Retrieving text from messy strings, rgding -regexp

[kenneal]

I have some html code over here which i would like to extract the .jpg from.
Sometimes it generates as many different forms

E.g.

<img src='http://123.com/12345.jpg' border=0>
<code='http://123.com/12345.jpg'>
<a code=1 file='http://123.com/12345.jpg' form=value >

Is there any definite way to just 100% get the http://123.com/12345.jpg out of the always randomed + garbled html tag? I do understand regexp only thing that since it could appear in many forms i'm not very sure what should be the best way.

[rosc2112]

Based on your examples:

Code: Select all

regexp {<.*?'(.*?\.jpg).*?>} $htmlInput fullstringmatchvar exactmatchvar

Or:

regexp {<.*?'(http://.*?\.jpg).*?>} etc

I would've thought html would use double quotes " tho, not '
Change the above example to use " if nec.

exactmatchvar is where your http.*jpg string will be stored.

[kenneal]

regexp {<.*?'(.*?\.jpg).*?>} $htmlInput fullstringmatchvar exactmatchvar

I tried messing around with it works fine. I still have a few questions.
I want to get lets say a certain string that could be called

www.abc.com/link.php?ref=123zxc

or

www.abc.com/link.php?ref=34zxz

or

www.abc.com/link.php?ref=SDASDASDASd

They are embedded in codes like,

<img src='www.abc.com/link.php?ref=SDASDASDASd' border=0>
<code='www.abc.com/link.php?ref=SDASDASDASd'>
<a code=1 file='www.abc.com/link.php?ref=SDASDASDASd' form=value >

1) How could I go about retrieving them since the last part is always random?
2) What if in the string there is also another www.abc.com/link.php?ref=ZXCZXC , regexp will only take out 1, is there another way I can take out the other?

I tried messing around with regexp but it still does not seem to work..

Do appreciate your help here

[user]

Code: Select all

{='([^']+)'}

To get all the matches, I'd use the -all and -inline options (check the regexp manual page) and a foreach loop.
BTW: your examples are not html.

[kenneal]

Code: Select all

if [string match *.jpg* $line] {
         foreach jpg [regexp -all -inline {<.*?'(http://.*?\.jpg).*?>} $line] {
                   puthelp "PRIVMSG #jtest :$jpg"		
        }
}	

It returns me oddly the before and after regexped results.

[00:29:06] <@PALL> <br><br><img src='http://hotimg2.fotki.com/b/222_88/97_98/25000.jpg ' border=0 onclick="window.open('http://www2.lookpipe.com/get.php?filepa ... /25000.jpg ');" onload='if(this.width>

[00:29:08] <@PALL> http://hotimg2.fotki.com/b/222_88/97_98/25000.jpg

[00:29:10] <@PALL> <br><br><img src='http://hotimg1.fotki.com/a/222_88/97_98/3353.jpg ' border=0 onclick="window.open('http://www2.lookpipe.com/get.php?filepa ... 8/3353.jpg ');" onload='if(this.width>

[00:29:12] <@PALL> http://hotimg1.fotki.com/a/222_88/97_98/3353.jpg


I'm quite sure this results procs from the foreach line. If I removed the PRIVMSG line nothing comes out.

Edit:
I seem to get the idea that the -inline thing parses out info as such, so should I be making a statement where 1st line dont parse, 2nd line parse, 3rd line dont parse, 4th line parse, 5th line dont parse, 6 line parse.. etc so on? Just wanna confirm.

Also user, yes it is results parsed out from a php forum so its not entirely html... I tried regexp {='([^']+)'} but am still unable to parse out the link.php from the garbled text as I have with the .jpgs...

[user]

You still get the entire match first (like you did without -inline), so add another variable in the loop to weed it out. Like this:

Code: Select all

foreach {garbage jpg} [regexp -all -inline {='(http[^']+)'} $data] {
	puthelp "PRIVMSG #chan :$jpg"
}

If you want to capture those relative urls starting with "www" too, replace "http" in the above rule with "(?:www|http)"

[kenneal]

im testing it and everything seems good, thanks a lot!

[kenneal]

Thx!