Convert urls from text to links even if no protocol

Question

Lets say that $content is the content of a textarea

/*Convert the http/https to link */
     $content = preg_replace('!((https://|http://)+[a-z0-9_./?=&-]+)!i', '<a target="_blank" href="$1">$1</a> ', nl2br($_POST['helpcontent'])." ");
/*Convert the www. to link prepending http://*/
     $content = preg_replace('!((www\.)+[a-z0-9_./?=&-]+)!i', '<a target="_blank" href="http://$1">$1</a> ', $content." ");

This was working ok for links, but realised that it was breaking the markup when an image is within the text...

I am trying like this now:

$content = preg_replace('!\s((https?://|http://)+[a-z0-9_./?=&-]+)!i', ' <a href="$1">$1</a> ', nl2br($_POST['content'])." ");
$content = preg_replace('!((www\.)+[a-z0-9_./?=&-]+)!i', '<a target="_blank" href="http://$1">$1</a> ', $content." ");

As is the images are respected, but the problem is that url's with http:// or https:// format won't be converted now..:

google.com -> Not converted (as expected)

www.google.com -> Well Converted

http://google.com -> Not converted (unexpected)

https://google.com -> Not converted (unexpected)

What am I missing?

-EDIT-

Current almost working solution:

$content = preg_replace('!(\s|^)((https?://)+[a-z0-9_./?=&-]+)!i', ' <a href="$2" target="_blank">$2</a> ', nl2br($_POST['content'])." ");
$content = preg_replace('!(\s|^)((www\.)+[a-z0-9_./?=&-]+)!i', '<a target="_blank" href="http://$2"  target="_blank">$2</a> ', $content." ");

The thing here is that if this is the input:

www.funcook.com http://www.funcook.com https://www.funcook.com funcook.com http://funcook.com https://funcook.com

All the urls I want (all, except name.domain) are converted as expected, but this is the output

www.funcook.com http://www.funcook.com https://www.funcook.com ; funcook.com http://funcook.com https://funcook.com

Note an ; is inserted, any idea why?

Maybe this question helps: http://stackoverflow.com/questions/161738/what-is-the-best-regular-expression-to-check-if-a-string-is-a-valid-url — Paulo Coghi, Jul 28 '12 at 22:30
You could look http://daringfireball.net/2010/07/improved_regex_for_matching_urls and http://mathiasbynens.be/demo/url-regex — Alexandre Khoury, Jul 29 '12 at 15:26

Brian Glaz · Answer 1 · 2012-07-31T01:01:52.520

8

try this:

preg_replace('!(\s|^)((https?://|www\.)+[a-z0-9_./?=&-]+)!i', ' <a href="$2">$2</a> ',$text);

It will pick up links beginning with http:// or with www.

Example

edited Jul 31 '12 at 01:01

answered Jul 17 '12 at 19:32

Brian Glaz

15,468
4
37
55

Hi there, was working fine but now i need more cases.. Also I offered a Bounty ;) – Toni Michel Caubet Jul 23 '12 at 14:21
in the very beginning, just add `s?` after http. This makes the `s` optional so it will work with both http and https. You don't need to use two preg replace like you have in your question. – Brian Glaz Jul 23 '12 at 14:45
I see, but i need to preg_replaces, right? because in the case of www.web.com i have to add the http:// (do you see any alternative for that second expresion?) – Toni Michel Caubet Jul 23 '12 at 19:07
hmm yeah, I guess you would need two... i haven't been able to come up with anything yet. – Brian Glaz Jul 23 '12 at 20:42
Hello again, I just realised that this would try to convert images that may be in the text! thats a real problem; I know it's not part of the question, but could you give me a hint? thanks – Toni Michel Caubet Jul 25 '12 at 20:50
Correct :( so i guess we could add ( i mean with your help) that the url cant be bettwen < and > ?? – Toni Michel Caubet Jul 25 '12 at 21:03
I love you bro', will test and let you know. Thanks – Toni Michel Caubet Jul 26 '12 at 00:59
Hello Again, I was able to test; but it doesn't seem to convert the urls (but it respects the images).. none of the 3 url cases are converted. Edited the cuestion with current code (mix of your last answer and the one to convert the www. ones and prepend the http://), Any idea? – Toni Michel Caubet Jul 27 '12 at 12:20
@ToniMichelCaubet are you trying this at the start of the string? ie "http://www.blah.com is a website" - this won't work as the above expression expects a \s (space) to start with the start of the code above should be changed from: !\s to !(\s|^) which will allow them at the start – williamvicary Jul 30 '12 at 13:21
BTW: Seems to detect them, but creates an empty link (with empty url and text) – Toni Michel Caubet Jul 30 '12 at 21:33
Funny just to test I tried $2 and it works now :) – Toni Michel Caubet Jul 30 '12 at 21:43
you're right, it should be $2 because i modified it with @williamvicary 's comment and added another capturing group – Brian Glaz Jul 31 '12 at 01:03
Aha my bad, if you don't want it to match you can do this: (?:\s|^) – williamvicary Jul 31 '12 at 08:12
Any idea why they cant be the first item on the text? they need an empty space right before – Toni Michel Caubet Aug 01 '12 at 12:56
Do you have `(\s|^)` at the beginning? this means it will match a space OR the beginning of the text. – Brian Glaz Aug 01 '12 at 13:56

