[ALL PLATFORMS] Extra REGEX for TV Show Episode matching [Archive]

View Full Version : [ALL PLATFORMS] Extra REGEX for TV Show Episode matching

xexe

2009-05-24, 12:36

In simple terms adding this to your XBMC config will match more TV shows than XBMC will by default.

After several months assisting users via IRC I decided to create a generic set of additional REGEX expressions to catch TV episodes XBMC does not by default (e.g. Topaz etc) for fun. Quite a bit of offline testing has been completed, and whilst I am confident I cannot guarantee this REGEX will not produce some false positives.

The method used in this REGEX differs from XBMC default methodology by extracting the season number from folder names rather than the file name. By doing this we can match stuff that otherwise could never be matched.

Examples:

13.show.dvd.avi only has the ep number but "MyShow/season 2/13.show.dvd.avi" has both.
"MyShow/season 5/4400513.show.dvd.avi has a show name that includes numbers making matching very difficult.

Installation

Adding additional TV episode matching is simply a matter of inserting the code listed later to the a file called advancedsettings.xml.
To locate and understand this XML file read the first part of this link http://xbmc.org/wiki/?title=AdvancedSettings.xml#List_of_available_sett ings
Remember by default advancedsettings.xml will NOT exist. Also note the name of this file IS CASE SENSITIVE and will require a XBMC restart to be applied.
End to end installation should take no more than 2 minutes.

Required Folder Structure

To make use of this REGEX you need to have a sensible folder structure for your TV shows as follows:

/showname/season x/episodes e.g. The Unit/season 2/the.unit.203.avi

Note: Case is irrelevant
Note: This is "Season 1" and NOT "Season 01"

If you do not have this structure most of these REGEX's will NOT work for you.

I have had some requests to support different structures. Whilst I am happy to accommodate some slight differences I cannot support multiple languages or weird ass structures. In the end trying to support this would make the REGEX ridiculously complicated for the majority of users whilst only helping a minority.

The chosen structure was decided on after months of seeing what users had developed on their own. Most came to this structure independently and I am happy with it.

Feedback

You are welcome to experiment with this REGEX and report back if you need help or have some suggestions.

I will maintain this first post for at least a while should it become popular. I will happily add new REGEX under a few small conditions:

1. The format you are trying to match allows for a good chance of no false positives i.e. its not my intention to try and deal with absolute rubbish naming dredged from the bowels of the internet.
2. The format you are trying to match will be useful to other people. If the matches will only ever be of use to you alone then this set is not the place for it.
3. If you are suggesting REGEX please supply a couple of examples of the full path you are matching against so I/we can test it.
4. I wont be adding REGEX or updating the existing ones with über complex/1337 REGEX just because it can be done more cleverly. Or put another way we need to keep these simple so normal users can get to grips with them.

Support

If you wish help debugging problems please do the following:

Update to the very latest SVN version. This is the version i stick with and will be testing against.
Post a COMPLETE debug log to pastebin and link it here (make sure this log catches the update library procedure, look for lines containing "Failed to Enumerate"). If you don't upload this log I cannot help and wont even bother. Sorry but time is to short to help those that wont help themselves :)

Have fun, and remember to refer back to this OP should this thread take off.

Additional Comments

Note: The test file names are pulled randomly from Google. Better lists would probably help if anyone can supply them.
Note: This is a re-post of the original thread which became read only in the Tips and Tricks forum.
Note: The forum cant handle all these words in one post so pease see my replies later for new version
Please, this thread is for discussions on THIS REGEX compilation only. It is not for random REGEX support, how do i setup advancedsettings, why doesn't my library work or anything else. For all other topics create a new thread.

spartacus51

2009-06-04, 05:44

You're awesome. Thank you!

The only thing you didn't get in my library is Real Time with Bill Maher. Anyone else want a date based regex?

T:\Real Time with Bill Maher\2009\Real.Time.With.Bill.Maher.2009.05.29.HD TV.avi

Daily Show, Colbert, and Conan (and well, Leno, until this week) are also EZTV releases and use the same naming convention.

Again, and really just this, THANK YOU!

xexe

2009-06-04, 11:32

spartacus51 I didn't know if this could be done so i consulted the powers to be and it looks like r20559 has added support directly to XBMC for Americanised dates.

