Change parse_html to parse strings, not files

[bookmarks_db.git] / parse_html / bkmk_ph_html5.py

diff --git a/parse_html/bkmk_ph_html5.py b/parse_html/bkmk_ph_html5.py
index d883bd7e5f4b98dd4307fc8c56390411ac347d3c..6400d02c20b924b0d77e65709c9136ab1647d41e 100644 (file)
--- a/parse_html/bkmk_ph_html5.py
+++ b/parse_html/bkmk_ph_html5.py
@@ -1,13 +1,11 @@
 """HTML Parser using html5
 
 This file is a part of Bookmarks database and Internet robot.
+
 """
 
-__version__ = "$Revision$"[11:-2]
-__revision__ = "$Id$"[5:-2]
-__date__ = "$Date$"[7:-2]
 __author__ = "Oleg Broytman <phd@phdru.name>"
-__copyright__ = "Copyright (C) 2010, 2011 PhiloSoft Design"
+__copyright__ = "Copyright (C) 2010-2014 PhiloSoft Design"
 __license__ = "GNU GPL"
 
 __all__ = ['parse_html']
@@ -17,12 +15,9 @@ from html5lib import HTMLParser as HTML5Parser
 from .bkmk_ph_util import HTMLParser
 
 
-def parse_html(filename, charset=None, log=None):
+def parse_html(html_text, charset=None, log=None):
     parser = HTML5Parser()
-    fp = open(filename)
-    parser._parse(fp, encoding=charset, parseMeta=bool(charset))
-    fp.close()
-    html_tree = parser.tree.getDocument()
+    html_tree = parser.parse(html_text, encoding=charset, parseMeta=bool(charset))
 
     for node in html_tree.childNodes:
         if (node.name == 'html') and (node.type != 3): # Skip DocType element
@@ -55,11 +50,8 @@ def parse_html(filename, charset=None, log=None):
                 else:
                     title = ''
 
-        if title is None:
-            return None
-
         for node in head.childNodes:
-            if node.name == 'meta' and \
+            if (node.name == 'meta') and \
                     ('http-equiv' in node.attributes) and \
                     (node.attributes['http-equiv'] == 'content-type'):
                 meta_content = node.attributes['content']
@@ -71,11 +63,14 @@ def parse_html(filename, charset=None, log=None):
                         meta_charset = False
                     else:
                         break
+            elif (node.name == 'meta') and ('charset' in node.attributes):
+                meta_charset = node.attributes['charset'].lower()
+                break
 
         if not charset:
             charset = parser.tokenizer.stream.charEncoding[0]
 
-        if charset or meta_charset:
+        if title and (charset or meta_charset):
             title = title.encode(charset or meta_charset)
 
         for node in head.childNodes:
@@ -101,7 +96,6 @@ def parse_html(filename, charset=None, log=None):
                 else:
                     title = ''
 
-        if title is None:
-            return None
-
+    if (title is None) and (refresh is None) and (icon is None):
+        return None
     return HTMLParser(charset, meta_charset, title, refresh, icon)