Moved parse_html.py and its submodules to a separate parse_html module.

[bookmarks_db.git] / parse_html / __init__.py

diff --git a/parse_html/__init__.py b/parse_html/__init__.py
new file mode 100644 (file)
index 0000000..c67f2e7
--- /dev/null
+++ b/parse_html/__init__.py
@@ -0,0 +1,179 @@
+"""
+   HTML Parsers wrapper
+
+   Written by Broytman. Copyright (C) 1997-2011 PhiloSoft Design
+"""
+
+import codecs
+
+universal_charset = "utf-8"
+DEFAULT_CHARSET = "cp1251" # Stupid default for Russian Cyrillic
+
+parsers = []
+
+try:
+   from . import beautifulsoup
+except ImportError:
+   pass
+else:
+   beautifulsoup.DEFAULT_CHARSET = DEFAULT_CHARSET
+   parsers.append(beautifulsoup.parse_html)
+
+try:
+   from .lxml import parse_html
+except ImportError:
+   pass
+else:
+    parsers.append(parse_html)
+
+try:
+   from .htmlparser import parse_html
+except ImportError:
+   pass
+else:
+    parsers.append(parse_html)
+
+try:
+   from . import html5
+except ImportError:
+   pass
+else:
+   parsers.append(html5.parse_html)
+
+# ElementTidy often segfaults
+#try:
+#   from . import etreetidy
+#except ImportError:
+#   pass
+#else:
+#   parsers.append(etreetidy.parse_html)
+
+import re
+from htmlentitydefs import name2codepoint
+
+entity_re = re.compile("(&\w+;)")
+num_entity_re = re.compile("(&#[0-9]+;)")
+
+def recode_entities(title, charset):
+   output = []
+   for part in entity_re.split(title):
+      if part not in ("&", "<", ">", """) and \
+            entity_re.match(part):
+         _part = name2codepoint.get(part[1:-1], None)
+         if _part is not None:
+             part = unichr(_part).encode(charset)
+      output.append(part)
+   title = ''.join(output)
+
+   output = []
+   for part in num_entity_re.split(title):
+      if num_entity_re.match(part):
+         try:
+            part = unichr(int(part[2:-1])).encode(charset)
+         except UnicodeEncodeError:
+            pass # Leave the entity as is
+      output.append(part)
+
+   return ''.join(output)
+
+
+def parse_html(filename, charset=None, log=None):
+   if not parsers:
+       return None
+
+   if charset:
+      try:
+         codecs.lookup(charset) # In case of unknown charset...
+      except (ValueError, LookupError):
+         charset = None         # ...try charset from HTML
+
+   charsets = [universal_charset, DEFAULT_CHARSET]
+   if charset:
+      charset = charset.lower().replace("windows-", "cp")
+      if charset in charsets:
+         charsets.remove(charset)
+      charsets.insert(0, charset)
+
+   for p in parsers:
+      parser = None
+      for c in charsets:
+         try:
+            parser = p(filename, c, log)
+         except UnicodeEncodeError:
+            pass
+         else:
+            break
+      if parser:
+         break
+      else:
+         if log: log("Parser %s.%s failed, trying next one." % (p.__module__, p.__name__))
+
+   if not parser:
+       return None
+
+   converted_title = title = parser.title
+   if title and (not parser.charset):
+      try:
+         unicode(title, "ascii")
+      except UnicodeDecodeError:
+         parser.charset = DEFAULT_CHARSET
+
+   if parser.charset:
+      parser.charset = parser.charset.lower().replace("windows-", "cp")
+
+   if title and parser.charset and (
+         (parser.charset <> universal_charset) or
+         ((not charset) or (charset <> parser.charset))):
+      try:
+         if parser.meta_charset:
+            if log: log("   META charset   : %s" % parser.charset)
+         elif (not charset) or (charset <> parser.charset):
+            if log: log("   guessed charset: %s" % parser.charset)
+         if log: log("   current charset: %s" % universal_charset)
+         if log: log("   title          : %s" % title)
+         if parser.charset <> universal_charset:
+            try:
+               converted_title = unicode(title, parser.charset).encode(universal_charset)
+            except UnicodeError:
+               if log: log("   incorrect conversion from %s, converting from %s" % (parser.charset, DEFAULT_CHARSET))
+               converted_title = unicode(title, DEFAULT_CHARSET, "replace").encode(universal_charset, "replace")
+               parser.charset = DEFAULT_CHARSET
+         if log and (converted_title <> title): log("   converted title: %s" % converted_title)
+      except LookupError:
+         if log: log("   unknown charset: '%s'" % parser.charset)
+   else:
+      if log: log("   title          : %s" % title)
+
+   if title:
+      final_title = recode_entities(converted_title, universal_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
+
+   icon = parser.icon
+   if isinstance(icon, unicode):
+       try:
+           parser.icon = icon.encode('ascii')
+       except UnicodeEncodeError:
+           if parser.charset:
+               parser.icon = icon.encode(parser.charset)
+   return parser
+
+
+def test():
+   import sys
+
+   l = len(sys.argv)
+   if l == 3:
+      filename = sys.argv[1]
+      charset = sys.argv[2]
+   elif l == 2:
+      filename = sys.argv[1]
+      charset = universal_charset
+   else:
+      sys.exit("Usage: %s filename [charset]" % sys.argv[0])
+
+   parser = parse_html(filename, charset, log=lambda s: sys.stdout.write(s + '\n'))
+   print "   refresh:", parser.refresh
+   print "   icon   :", parser.icon