Image Status Checker / Dead Image Finder

Note this hack works with vb3.6


What does this do?
It scans all your posts, extracts all the img tags, and scans each of the images to see if they're still valid.


Why?
I had a look at all the images on my site and was alarmed at how many were now gone - deleted from photobucket accounts etc. Since the only way you can check the images on your board is to manually read every post, I decided to come up with a better way... and this is it.


How does it work?
The first part: In the AdminCP, under Maintenance and Update Counters... right at the bottom is this hack. It works by looking up every img tag, then requesting the image, and reading the http status code. So code 200 means 'image ok', 404/410 means 'image gone' etc. That then gets stored in a database table. A server has 15 seconds to reply to the request or the status is labelled as "Unknown"
The second part: The browsing element, imagestatuscheck.php (original filename huh!). This allows you to browse all the images found in the last scan using some powerful filtering (statuses to display, search, order by).


Hack features

• General
• Fully phrased.
• Templates are grouped. Who's online handled.
• Part 1 - Admin
• Reads the post table, scans all the [img] tags on demand and records the actual http status code returned.
• If it gets stuck during the scan, you can restart the section it's currently doing.
• If an image appears in more than one post, it's only checked once.
• Start from, per page and timeout options for scanning.
• Part 2 - Browser
• Status codes are put into one of three descriptions for simplicity: Working, Dead, Unknown. Unknown is if the server didn't respond or similar - on the basis that a temporary timeout doesn't necessarily mean the image has gone.
• In the browser, image urls are force wrapped. Unless people post using all caps, you have a low screen resolution, or the font size is big, the table should never stretch.
• Filtering allows you to show just the working/dead/unknown images, and there's a search facility for a variety of fields.
• Convenient link to edit the post (if a dead link is found). This works by can_moderate - edit links only appear for people who own the post, or can moderate the forum it's in.
• Works by canview - if someone can't view a particular forum (e.g. staff forum) normally, they can't view the images within it.
• Uses css for common stuff to reduce the size of the outputted pages.

Bad Things
It's far from a perfect hack, there are many things to do. Please be aware that I won't be doing them, but if anyone else wants a crack, feel free!

• Only supports http://, not https://
• Can only handle replies like: HTTP 1.x 200 as the first line.
• Only supports [img] tags. If you have HTML turned on in any forums it won't see <img src=> images.
• Biggie: There's no way to update a single post or image without a full re-scan. That means if someone edits their post to update or remove a dead link, it will not change on the browser until a full re-scan is done. I did play with various update methods but most are flawed in one way or another. A planned feature will be to update the table dynamically whenever a post is made, edited or deleted, and on demand using a link.
• No cron job.
• No session variables. (People without cookies will be logged out a lot).

Footnotes
Originally I planned to throw something together quickly just for me to use but it turned into a "I may as well make a nice interface... oh and I may as well put some filtering controls in and I ..."


A [url] link checker can be found here


Installation
Upload imagestatuscheck.php to your vB directory. Install the product, set overwrite to yes.


Customizing

• By default it's set to only allow moderators, super-moderators and administrators to view the browser. This can be changed with the setting in AdminCP > vB Options.
• The phrases all start with ics_ if you want to change them.
• You can add a link to imagestatuschecker.php on the navbar (or anywhere) if you want your members to be able to view it.

Screenies
Shot 1 is AdminCP during scan
Shot 2 is a typical Browser section output
Shot 3 is no results output


Changelog
See attached file for specific changes.
1.00 - 16th August 06
1.01 - 17th August 06
1.02 - 27th December 06

[Gryphon]

There was also another issue, I edited my post.

[rmxs]

Thanks installed

[rmxs]

OK i try it it works byt i get many worning links with Unknown status

Y this happents?

Can you tell me how can i add it also to navbar for moder admin smoder groups only??
i mean if there is no 5,6,7 dont show the link

EDIT:

Ok i made it its easy LOL

<if condition="$bbuserinfo[usergroupid] == 6">

<td class="vbmenu_control"><a href="imagestatuscheck.php">DIF</a></td>

</if>

[bairy]

Chad : Each image is requested one by one and only the first 12 characters of the return are read, as they are the ones with the status code in them. After that the connection is closed. Theoretically it will use output about 200 bytes and input 12 bytes per request. Practically I don't know how web servers work, but I suspect once php has sent a close to the other server the transfer will stop. So no, not much bandwidth

[ForYou]

Hello ,

there is error ,

Database error in vBulletin 3.6.0:

Invalid SQL:
INSERT INTO imagestatus VALUES (NULL, 172959, 3498, 'http://www.dohaeye.com/lyrics/3'air%20elnass.jpg', '');

MySQL Error : You have an error in your SQL syntax; check the manual that corresponds to your MySQL server version for the right syntax to use near 'air%20elnass.jpg', '')' at line 1
Error Number : 1064
Date : Wednesday, August 16th 2006 @ 09:21:55 PM
Script : http://www.dha.net/moda/admincp/misc...tus&cis=findem
Referrer : http://www.dha.net/moda/admincp/misc...k_image_status
IP Address : 213.6.1.100
Username : Admin
Classname : vb_database

[bairy]

The sql errors are because the image url isn't escaped. Silly oversight. I'll probably get an updated product out tomorrow along with a couple of other changes.

[EasyTarget]

on a related note, what about scanning all posts for img tags, rehosting all remote images at a local location and editting the posts with the new url so that you don't have to worry about the images going away?

[bairy]

Originally Posted by rmxs

OK i try it it works byt i get many worning links with Unknown status

Y this happents?

Sorry. Missed this question.
Unknown generally means one of two things:
1. Server didn't reply within 15 seconds
2. Server didn't send a nice http header back

Originally Posted by EasyTarget

on a related note, what about scanning all posts for img tags, rehosting all remote images at a local location and editting the posts with the new url so that you don't have to worry about the images going away?

Copyright issues. The idea is quite good though. I might include a way to allow you to manually do that.

[rmxs]

Thanks bairy

[Mr Chad]

Originally Posted by bairy

Chad : Each image is requested one by one and only the first 12 characters of the return are read, as they are the ones with the status code in them. After that the connection is closed. Theoretically it will use output about 200 bytes and input 12 bytes per request. Practically I don't know how web servers work, but I suspect once php has sent a close to the other server the transfer will stop. So no, not much bandwidth

ahh thanks that clears it up, good job coding it.