Madara's Ghost · Answer 2 · 2012-07-24T11:03:08.730

2

You can't at 100%. Becuase there may be links such as stackoverflow.com which do not have www..

If you're only targeting those links:

!(www\.\S+)!i

Should work well enough for you.

EDIT: As for your newest question, as to why http links don't get converted but https do, Your first pattern only searches for https://, or http://. which isn't the case. Simplify it by replacing:

(https://|http://\.)

With

(https?://)

Which will make the s optional.

edited Jul 24 '12 at 11:03

answered Jul 17 '12 at 19:30

Madara's Ghost

172,118
50
264
308

For www-less urls couldn't you just use `#\b\S+\.\S{3}\b#`? – Bailey Parker Jul 23 '12 at 23:45
1

@PhpMyCoder: "hi.my name is truth!my english isn't perfect!" – Madara's Ghost Jul 24 '12 at 11:00
Urls without the `www.` can be matched with `#\b\S+\.\S{3}\b#`. This doesn't account for tlds like `.co.uk`, but it will cover your basics: `.com`, `.org`, `.net`, etc. – Bailey Parker Jul 25 '12 at 02:57
Read my comment. It will match even though there's no url there. Also, you'll match `..........` – Madara's Ghost Jul 25 '12 at 13:02
Check with Regexr. Unless I'm copying something wrong neither of those match for me. – Bailey Parker Jul 26 '12 at 01:12
@DanLee Point taken. Then I'd modify to explicitly cover the popular TLDs + http(s) + subdomains: `#\b(https?://)?([a-z0-9-]+\.)+(com|org|net|edu)\b#i` – Bailey Parker Jul 29 '12 at 22:28

score 1 · Answer 3 · answered Jul 17 '12 at 19:38

Another method to go about adding hyperlinks is that you could take the text that you want to parse for links, and explode it into an array. Then loop through it using foreach (very fast function - http://www.phpbench.com/) and change anything that starts with http://, or https://, or www., or ends with .com/.org/etc into a link.

I'm thinking maybe something like this:

$userTextArray = explode(" ",$userText);
foreach( $userTextArray as &$word){
    //if statements to test if if it starts with www. or ends with .com or whatever else
    //change $word so that it is a link
}

Your changes will be reflected in the array since you had the "&" before $userText in your foreach statement. Now just implode the array back into a string and you're good to go.

This made sense in my head... But I'm not 100% sure that this is what you're looking for

score 1 · Answer 4 · answered Jul 24 '12 at 04:12

I had similar problem. Here is function which helped me. Maybe it will fit your needs to:

function clHost($Address) { 
   $parseUrl = parse_url(trim($Address)); 
   return  str_replace ("www.","",trim(trim($parseUrl[host] ? $parseUrl[host].$parseUrl[path] :  $parseUrl[path]),'/'));
}

This function will return domain without protocol and "www", so you can add them yourself later.

For example:

$url = "http://www.". clHost($link);

I did it like that, because I couldn't find good regexp.

But that would be applied to the direct url right, i need to scan for urls within a text. Thanks — Toni Michel Caubet, Aug 01 '12 at 13:47

score 0 · Answer 5 · answered Jul 30 '12 at 13:37

\s((https?://|www.)+[a-z0-9_./?=&-]+)

The problem is that your starting \s is forcing the match to start with a space, so, if you don't have that starting space your match fails. The reg exp is fine (without the \s), but to avoid replacing the images you need to add something to avoid matching them.

If the images are pure html use this: (?<!src=")((https?://|www.)+[a-z0-9_./?=&-]+)

That will look for src=" before the url, to ignore it.

If you use another mark up, tell me and I'll try to find another way to avoid the images.

Hi tried that, but it will convert to empty links with empty text; even if I use $1 or $2.. — Toni Michel Caubet, Aug 02 '12 at 16:30

Convert urls from text to links even if no protocol

5 Answers5

Linked

Related