You can see it:

http://xbmc.org/trac/changeset/20559
http://xbmc.org/trac/ticket/5143

As XBMC handles it natively I will not need to here. All you need to do is update beyond this build number and rescan.

xexe

2009-06-06, 12:58

Version 0.6

This version marks the 99% completion point on every TV ep that i managed to find on Google that wasn't matched with XBMC defaults (something like 10,000+ matches).

One word of warning. The last REGEX in the TV section is very loose and needs to be to catch the last common bad naming scheme. I recommend you use that one with caution and only have it at the very end of the match list.

With this release I also include some Movie stacking REGEX to match some names Ive seen reported as missed via IRC.

Due to time constraints this has not been tested with XBMC directly so I would appreciate any reports.

Edit: managed to do some quick testing and so far performing as expected.

<advancedsettings>
<tvshowmatching append="yes">






<regexp>[.a-z](\d\d?)(\d\d)-notv([^/\\]*)</regexp>




<regexp>[/\\]season[^\w]?(\d{1,2})[^\d]*[/\\]tpz-.*?\1(\d\d)([^/\\]*)</regexp>


<regexp>[/\\]season[^\w]?(\d{1,2})[^\d]*[/\\](\d\d?)[^\w]([^/\\]*)</regexp>


<regexp>[/\\]season[^\w]?(\d{1,2})[^\d]*[/\\].*?[^\w]ep?\.?(\d\d?)[^\w]([^/\\]*)</regexp>


<regexp>[/\\]season[^\w]?(\d{1,2})[^\d]*[/\\].*?[^\w]episode[^\w]?(\d\d?)[^\w]([^/\\]*)</regexp>


<regexp>[/\\]season[^\w]?(\d{1,2})[^\d]*[/\\].*?[^\w]part[^\w]?(\d\d?)[^\w]([^/\\]*)</regexp>


<regexp>[/\\]season[^\w]?(\d{1,2})[^\d]*[/\\].*?[^\w]chapter[^\w]?(\d\d?)[^\w]([^/\\]*)</regexp>



<regexp>[/\\]season[^\w]?(\d{1,2})[^\d]*[/\\].*?\1(\d\d)(?!.*])</regexp>


<regexp>[/\\]season[^\w]?(\d{1,2})[^\d]*[/\\].*?\1[^\w]?x[^\w]?(\d\d?)([^/\\]*)</regexp>


<regexp>[/\\]season[^\w]?(\d{1,2})[^\d]*[/\\].*?s0?\1[ex]{0,2}(\d{1,2})([^/\\]*)</regexp>


<regexp>[/\\]season[^\w]?(\d{1,2})[^\d]*[/\\][^\d]+[\. _-](\d{1,2})[\. _-][^\d]+</regexp>
</tvshowmatching>

<moviestacking append="yes">



<regexp>done[ _\.-](\d)[ _\.-]</regexp>
</moviestacking>

</advancedsettings>

xexe

2009-06-10, 13:44

Version 0.8

I have hit the forum word limit on the OP so from now on I will post one version per post.

This new version is a big change. In theory it should handle most of TPZ releases WITHOUT needing a season folder.

It also adds a config that should in theory stop movie extras being added to the library as well.

It is released as experimental awaiting user feedback.

<advancedsettings>
<tvshowmatching append="yes">






<regexp>[.a-z](\d\d?)(\d\d)-notv([^/\\]*)</regexp>





<regexp>[/\\]tpz-(?:24|30rock|4400)(\d)(\d{2})(\d{2})?(?:r|fix|dc|-repack|int|d)?\.</regexp>


<regexp>[/\\]tpz-[^\d]*(\d)(\d{2})(\d{2})?(?:r|fix|dc|-repack|int|d)?\.</regexp>






<regexp>[/\\]season[^\w]?(\d{1,2})[^\d]*[/\\](\d\d?)[^\w]([^/\\]*)</regexp>


<regexp>[/\\]season[^\w]?(\d{1,2})[^\d]*[/\\].*?[^\w]ep?\.?(\d\d?)[^\w]([^/\\]*)</regexp>


<regexp>[/\\]season[^\w]?(\d{1,2})[^\d]*[/\\].*?[^\w]episode[^\w]?(\d\d?)[^\w]([^/\\]*)</regexp>


