Regex c# remove white space but not inside <pre> and </pre>

[Seabizkit]

Hi all

 

I am trying to write some regex to remove all white space from html.

 

but the regex i am currently using doesn't factor in that in the pre tags there maybe opening "<" and closing ">" tags

 

This matches everything inside the pre tags

 

(<)\s*?(pre\b[^>]*?)(>)([\s\S]*?)(<)\s*(/\s*?pre\s*?)(>)

 

e.g.

(009)156 (010) (0)<pre> <test> edehofo<w<dieoj >  ></pre>     yuui u    ji 

will match 

<pre> <test> edehofo<w<dieoj >  ></pre>

 

and 

(?<=\s)\s+(?![^<>]*</pre>)

 

eg will almost work but does not work if there is an "<" or ">" in the mark up.

 

space[     ]spaces <pre>[          ]spaces</pre>space[      ]spaces 

 

will result in 

space[ ]spaces <pre>[ ]spaces</pre>space[ ]spaces 

 

but if there is a "<" or ">" in the pre tags then it will not work.

 

Could anyone help me

 

 

 

 

 

 

 

 

 

 

[Lant]

A warning: http://stackoverflow.com/questions/1732348/regex-match-open-tags-except-xhtml-self-contained-tags

 

A solution: http://stackoverflow.com/questions/8762993/remove-white-space-from-entire-html-but-inside-pre-with-regular-expressions (same as your regex)

 

Just be careful with regex and HTML, parsing it to a C# object and then outputting it without whitespace is probably a better way to do this (likely to be slower though).

 

You shouldn't include < or > in <pre>, they should be escaped to &lt; and &gt;

[Seabizkit]

Thanks, but I've already been to those sites and a lot of others trying to find an solution.

 

From that its highlighted that I maybe able to do it in to step then...

 

1. encode all html with in pre tags

2. then use (?<=\s)\s+(?![^<>]*</pre>) to remove white space.

 

I am finding it hard to now find regex which will encode html with in a pre tag.

I love cats and dogs but dogs > than cats

<spaces>       <spaces>
<pre>
    1 > 0  = ?

    <code>       dedw</code>
</pre>

should be 

I love cats and dogs but > than cats <spaces> <spaces>
<pre>
    1  &gt; 0 = ?
    &glt;code&gt;       dedw &gt;/code &gt;
</pre>

It would be easier if i had some regex which basically ignored everything between a given tag.

 

I am doing my own research and trying to find the answer... but thought i would clarify what I am trying to do.

 

Thanks if you can help me find a solution

[Eric Veteran]

What language is this for? You should be able to walk through the elements and strip everything that's not a <pre> element.

[Seabizkit]

Hi Eric

 

I'm trying to do it for c#

 

Basically I can get what i want if I can strip out white-space but not between two sets of chars

 

e.g.  <pre> anything </pre> or <code> anything </code>

 

seems more challenging that i thought it would be.

 

 (?<=\s)\s+(?![^<>]*</pre>)  is awesome but if i could get it to basically ignore everything inside of a given tag

 

My understanding of how to get the right syntax for regex is limited.

 

Match everything but not between tag(pre or code) and remove white space.

 

any help with getting this to work would be so awesome.

[Lant]

How about the following: http://snipd.net/parsing-xhtml-into-a-dom-tree-in-c

You avoid using regex as you simply parse then print the HTML instead.

 

 

You might also want to re-read my first link, which points out that you cannot parse arbitrary (X)HTML using a regex. It is not possible as regex is not powerful enough to parse (X)HTML. (Parsing a limited subset is possible)

[Seabizkit]

to anyone who is good with regex please help

 

(?<=\s)\s+(?![^[<p]+?[e>])

 

cant not get this to work but must be close?

 

to: Lant

 

Sorry but i don't need to parse the HTML... i am not trying to validate it or read it really... 

 

I am wanting to remove white space but only when outside of a given tag like say <pre> or <code>

 

It has been working fine with (?<=\s)\s+(?![^<>]*</pre>)  until i hit a new requirement which.... is that the <pre> tag can contain < or >

 

which this regex does not support. One suggestion was to replace < and > with their HTML versions &lt; and &gt;

 

but again i would need to have something which only applied this on html with in tag the <pre> tag and I would need to do it in c#

 

seems easier to just get the regex to ignore things between tags some how...

 

Thanks

[Squirrelington]

If you don't mind doing the replace in C#, you can use capture groups with this pattern.

(?:\<pre\>)(.*)(?:\<\/pre\>)

It should capture everything between the pre tags and the value of the capture can then be string.Replaced on. I am working on a Regex replace version so you can use RegexOptions.Compiled. This regex was tested with CaseInsensitive and SingleLine (despite it being multi-line, using that to adjust how it handles .*)

 

 

 

so for example:

 

private static Regex preBlockMatch = new Regex(@"(?:\<pre\>)(.*)(?:\<\/pre\>)", RegexOptions.Compiled | RegexOptions.Singleline | RegexOptions.IgnoreCase);

 

(code block)

string newValue = preBlockMatch.Captures[0].Value.Replace("<", "&lt;");

(code block)

 

obviously room for some enhancement but it gets you started

[Seabizkit]

Thought I would share what I got working......

 

Also thanks!!! to Squirrelington for the regex for the between stuff

 public class MinifiedStream : MemoryStream
        {
            private readonly Stream _output;
            public MinifiedStream(Stream stream)
            {
                _output = stream;
            }

            private static readonly Regex Whitespace = new Regex(@"(?<=\s)\s+(?![^<pre>]*</pre>)",RegexOptions.Compiled);

            private static readonly Regex PreBlockMatch = new Regex(@"(?:\<pre\>)(.*)(?:\<\/pre\>)", RegexOptions.Compiled | RegexOptions.Singleline | RegexOptions.IgnoreCase);

            Dictionary<string, string> result = new Dictionary<string, string>();

            public override void Write(byte[] buffer, int offset, int count)
            {
                var html = Encoding.UTF8.GetString(buffer);

                //------------------------------------
                var matches = PreBlockMatch.Matches(html);
                int loopcount = 1;
                foreach (Match match in matches)
                {
                    var token = Guid.NewGuid().ToString() + "_"+ loopcount;
                    html = html.Replace(match.Value, token);
                    result.Add(token, match.Value);
                    loopcount ++;
                }
                //------------------------------------



                html = Whitespace.Replace(html, string.Empty);
                html = html.Trim();

                //-----------------------------------
                foreach (var match in result)
                {
                    html = html.Replace(match.Key, match.Value);
                }

                //-----------------------------------

                _output.Write(Encoding.UTF8.GetBytes(html), offset, Encoding.UTF8.GetByteCount(html));
            }
        }

But still feels as if you should be able to create a straight regex version.

 

if anyone can help me improve that would be awesome

[Squirrelington]

I agree that there should be a way to get it exclusively in regex but I am only really good at matching things, I don't have a lot of experience in replacement syntax. I know you can do $1 for the first group, $2 for 2nd or ${name} for named groups but how to modify the contents of those groups in the replacement is beyond me atm. Maybe someone with more experience will be able to chime in. Glad it is working thus far though. \o/

 

btw the regex I gave you, I probably went a little too crazy with the non-capture groups. I think at the time I was trying some funky magic to try and capture parts of it for replacement purposes but ended up with that.

<pre>(.*)</pre>

would probably be fine as well. lol

[saad90]

s = Regex.Replace(s, @"\s+", string.Empty);

works 100%