[jbsoum]

http://code.google.com/p/jmm/

Hello all. I currently use Mediaportal for my HTPC software, and I was wondering if it was possible for XBMC to support it's functionality. AFAIK, XBMC doesn't rely on plugins to parse episode titles and add them to the database, this is a central function. The above link is to a plugin for Mediaportal that also has external functionality outside of Mediaportal. The plugin uses file hashes to add episodes to its own database (either sqlite or mysql), via information available on anidb.net. I have always found this to be a much more efficient and reliable way to add anime to the database, since the file names tend to confuse most regex expressions.

Is there any possibility of a plugin like this being functional in XBMC?

[Bstrdsmkr]

A scraper could be made, but I can't see where they're referencing the hashes against. I didn't see them available on anidb

[jbsoum]

Here's an example:

Let's take Fullmetal Alchemist. Here is a link to the page on anidb.net for that series:

http://anidb.net/perl-bin/animedb.pl?show=anime&aid=979

If you go down on the page, you'll see a section called 'Group Status.' Here is listed every ripping/fansubbing/muxing group that has released this series, and their status of completion, as well as the available audio/subtitle languages from that group. Under the 'State' column, you'll see the status the group has achieved on the project. Examples are 'finished,' complete,' 'stalled,' etc. Click on the 'state' for a group (Let's use AHQ as an example), and that will take you to this page:

http://anidb.net/perl-bin/animedb.pl?sho...08&aid=979

Here, a CRC hash is listed for each file, as well as other attributes of the file. JMM (the name of the anime plugin I adore), compares your file hashes to the hashes on anidb and adds anime to your library in this fashion. Obviously, this couldn't be done for regular TV shows or movies, since there is not a database online that provides all that information for each online release. However, since naming conventions are pretty standard for most TV shows (except for shows that are labeled with dates instead of SxxExx, which I'm sure XBMC can handle since MP-TVSeries for Mediaportal has been able to do that for awhile), I don't think this is a necessary function for other kinds of media.

However, naming conventions for anime are very different, and I have found that these titles confuse regex expressions, since the CRC is often part of the filename. Also, the names of folders could be either in Romanji or in English, and the TVDB may have one name or the other, or have different season placement for specials, and it becomes an arduous process getting anime added to your library.

JMM does not care about folder name or file name. As long as the file is ANYWHERE in your import folder, JMM will hash the file and add it to your library appropriately. You could literally name the subfolder Poop, and name the file in the folder poop.mkv, and JMM would know what anime series it's from and what season/episode number it is, as well as other useful information specific to that file.

The plugin for Mediaportal is part of the JMM project, and JMM itself acts as an independent program that many anime watchers use just to manage their collections. There are too many features to list.

The point is that JMM is the only reason I continue to use Mediaportal. If XBMC supported the functionality of this plugin, I would switch to XBMC in a heartbeat. I believe it would be required that JMM used a database separate from XBMC, but if there's another way for this software to be supported by XBMC, I would love it.

[HenryFord]

Okay - me thinks that should be possible with a scraper polling the local JMM-Server to get information about specific shows. The scraper could parse the hash of a file and send it to the Server which then in turn would answer the request based on the hash.
You'd only need to find out on which basis they are hashing the files, but that should be easy to find out since the program is open-source.

However:
Wouldn't it probably more feasible to shoot up a renamer-program which relies on the information it gets from the JMM-Server? I mean - there is a scraper for anidb (isn't there?). So, what the program should do, is getting the hash for a file, asking JMM what the hash stands for and then rename the file to the given result. This would in turn enable the existing scraper to work with the filename, you have your collection nice and clean looking and you wouldn't have to think up a scraper that is platform dependend (technically - you could probably talk to the server within Linux for example, but the server-component is still windows only.

[jbsoum]

JMM does have this functionality. There is a drop folder that organizes your anime into appropriate folders, and recently, there has been file renamer functionality added.

However, everyone doesn't use it. The main reason for not wanting to use it is that renaming files makes it more difficult to seed these files to private trackers. Either you have to relocate every file in your torrent software (a royal pain), make copies of each file (waste of space), or just accept that you won't be seeding your files. Granted, this isn't important to all JMM users, but I'd imagine it is important to a great deal of users.

[HenryFord]

Mhm... okay, didn't think of that.
There is however another method, if the torrent-program supports it: Junction Points (the NTFS-equivalent of SymLinks).
Would it be possible to link to those files with a symbolic link and the torrent-program would still be able to pick it up correctly?
But that would probably require even more work.

Since I'm really not great with Python I can only try. Plus: My time is limited, so don't expect anything from me
But if you find someone with some anime love and python skills, you could probably bug him about building a JMM-scraper.