find and replace in large formated text

Lcousins

hi all,

i am looking to move all images for my site from one location to another. I can change all the ones that make up the static site easily but am looking for a way to achieve it with all the mentions in the dynamic content that is called up from a mysql database.

say for example i had $maincontent which is a 'text' entry containing a lot of html formatting and I want to replace all referance to 'http://mysite1/images/' with 'http://mysite2/image_directory/'
how would I do it and what would I have to watch out for so as not to mess up the html code?

thanks for any help
lcousins

Shrike

$text = preg_replace("/http:\/\/mysite1\/images\//si", "http://mysite2\/image_directory", $text);

If you need more conditions, a more complex regexp would be needed.

Shrike

$text = preg_replace("/http:\\/\\/mysite1\\/images\\//si", "http://mysite2\\/image_directory/", $text);

If you need more conditions, a more complex regexp would be needed.

Lcousins

thanks, I had actually considered trying this with almost the identical solution as you suggested just without the '//si' bit. could you please explain to me what that bit does??

thanks

Luke

Weedpacket

Originally posted by Lcousins
thanks, I had actually considered trying this with almost the identical solution as you suggested just without the '//si' bit. could you please explain to me what that bit does??

http://www.php.net/manual/en/pcre.pattern.modifiers.php

Shrike

Ah beat me to it.

Hmm, 'It was like that when I found it'....which book is that from ? 🙂

Pop quiz - which GSV/mind was called Meatf*cker? (no Google-ing)

Weedpacket

Originally posted by Shrike
Ah beat me to it.

Hmm, 'It was like that when I found it'....which book is that from ? 🙂

Pop quiz - which GSV/mind was called Meatf*cker? (no Google-ing)

The Grey Area, and if memory serves it was a GCU 🙂

Shrike

Ah possibly. It's a long time since I read them but didn't the Grey Area become GSV Sleeper Service?

rangalo

hi friends,

I have a similar problem to this the only difference is that i dont know the name of the old image files.....i want to create a generall function that changes src attribute of an image tag in a given html file...

can you help me with the regular expression shrike?

Shrike

Something along the lines of:

$htmlstring = "<img src=\"test.gif\" border=\"0\">";
$newname = "test.jpg";
$htmlstring = preg_replace("/(<img.*?src=\")[a-z0-9-_\\.]+(\".*?>)/si", "$1$newname$2", $htmlstring);
echo $htmlstring;

To replace the src attribute of every image tag (not that that's of much use) 😉

bubblenut

<bubble>put's on his sci-fi baddie mask and pretends to be shrike, hoping no one will notice</bubble>

you'll need something along the lines of

preg_replace('/<img(.*?)src=(["\'])(.*?)\/([^\/]*?)\2','<img\1src="http://www.site.com/new/pic/path/\4"',$string);

I haven't tested it yet but here's a rough explination.
<img - match <img (simple enough so far
(.*?) - do a non greedy match of any characters and store the result in the bank so we can replace it in the replace string (could contain things such as width and height attributes)
src= - match src= (ie the begining of the source attribute)
(["\']) - match either a single or a double quote and place it in the bank so we can match the same on the close
(.*?) - do a non greedy match on any character and store the result (this will contain the path of the directory that the file is contained in)
\/ - match a /
([^\/]) match anything other than a / and place it in the bank, this will match the file name.
\2 - match whatever the second thing we placed in the bank was (the opening quote, be it single or double)

HTH
Bubble

Lcousins

bubblenut, thankyou for that, I think it does contain some errors, but since I dont totally understand how this works yet im not sure where they are 😛

really this is a whole new world to me, having spent the night reading the relevant pages in the php manual (thanks weedpacket for the link) I'm more confused than ever.

preg_replace('/<img(.*?)src=(["'])(.*?)\/([^\/]*?)\2','<img\1src="http://www.site.com/new/pic/path/\4"',$string);

ok, what exactly does (.*?) capture. what doesnt it capture (what does it stop at?)

(.*?)\/ how does that capture http://www.adomain.com/randomfile/images bit and how does it know when to stop?

([^\/]*?) how does this identify just the file name?

Weedpacket

Originally posted by Shrike
Ah possibly. It's a long time since I read them but didn't the Grey Area become GSV Sleeper Service?

Nah, you're thinking of the Quietly Confident. You can recall which one is the Grey Area by remembering what its - er - perversion involved.

bubblenut

Well, in order to understand how it matches you will have to do quite a lot of reading (Mastering Regular Expressions is a very good book) but basically the regex will go along untill it finds a <img and then it will chomp up all the characters untill (in this case) the end of the line with the .? (any number of any character and then it will do what's known as 'backtracking' to step backwards through the string untill it finds where it should actually be continuing from. At each step back it will try to go forwards again untill it can't and then it will step back again.
([^\/]?) matches the file name largely because of the (.*?)\/ before it and the \2 after it. A file path is of the form
/home/httpd/maintenance/directory/file.php
therefore all the non backslash characters between the last backslash and the closing quote must make the file name yes? and all the characters between the last backslash and the first quote must make the directory path. This is all the two statements are saying.
HTH
Bubble

Lcousins

thankyou Bubble, I still dont understand it fully but Ive understood enough to get it working! one final question, is there a way I can check to see if anything was changed, eg so if it did not find any instances to change It doesnt continue, coming up with a message??

luke

bubblenut

do a plain preg_match with the same regex first?

Bubblenut

Weedpacket

Originally posted by bubblenut
do a plain preg_match with the same regex first?

Bubblenut

$altered = preg_replace('...',$original);
if($altered==$original)
{ // No real change
}

perhaps? I know I used bubblenut's idea in some code in the past.