85801b027dbfec8669d4 100755 (executable)
--- a/Robots/parse_html.py
+++ b/Robots/parse_html.py
@@ -2,13 +2,13 @@
 """
    HTML Parser
 
-   Written by BroytMann, Jun 2002 - Aug 2002. Copyright (C) 1997-2002 PhiloSoft Design
+   Written by BroytMann, Jun 2002 - May 2003. Copyright (C) 1997-2003 PhiloSoft Design
 """
 
 
 import sys
 current_charset = sys.getdefaultencoding()
-DEFAULT_CHARSET = "windows-1251"
+DEFAULT_CHARSET = "windows-1251" # Stupid default for Russian Cyrillic
 
 
 from HTMLParser import HTMLParseError
@@ -46,7 +46,8 @@ class HTMLParser(_HTMLParser):
          try:
             # extract charset from "text/html; foo; charset=UTF-8; bar;"
             self.charset = content.lower().split('charset=')[1].split(';')[0]
-            self.meta_charset = 1
+            self.meta_charset = 1 # Remember that the charset was retrieved from
+                                  # META tag, not from the Content-Type header
          except IndexError:
             pass
 
@@ -56,11 +57,25 @@ class HTMLParser(_HTMLParser):
 
    def start_title(self, attrs):
       self.accumulator = ''
+
    def end_title(self):
       if not self.title: # use only the first title
          self.title = self.accumulator
 
 
+import re
+entity_re = re.compile("(&#[0-9]+;)")
+
+def recode_entities(title, charset):
+   output = []
+   for part in entity_re.split(title):
+      if entity_re.match(part):
+         part = unichr(int(part[2:-1])).encode(charset, "replace")
+      output.append(part)
+
+   return ''.join(output)
+
+
 def parse_html(filename, charset=None, log=None):
    infile = open(filename, 'r')
    parser = HTMLParser(charset)
@@ -78,8 +93,9 @@ def parse_html(filename, charset=None, log=None):
    except (HTMLParseError, HTMLHeadDone):
       pass
 
+   title = parser.title
+
    if not parser.charset:
-      title = parser.title
       ascii = 1
       for c in title:
          if not (32 <= ord(c) <= 127): # non-ASCII character
@@ -87,19 +103,21 @@ def parse_html(filename, charset=None, log=None):
             break
       if not ascii:
          parser.charset = DEFAULT_CHARSET
+
    if parser.charset and (parser.charset <> current_charset):
       try:
          if parser.meta_charset:
             if log: log("   META charset   : %s" % parser.charset)
          else:
             if log: log("   charset        : %s" % parser.charset)
-         if log: log("   title          : %s" % parser.title)
-         parser.title = unicode(parser.title, parser.charset, "replace").encode(current_charset, "replace")
+         if log: log("   title          : %s" % title)
+         title = unicode(title, parser.charset, "replace").encode(current_charset, "replace")
          if log: log("   current charset: %s" % current_charset)
-         if log: log("   converted title: %s" % parser.title)
+         if log: log("   converted title: %s" % title)
       except LookupError:
          if log: log("   unknown charset: `%s' or `%s'" % (parser.charset, current_charset))
 
+   parser.title = recode_entities(title, current_charset)
    return parser