[Gamester17]

Anyone wanting to do long term enhancements to the database should look at how Meedio (and now MeediOS) works. Would be a lot of work to implement, but it still uses SQLite and is far more flexible.

For reference:
http://www.meedios.com
http://sourceforge.net/projects/meedios/

http://www.meedios.com/wiki/index.php?ti...management

MeediOS Database Core

The MeediOS database engine is one of the defining characteristics among all other Media Center software. With that said, most of the documentation is outdated or stuck in my head. So to rectify that, this page hosts a growing collection of articles that explains just about everything users and developers would want to know.

[djdafreund]

Yeah, that used to be cool. I used it way back since it was originally MyHTPC, but like a lot of other software (GBPVR) once they got bought out, they usually end up going down. THat's why i'm glad XBMC is open-sourced and staying that way. That's what's so great about it. And keeps getting better.

[tvdb-scott]

Yeah, that used to be cool. I used it way back since it was originally MyHTPC, but like a lot of other software (GBPVR) once they got bought out, they usually end up going down. THat's why i'm glad XBMC is open-sourced and staying that way. That's what's so great about it. And keeps getting better.

They're rewriting Meedio as MeediOS, which is open source and designed to be completely modular. The library (database) system in Meedio/MeediOS is definitely its best feature, but also the ability for the menus to be completely customizable with very little effort.

For anyone wishing to take this on, a quick explanation of how the Meedio library works:
- Each library is set up as a separate SQLite database. Adding "tags" is basically the same as adding a database field, except that it can be done by developers, plugins, or end users. This means anything can easily have a library as long as someone makes an importer (scraper) and display module (plugin/script) for it: emulators, trailers, sports scores, movie times, etc.

- Each library can have importers designed specifically for it. In many cases, someone will make both an importer (this translates to an XBMC scraper) and module (this is an XBMC plugin or script) since using the same fields can be important (scraping/displaying TV information, for example). However, there are other importers that allow you to modify the data in a table as well. For example, say you need to apply a search/replace regular expression on your imported TV data for some reason... you can just set up your importers so the first is the TV scraper and the second is the regex "importer".

- Each library can be given views, which allow you to group and sort by any field in the database. These views also allow you to define the default view mode for that step (each view can be multiple steps). This means that an end-user could set up a view that would first show Genres, then widescreen banners, then Seasons, then episodes. Or perhaps sort by year, then genre, then actors, etc. 100% customizable.

- On the front-end, the modules (plugins/scripts) are then given one setting: which library to use. By using a standard media module, you're able to browse a library using the views/steps explained above.

- All data fields are varchar (except id), but can be casted to sort as text/date/int.

I know MeediOS has been considering implementing the XBMC scraper format for one of their importer types, so it might be worthwhile for someone to work closely with them so both projects benefit.

http://www.meedios.com/forum/viewtopic.php?t=2238

[spiff]

please discuss medios in an appropriate thread kthx.

this is my first take at tvshow nfo support. it has its limitation, this is what you get from me.

1) tvshow nfo file is called tvshow.nfo - this and only this file will be used for a tvshow's info.

this file can either hold an url like for movies, or complete xml info.
if it holds complete info, there are two modes of operation; either you have an episode guide url in there, meaning additional online lookups is possible, or you dont. if you do not, only episodes with xml files can be added.

2) each episode has a separate xml - this can only hold info, no url overrides here.

3) IMPORTANT. the episode file names still needs to enumerate properly. this is NOT a way around that. i will NOT change the code to support random file names as it means large changes to the structure of things, something i am NOT willing to do since i have absolutely no need for these features myself.

as i'm certain i have introduced issues, i wont apply it to trunk just yet. and i'm not in the mood for extensive testing, so please help me out.

the diff is available Here
and a win32 directx xbmc_pc exe file is available Here.

diff is created against trunk but should apply to the branch (possibly with a bit larger fuzz setting)

as these are hosted on my dsl, please do not grab unless you plan to help out.
in particular i want help on making sure there's no regressions for movies or mvids (in addition to the new features working as expected ofc). this means both manual and automated scans, full info or nfo urls (not supported for mvids atm for some reason). there are so many corner cases, so please be thorough and state exactly what you have tested. note that i have only updated the tvdb scraper.

if noone responds, the diff will most likely go to /dev/null or atleast get stuck in my local tree.

cheers

[agathorn]

please discuss medios in an appropriate thread kthx.

this is my first take at tvshow nfo support. it has its limitation, this is what you get from me.

