X-Git-Url: https://git.phdru.name/?a=blobdiff_plain;f=Robots%2Fparse_html.py;h=9f924c49597a3f4436fd477363ee7cd1fe1b5c32;hb=7ebd86416b89bf67bb97f7680b87660f8cdc0cf7;hp=c6a72f62fcff0f4987357830372b38aae03dd17a;hpb=a6dd530445b282c35407f40bf83160bb6cf8dfb3;p=bookmarks_db.git

diff --git a/Robots/parse_html.py b/Robots/parse_html.py
index c6a72f6..9f924c4 100755
--- a/Robots/parse_html.py
+++ b/Robots/parse_html.py
@@ -1,6 +1,6 @@
 #! /usr/bin/env python
 """
-   HTML Parser
+   HTML Parsers wrapper
 
    Written by BroytMann. Copyright (C) 1997-2007 PhiloSoft Design
 """
@@ -11,73 +11,16 @@ from m_lib.defenc import default_encoding
 current_charset = default_encoding.replace("windows-", "cp")
 DEFAULT_CHARSET = "cp1251" # Stupid default for Russian Cyrillic
 
-from HTMLParser import HTMLParseError
-from m_lib.net.www.html import HTMLParser as _HTMLParser
+parsers = []
+try:
+   from parse_html_beautifulsoup import parse_html
+except ImportError:
+   pass
+else:
+   parsers.append(parse_html)
 
-
-class HTMLHeadDone(Exception): pass
-
-
-class HTMLParser(_HTMLParser):
-   def __init__(self, charset=None):
-      _HTMLParser.__init__(self)
-      self.charset = charset
-      self.meta_charset = 0
-      self.title = ''
-      self.refresh = ''
-      self.icon = None
-
-   def end_head(self):
-      raise HTMLHeadDone()
-
-
-   def do_meta(self, attrs):
-      http_equiv = ""
-      content = ""
-
-      for attrname, value in attrs:
-         if value:
-            value = value.strip()
-            if attrname == 'http-equiv':
-               http_equiv = value.lower()
-            elif attrname == 'content':
-               content = value
-
-      if (not self.charset) and (http_equiv == "content-type"):
-         try:
-            # extract charset from "text/html; foo; charset=UTF-8; bar;"
-            self.charset = content.lower().split('charset=')[1].split(';')[0]
-            self.meta_charset = 1 # Remember that the charset was retrieved from
-                                  # META tag, not from the Content-Type header
-         except IndexError:
-            pass
-
-      if http_equiv == "refresh":
-         self.refresh = content
-
-
-   def start_title(self, attrs):
-      self.accumulator = ''
-
-   def end_title(self):
-      if not self.title: # use only the first title
-         self.title = self.accumulator
-
-
-   def do_link(self, attrs):
-      has_icon = False
-      href = None
-
-      for attrname, value in attrs:
-         if value:
-            value = value.strip().lower()
-            if (attrname == 'rel') and (value.lower() in ('icon', 'shortcut icon')):
-               has_icon = True
-            elif attrname == 'href':
-               href = value
-
-      if has_icon:
-         self.icon = href
+from parse_html_htmlparser import parse_html
+parsers.append(parse_html)
 
 
 import re
@@ -87,7 +30,10 @@ 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")
+         try:
+            part = unichr(int(part[2:-1])).encode(charset)
+         except UnicodeEncodeError:
+            pass # Leave the entity as is
       output.append(part)
 
    return ''.join(output)
@@ -100,24 +46,14 @@ def parse_html(filename, charset=None, log=None):
       except (ValueError, LookupError):
          charset = None         # ...try charset from HTML
 
-   infile = open(filename, 'r')
-   parser = HTMLParser(charset)
-
-   for line in infile:
-      try:
-         parser.feed(line)
-      except (HTMLParseError, HTMLHeadDone):
+   for p in parsers:
+      parser = p(filename, charset)
+      if parser:
          break
+      else:
+         if log: log("Parser %s.%s failed, trying next one." % (p.__module__, p.__name__))
 
-   infile.close()
-
-   try:
-      parser.close()
-   except (HTMLParseError, HTMLHeadDone):
-      pass
-
-   title = parser.title
-
+   converted_title = title = parser.title
    if not parser.charset:
       try:
          unicode(title, "ascii")
@@ -132,20 +68,37 @@ def parse_html(filename, charset=None, log=None):
          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" % title)
-         title = unicode(title, parser.charset, "replace").encode(current_charset, "replace")
+            if log: log("   HTTP charset   : %s" % parser.charset)
          if log: log("   current charset: %s" % current_charset)
-         if log: log("   converted title: %s" % title)
+         if log: log("   title          : %s" % title)
+         save_title = title
+         try:
+            converted_title = unicode(title, parser.charset).encode(current_charset)
+         except UnicodeError:
+            if parser.meta_charset and parser.charset.endswith("1252") and \
+                  not DEFAULT_CHARSET.endswith("1252") and (DEFAULT_CHARSET <> current_charset):
+               parser.charset = DEFAULT_CHARSET
+               if log: log("   incorrect conversion from cp1252, converting from %s" % DEFAULT_CHARSET)
+               converted_title = unicode(save_title, DEFAULT_CHARSET, "replace").encode(current_charset, "replace")
+            else:
+               converted_title = unicode(title, parser.charset, "replace").encode(current_charset, "replace")
+         if log and (converted_title <> title): log("   converted title: %s" % converted_title)
       except LookupError:
          if log: log("   unknown charset: `%s' or `%s'" % (parser.charset, current_charset))
-
-   parser.title = recode_entities(title, current_charset)
+   else:
+      if log: log("   title          : %s" % title)
+
+   final_title = recode_entities(converted_title, current_charset)
+   parts = [s.strip() for s in final_title.replace('\r', '').split('\n')]
+   final_title = ' '.join([s for s in parts if s])
+   if log and (final_title <> converted_title): log("   final title    : %s" % final_title)
+   parser.title = final_title
    return parser
 
 
 if __name__ == '__main__':
-   parser = parse_html(sys.argv[1])
+   import sys
+   parser = parse_html(sys.argv[1], current_charset)
    print parser.charset
    print parser.title
    print parser.refresh