Mercurial Repositories > Web Project and Tools / file diff

--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/ProjectWeb/Html5Prettifier.py	Thu Jan 01 17:14:39 2015 +0100
@@ -0,0 +1,77 @@
+# -*- coding: utf-8 -*-
+
+# Copyright (c) 2015 Detlev Offenbach <detlev@die-offenbachs.de>
+#
+
+"""
+Module implementing a class to prettify HTML code.
+"""
+
+from __future__ import unicode_literals
+
+import re
+
+
+from PyQt5.QtCore import QObject
+
+import Preferences
+
+
+class Html5Prettifier(QObject):
+    """
+    Class implementing the HTML5 prettifier.
+    """
+    def __init__(self, html, parent=None):
+        """
+        Constructor
+        
+        @param html HTML text to be prettified (string)
+        @param parent reference to the parent object (QObject)
+        """
+        super(Html5Prettifier, self).__init__(parent)
+        
+        self.__html = html
+        self.__indentWidth = Preferences.getEditor("IndentWidth")
+    
+    def getPrettifiedHtml(self):
+        """
+        Public method to prettify the HTML code.
+        
+        @return prettified HTML code (string)
+        """
+        from bs4 import BeautifulSoup
+        soup = BeautifulSoup(self.__html, "html.parser")
+        prettyHtml = soup.prettify(formatter=self.tagPrettify)
+        # prettify comments
+        prettyHtml = re.sub("<!--(.*?)-->", self.commentPrettify, prettyHtml,
+                            flags=re.DOTALL)
+        # indent all HTML
+        prettyHtml = re.sub("^( *)(.*)$",
+                            r"{0}\2".format(r"\1" * self.__indentWidth),
+                            prettyHtml, flags=re.MULTILINE)
+        
+        return prettyHtml
+    
+    def tagPrettify(self, tag):
+        """
+        Public method to prettify HTML tags.
+        
+        @param tag tag to be prettified (string)
+        @return prettified tag (string)
+        """
+        result = re.sub(" {{1,{0}}}".format(self.__indentWidth), " ", tag,
+                        flags=re.MULTILINE)
+        return result
+    
+    def commentPrettify(self, matchobj):
+        """
+        Public method to prettify HTML comments.
+        
+        @param matchobj reference to the match object (re.MatchObject)
+        @return prettified comment (string)
+        """
+        if re.search("\n", matchobj.group()):
+            result = self.tagPrettify(matchobj.group())
+        else:
+            result = matchobj.group()
+        return result