<regexp>[/\\]season[^\w]?(\d{1,2})[^\d]*[/\\].*?[^\w]part[^\w]?(\d\d?)[^\w]([^/\\]*)</regexp>


<regexp>[/\\]season[^\w]?(\d{1,2})[^\d]*[/\\].*?[^\w]chapter[^\w]?(\d\d?)[^\w]([^/\\]*)</regexp>



<regexp>[/\\]season[^\w]?(\d{1,2})[^\d]*[/\\].*?\1(\d\d)(?!.*])</regexp>


<regexp>[/\\]season[^\w]?(\d{1,2})[^\d]*[/\\].*?\1[^\w]?x[^\w]?(\d\d?)([^/\\]*)</regexp>


<regexp>[/\\]season[^\w]?(\d{1,2})[^\d]*[/\\].*?s0?\1[ex]{0,2}(\d{1,2})([^/\\]*)</regexp>


<regexp>[/\\]season[^\w]?(\d{1,2})[^\d]*[/\\][^\d]+[\. _-](\d{1,2})[\. _-][^\d]+</regexp>
</tvshowmatching>

<moviestacking append="yes">



<regexp>done[ _\.-](\d)[ _\.-]</regexp>
</moviestacking>

<video>
<excludefromscan append="yes">
<regexp>[/\\]extras[/\\](?!season)</regexp>
</excludefromscan>
</video>

</advancedsettings>

xexe

2009-06-28, 14:00

Version 0.9

Default XBMC REGEX matching some well known files as season 247 incorrectly. To deal with this we now have both prepend and append REGEX. Use with care I have only lightly tested this.

Edit: I ahve done some more testing and it is working as expected.



<advancedsettings>
<tvshowmatching action="prepend">



<regexp>[/\\]tpz-(?:24|30rock|4400)(\d)(\d{2})(\d{2})?(?:r|fix|dc|-repack|int|d)?\.</regexp>


<regexp>[/\\]tpz-[^\d]*(\d)(\d{2})(\d{2})?(?:r|fix|dc|-repack|int|d)?\.</regexp>

</tvshowmatching>

<tvshowmatching action="append">




<regexp>[.a-z](\d\d?)(\d\d)-notv([^/\\]*)</regexp>





<regexp>[/\\]tpz-(?:24|30rock|4400)(\d)(\d{2})(\d{2})?(?:r|fix|dc|-repack|int|d)?\.</regexp>


<regexp>[/\\]tpz-[^\d]*(\d)(\d{2})(\d{2})?(?:r|fix|dc|-repack|int|d)?\.</regexp>






<regexp>[/\\]season[^\w]?(\d{1,2})[^\d]*[/\\](\d\d?)[^\w]([^/\\]*)</regexp>


<regexp>[/\\]season[^\w]?(\d{1,2})[^\d]*[/\\].*?[^\w]ep?\.?(\d\d?)[^\w]([^/\\]*)</regexp>


<regexp>[/\\]season[^\w]?(\d{1,2})[^\d]*[/\\].*?[^\w]episode[^\w]?(\d\d?)[^\w]([^/\\]*)</regexp>


<regexp>[/\\]season[^\w]?(\d{1,2})[^\d]*[/\\].*?[^\w]part[^\w]?(\d\d?)[^\w]([^/\\]*)</regexp>


<regexp>[/\\]season[^\w]?(\d{1,2})[^\d]*[/\\].*?[^\w]chapter[^\w]?(\d\d?)[^\w]([^/\\]*)</regexp>



<regexp>[/\\]season[^\w]?(\d{1,2})[^\d]*[/\\].*?\1(\d\d)(?!.*])</regexp>


<regexp>[/\\]season[^\w]?(\d{1,2})[^\d]*[/\\].*?\1[^\w]?x[^\w]?(\d\d?)([^/\\]*)</regexp>


<regexp>[/\\]season[^\w]?(\d{1,2})[^\d]*[/\\].*?s0?\1[ex]{0,2}(\d{1,2})([^/\\]*)</regexp>


<regexp>[/\\]season[^\w]?(\d{1,2})[^\d]*[/\\][^\d]+[\. _-](\d{1,2})[\. _-][^\d]+</regexp>
</tvshowmatching>



