PHP: Create your own url shortener with this easy to use PHP Class

After using Twitter some time now, i saw several URL shortening services and a few scripts that would do the work. But i did not see any PHP Class for doing this job easily. So i wrote one myself ...

The most mentioned code example for this application floating around on Twitter is this one written by jaisen. Its a good start but had some things i would do different.

Goals

Let me start with what i wanted to achieve:
- Write a simple to use PHP Class for saving, shortening and redirecting URLs.
- URLs should be stored in a MySQL database.
- Number of URLs should not be limited too early.
- No duplicate URL entry's in database.
- A simple hit counter for each URL.
- Short URL should be case-insensitive.

Source

Link to the source: http://juliusbeckmann.de/classes/src/url_shortener/url_shortener.zip
And direct source view: http://juliusbeckmann.de/classes/src/url_shortener/

Documentation

Here you can find the documentation: http://juliusbeckmann.de/classes/

Demo

A live demo of my class: http://juliusbeckmann.de/url_shortener/

Process of developing

Some people think that a URL shortener somehow shortens a long URL like "http://google.com" to something like "b7ic3". There is no compression done, the URL will be stored in a database with the unique key "b7ic3". This is the whole magic.

Main problem of this concept, how do we generate the key "b7ic3"? I saw some functions that can do that but i decided to use the PHP build in function base_convert(). I use this function to convert a number like 1234 to its base 36 (a..z and 0..9) equivalent "ya". One downside of this function is, that only lowercase chars will be generated which will result in a lower "compress ratio". But this is wanted because some programs like the original ICQ Chat client sometimes handles case-sensitive URLs incorrect. It is also easier to spell case-insensitive URL on the phone.
The principle with the short URL key works like this:
We insert a new URL to our database and use mysql_insert_id() to get the next id given away from AUTO INCREMENT. This decimal number will be converted to the equivalent base 36 one, and viola - we have our unique key. I know this method is not the best if old URLs get deleted, but it will work always and there will be no duplicate keys.

I build in a special security feature that is often forgotten or even unknown.

This is important. If an attacker would be able to inject newlines to our url, he could change the "Location: "-Header and create a second request answer.

Options

If you want to use a different base, just configure this parameter:

Possible values are from 2 up to 36.

I also build in a URL validation. This is to prevent that URLs other than http or https are shortened. To change this validation configure this parameter:

If you want to define bad words or even whole URLs, use this parameter. Each URL that contains a word from this list will not be shortened:

Methods

The whole class has only 5 public methods.

This function creates a new short URL. If the URL is already in the database no new entry will be made. This function will return a array with all relevant data inside, or an empty array on error.

This function will give you the data for a specific key. The return is the same data like new_redirect().

If you want to log redirect hits, just call this method and the hit counter for this key will be +1.

If you want to know how much hits a key already got, use this method.

Simply returns the number of URLs in our database.

Table structure

Here is the table structure my class uses:

The id column will be used to identify a URL. The principle is easy. If the key is "ya", the URL with the id 123 is asked. There is also a md5 column in which the md5 of the URL will be stored. This is done for faster searching on duplicate URLs and to validate if a URL is still correct.
Some additional data like timestamp and ip are also stored. ip is stored in a shorter format.

If you want to log the number of hits, this is the used table structure:

It is important to say that we need a second table for this because otherwise we would write instantly to the table MySQL needs to read from. This would lead to serious performance troubles.

I inserted about 100.000 URLs, each with at least one hit. The whole database needed about 17MB of space, which are ~180 Byte per URL. This is a good value if you think how much information is saved.

Last Words

Es you can see, the class has all functions needed for a simple and secure URL shortener. You can find a short example index.php with htaccess in the source dir. If you find mistakes or errors, please write a comment.

No related posts.

Geschrieben am Donnerstag, 10. September 2009 und abgelegt unter PHP. Verfolgen Sie die Diskussion zu diesem Beitrag per RSS 2.0 Feed. Die Kommentarfunktion wurde deaktiviert. Auch das "Anpingen" des Beitrages ist nicht möglich.