VB Regex help

Author	Message

TheQuestor

Total Posts : 5
Scores: 0
Reward points : 0
Joined: 1/18/2011
Status: offline

VB Regex help Tuesday, January 18, 2011 1:20 AM (permalink)

0

I am currently trying to figure out to filter out all a hrefs [and whatever other html i need later] from XML that I have converted to standard html and stored in a database. I can use the following:

Function RemoveHTML( strText )
 Dim RegEx
 Set RegEx = New RegExp
 RegEx.Pattern = "<[^>]*>"
 RegEx.Global = True
 RemoveHTML = RegEx.Replace(strText, "")
End Function

to strip out ALL html but I would like to keep the image tags in tact but I can not using the above code. Can somebody give me a good regex expression that can remove just say the following:
<a href="someurl">Some words here</a>
and better yet the above and also something like:
<div style="blah;blah;">Leave the words alone </div>
I guess above I would just need to remove <div style="blah;blah;"> and then </div> after all div calls are removed.
Anyway before I get even more confused maybe just start with the a href one :)

HELP!
r

#1

TNO Total Posts : 2094 Scores: 36 Reward points : 0 Joined: 12/18/2004 Location: Earth Status: offline	Re:VB Regex help Saturday, January 29, 2011 9:48 AM (permalink) 0 Probably easier to use the DOM (XML/HTML) and the innerText property of the elements in question to get the contents you need. Don't parse XML/HTML with regular expressions: http://stackoverflow.com/...d-tags/1732454#1732454 *To iterate is human, to recurse divine.* -- L. Peter Deutsch
	#2

mcnd Total Posts : 50 Scores: 2 Reward points : 0 Joined: 4/27/2008 Status: offline	Re:VB Regex help Wednesday, March 23, 2011 12:28 AM (permalink) 0 RegEx.Pattern = "</?(?!img)[^>]*>"
	#3

TNO

Total Posts : 2094
Scores: 36
Reward points : 0
Joined: 12/18/2004
Location: Earth
Status: offline

Re:VB Regex help Wednesday, March 23, 2011 8:22 AM (permalink)

0

mcnd

RegEx.Pattern = "</?(?!img)[^>]*>"

That fails to capture things like:

<a href="http://example.com/hello<world>/littlejohnny.html"><img id="foo" src="foo.png"/></a>

Writing the html to CreateObject("htmlfile") followed by .getElementsByTagName("img") is probably wiser.

To iterate is human, to recurse divine. -- L. Peter Deutsch

#4

mcnd

Total Posts : 50
Scores: 2
Reward points : 0
Joined: 4/27/2008
Status: offline

Re:VB Regex help Wednesday, March 23, 2011 9:23 PM (permalink)

0

Too busy to read RFC 3986 or W3C about the characters allowed.
But you are right. You can find anything in an HTML page.
But since it is easier to do it the hard way, it's probably faster to do a little previous cleanup to handle those cases.
What about

<\/?(?!img)\w(\s+(?:\w+\s*=\s*\x22[^\x22]*\x22|\w+\s*=\s*[^ ]+|\w+)){0,}[^>]*>

? I know it's not perfect, but it's amusing

#5

TNO

Total Posts : 2094
Scores: 36
Reward points : 0
Joined: 12/18/2004
Location: Earth
Status: offline

Re:VB Regex help Thursday, March 24, 2011 8:33 AM (permalink)

0

mcnd

Too busy to read RFC 3986 or W3C about the characters allowed.

It is not just about the url. Plus on page 50 of the rfc you can see the regex to capture a valid url.

mcnd

But you are right. You can find anything in an HTML page.

I'm not trying to show willfully broken pages if that's what you mean. That url from before is legitimate AFAICT.

mcnd

But since it is easier to do it the hard way,

I'm not sure what that means

mcnd

it's probably faster to do a little previous cleanup to handle those cases.
What about
<\/?(?!img)\w(\s+(?:\w+\s*=\s*\x22[^\x22]*\x22|\w+\s*=\s*[^ ]+|\w+)){0,}[^>]*>
? I know it's not perfect, but it's amusing

Nope, (X)HTML is not a regular language, so a regular expression won't help you on its own short of Perl/.NET extensions (which makes them irregular expressions I guess).

 <!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Strict//EN"
 "http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd">
<html xmlns="http://www.w3.org/1999/xhtml" xml:lang="en">
     <head>
         <title>foo</title>
         <meta http-equiv="Content-Type" content="text/html; charset=UTF-8" />
     </head>
     <body>
         <!-- I like turtles -->
         <img src="example.png"/>
         <![CDATA[<img src="example.png"/>]]>
     </body>
</html>

This is one way which will work:

 Dim Document : Set Document = CreateObject("htmlfile") Document.Write htmlString Dim Images : Set Images = Document.getElementsByTagName("img") Dim Image For Each Image In Images     WScript.Echo Image.OuterHtml Next

Another way is using CreateObject("MSXML2.DomDocument.3.0") to load the file as XML (assuming it is valid) and using xpath. Its about the same amount of code

To iterate is human, to recurse divine. -- L. Peter Deutsch

#6

mcnd

Total Posts : 50
Scores: 2
Reward points : 0
Joined: 4/27/2008
Status: offline

Re:VB Regex help Friday, March 25, 2011 12:00 AM (permalink)

0

TNO

I'm not trying to show willfully broken pages if that's what you mean. That url from before is legitimate AFAICT.

I know. Not what i was trying to mean. Just talking about (X)HTML, not your answer.

TNO

mcnd

But since it is easier to do it the hard way,

I'm not sure what that means

Just a not so good translation of an spanish proverb : "es más facil hacerlo de la forma más dificil"

TNO

Nope, (X)HTML is not a regular language, so a regular expression won't help you on its own short of Perl/.NET extensions (which makes them irregular expressions I guess).
....

As said before, a little previous cleanup (remove anything not needed)

TNO

This is one way which will work:
 Dim Document : Set Document = CreateObject("htmlfile") Document.Write htmlString Dim Images : Set Images = Document.getElementsByTagName("img") Dim Image For Each Image In Images     WScript.Echo Image.OuterHtml Next 
 
Another way is using CreateObject("MSXML2.DomDocument.3.0") to load the file as XML (assuming it is valid) and using xpath. Its about the same amount of code

Sure, this retrieves all the image tags, but

TheQuestor

... strip out ALL html but I would like to keep the image tags in tact ...

Regular expressions ONLY does not solve the problem. Maybe i'm wrong, but using htmlfile or DomDocument seems less efficient.
BUT
well, the original post says