Importer for Invision's IPB/IPS software

[KeiroD]

This is a feature request for the IPB software to create an importer for.

I think I might be able to get you guys access to the software to try creating an importer for. Let me know what you guys need and I'll see if I can make it happen.

[kode54]

I, too, would like to see this. I have access to a 15 year backlog on IPB 2.x, which dates back to an initial import from vBulletin, I think.

[emanuele45]

Yep, yep, have access to a database is the first and "must have" step in order to start building an importer! If you have a way to give access to the database feel free to drop me an email at emanuele45 at gmail.com

[KeiroD]

Got it.

We'll provide a copy of a 3.x install. We don't have a 2.x install on hand, but I think @kode54 might be able to help you out with that.

[emanuele45]

An IPB 3.4 converter should be already available in the development branch, I used it to convert a forum, so it should work, but improvements can always be applied. ;D

[KeiroD]

@emanuele45 Sounds like you don't have an importer for the 2.x branch, then?

Welp... looks like I need to entice a 2.x license-holder to cough up a copy.

[kode54]

Database link emailed, please tell me where I may securely send the password.

E: Forum software link emailed too, same password.

[kode54]

Here, have a mostly finalized SMF importer SQL file:

https://gist.github.com/kode54/8a296da37c23391d020c

You can check the revision list to compare what I changed versus the original script that was attached on the topic at SMF's forum.

Things this needs added:

• Fix up backslash escapes in forum descriptions
• Fix up attached avatar references for uploaded avatars
  • This needs to be done with care. Basically, any user with an avatar named 'av-' . $id_member . '.' . $ext needs to be processed to include a reference in the attachments table as an avatar belonging to that user. The file names are all that format, and they are located in the IP Board's common uploads folder. They are unlikely to be referenced in the attachments table on IP Board.

[kode54]

Minor comment, I notice the SMF to IP Board conversion is way faster, most likely because it doesn't have to run a complex regex process on every post and private message to undo IP Board's pre-processing to hash tagged HTML, which the SMF converter already spent pretty much all night doing.

[kode54]

Here's another thing I did. It kind of breaks through licensing requirements, and requires a working copy of IP Board 2.3.x to run, but it will kindly use IP Board's own BBcode parser to translate the database entries of the posts table back to plain editor-friendly BBcode, and also translates the newlines to <br> tags, which is what ElkArte wants.

https://gist.github.com/kode54/a6e4117e81c329534966

Next up, I write another one to do the private messages table.

Note, that this script has no mention of any IPS proprietary code, except for function and class names, and script filenames. The init.php path must be specified by hand, but the rest can be generated from its settigs. It assumes your IPS setup is working, and can access the forum database.

[kode54]

Further note on the importer, recent version. It appears that the IP Board 2.3.x script, or at least the portions that I use in my dump script, work on a server running PHP 5.5.x, which means I can dump it from a server I've fed the old database to, and much faster at that.

[emanuele45]

Useful notes and great job @kode54!

[kode54]

Further notes:

• Some of the newer or more recently changed uploaded avatars' filenames in the IP Board database contain "upload:" before the actual filename. This may or may not signal a proper uploaded avatar. A lot of the older ones are missing this indicator.
• Uploaded avatars are in the IP Board uploads folder, named 'av-' . intval($id_member) '.' . $ext. And for some reason, due to errors with random browsers, a lot of the avatars on my existing setup are named .txt, so that may need some GD extension detection and changing/adding an extension.
• I used the IP Board BBcode processor for a conversion pass, mainly because it handles nested tags, and tags with random parameters in random orders, better than the Regex in SMF's importer.

Good luck with that mess I gave you!

[kode54]

New issue, related to my own translator: It needs to remove double-quotes from [url] tags, since those break ElkArte's parser.

[kode54]

Here's a new addition. It seems it was pointless of me to try to convert the date stamps from the original IPS Board tags, as those strings are in local time or something stupid like that. So, the proper regex replacement for quoteo/quotec with the full info is to generate the author and link=msg values, then fill in a zero for the date, and let this script fix it:

https://gist.github.com/kode54/80e3be8698f521109b6f

[kode54]

And here's an attempt to fix up polls, which only supports single question polls, and only supports single choice polls:

https://gist.github.com/kode54/a4ec219b9ce4b1b750d3