1) tvshow nfo file is called tvshow.nfo - this and only this file will be used for a tvshow's info.

this file can either hold an url like for movies, or complete xml info.
if it holds complete info, there are two modes of operation; either you have an episode guide url in there, meaning additional online lookups is possible, or you dont. if you do not, only episodes with xml files can be added.

2) each episode has a separate xml - this can only hold info, no url overrides here.

3) IMPORTANT. the episode file names still needs to enumerate properly. this is NOT a way around that. i will NOT change the code to support random file names as it means large changes to the structure of things, something i am NOT willing to do since i have absolutely no need for these features myself.

as i'm certain i have introduced issues, i wont apply it to trunk just yet. and i'm not in the mood for extensive testing, so please help me out.

the diff is available Here
and a win32 directx xbmc_pc exe file is available Here.

diff is created against trunk but should apply to the branch (possibly with a bit larger fuzz setting)

as these are hosted on my dsl, please do not grab unless you plan to help out.
in particular i want help on making sure there's no regressions for movies or mvids (in addition to the new features working as expected ofc). this means both manual and automated scans, full info or nfo urls (not supported for mvids atm for some reason). there are so many corner cases, so please be thorough and state exactly what you have tested. note that i have only updated the tvdb scraper.

if noone responds, the diff will most likely go to /dev/null or atleast get stuck in my local tree.

cheers

This is very good news spiff, and such nfos will also make local fanart support much better, as the colors can be stored in the nfo.

[spiff]

oh, and one more thing.

it would be wise to NOT criticize my db work if you want to stay on my good side. it seriously pisses me off. i had, and still have, absolutely NO experience with db design or sql. i called out for help, but nobody stepped up so i fumbled my way around it to the best of my abilities. it's not that i cant take critique, but i will only accept it in the language of a diff. unless you supply a diff you are a bunch of hypocritical whiners, bashing the work of the only person that actually stepped up and added what all of you wanted for years.

you want something better? step up. if not, please stfu.

[billyad2000]

This sounds great, i've dl it and will test it out over the next few days.

Just to get this clear, tvshow.nfo is the only name that can be used. I assume that this is placed in the root directory of a given TV show. And each single tvshow.nfo file is to contain all the xml metadata for all seasons of a given show?

btw - I didn't mean to criticise your hard work. I didn't even realise that you were the sole developer behind the XBMC library. IMHO the library features are what make XBMC much more than just another media centre. An excellent job overall. It's just too easy to take people for granted.

[spiff]

correct on the nfo placement part.

however, not on the info. tvshow.nfo only holds the TVSHOW's info.
each episode needs a separate nfo file. there is no season info as such in the database

and don't get me wrong, i realize that you did mean no offense. it's just so frustrating getting told 'this could be done so much better' from everybody with experience, yet nobody is willing to help.

[Gamester17]

I apologise as well, I did not mean my comments as criticism, I am sorry if it sounded that way spiff

[billyad2000]

correct on the nfo placement part.

however, not on the info. tvshow.nfo only holds the TVSHOW's info.
each episode needs a separate nfo file. there is no season info as such in the database

and don't get me wrong, i realize that you did mean no offense. it's just so frustrating getting told 'this could be done so much better' from everybody with experience, yet nobody is willing to help.

OK I've got it now, sounds like a good system.

[NineT9mustang]

I'll be testing this when I can, a bit confused, but thats simply because im a complete noob at everything in xbmc...

Looking forward to learning though, like I said before in this thread, i'll be making a bunch of NFO's for episodes/shows, so ill be sure to share the nfo's with everyone if anyone cares

[djdafreund]

Yeah, that's not a bad idea about people posting the files they already have made. I find this definitely usefull since i also collect a lot of Tv shows, and already put movie info files to good use for times when i have to redo things, and don't want to wait a while to re-grab all info.
THanks a bunch Spiff, your work is always respectfully appreciated!!! Also thanks for being the one to take this on and make it happen.

[Nick8888]

There might even be a use for a standalone "thetvdb -> .nfo" windows program if someone was bored. Which would allow people to manually add tv show info to un-networked xboxs via cd/dvds.

Thanks spiff for adding this feature, I know it is something you didn't code for your own personal use so it is very much appreciated.

[djdafreund]

Hopefully Billyad2000 will add it to his current nfo exporter program. I just asked him in his thread. Would be VERY nice indeed!!!

[spiff]

i certainly hope so as i added this out of respect for his work.

and thanks for your kind words guys