<moviestacking action="append">
<regexp>done[ _\.-](\d)[ _\.-]</regexp>
</moviestacking>



<video>
<excludefromscan action="append">
<regexp>[/\\]extras[/\\](?!season)</regexp>
</excludefromscan>
</video>

</advancedsettings>

havix

2009-08-10, 23:22

Are these regex's built into XBMC now? And if they aren't why not?

jmarshall

2009-08-11, 01:07

Some of them (as clearly commented in the file) would cause too many false positives, others are specific to scene groups such as tpz which we clearly don't want by default.

xexe

2009-08-16, 13:16

jmarshall is obviously is spot on. Its fine for users with the skill and motivation to add these but if they were default the false positives would eat up valuable dev support time epecially since users wouldnt have a clue why it was happening.

I only did this for fun and to save time answering the same question over and over via the support IRC.

Now dont get me wrong it IS surprisingly accurate and in my simulations we are talking only 1 ep in a 1000 errors BUT a user with a naming scheme that triggers one of these errors will likely trip hundreds of them.

Lastly the TPZ naming scheme matching will NEVER be 100% accurate as their naming is quite simply completely useless.

j3ff

2009-09-10, 05:37

Thanks for this - before I found this thread I was beating my head against the wall wondering why XBMC was not finding my stuff. This made it all 100% perfect in 2 minutes.

dc_williamson

2009-09-10, 16:54

Most of my TV shows are stored in the folder structure:
show name\season n\show name Snn Enn.extn
(where nn=two digits 0-9).
None of the above regex work for me. However I use this in my advancedsettings.xml file:

<tvshowmatching action="prepend">
<regexp>[\\/]*S([0-9]+) E([0-9]+)[^\\/]*</regexp>
</tvshowmatching>

Which works fine :)

Only issue comes when I have shows with no season number (like they were one season one) in the format show name\show name Enn.extn

I'm still trying to work out a regex that'll work for this - for now I'm fudging it by renaming all the files in a Snn Enn format :) Anyway, feel free to use or add this to the above it it's useful.

locust

2009-09-15, 00:41

Boom!

thanks so much for your hard work.

I have lots of -TOPAZ releases and this was pissing me off to no end how it wouldn't scan in their episodes

Nicely done!

locust

2009-09-22, 02:56

Actually I am not having luck with -MEDiEVAL rips though, for example

Arrested.Development.S01E05.WS.DVDRip.XviD-MEDiEVAL has an archive name of med-ad105.rar and it's not picking it up (and i cant scan it into my library)

any chance of getting a solution for that?

thanks!

EvilMatt666

2009-11-07, 01:03

I'm kinda new here but I've been messing about with XBMC for a while now. I have just installed the newest version and thought maybe it might sort my library problem for the tv show rips on my system but it hasn't and looking at the REGEX code I would doubt it would sort my naming system properly.

Basically the way I rename all my files (because I'm anal) is as follows:-

Tv Hard drive/tv show title/season 1/tv.show.-.101.-.episode.title.(DVD).avi/mkv/etc

"101" would mean season 1 episode 01 and then "2024" would be season 20 episode 24. It just doesn't seem to pick up the episode details at all. In fact this last pass with the REGEX file utilised has just given me files numbered 1-30 and nothing inside in Library view. I can and have been using the file view to go through my TV shows but it's not perfect. Anyone got any ideas?

Thanks in advance.

jmarshall

2009-11-07, 04:18

@EvilMatt666. Do a debug log while refreshing a show. It'll tell you right off whether it's detecting them properly or not.

Don't just guess that things aren't working due to your file naming!

xexe

2009-11-10, 14:57

Can users confirm the the "Extras" removal regex is working for them. I have had reports that it works but i cannot get it working for myself which is curious.

If it does work can you confirm if you use the tvshow.nfo and movie.nfo URL tag method.

xexe

2009-11-11, 14:22

xexe

2009-11-16, 12:06

Version 1.2

See file for some more details of whats changed. Small change to follow new config options.

See it here: http://pastebin.com/f6edebc7a

Remember I am also looking for feedback. Without it I cannot know if this is working for you or if anyone is using it. Eventually i will lose interest otherwise :)

biggles

2009-11-17, 11:29

Version 1.1

See file for some more details of whats changed. In essence this is a cleanup release.

In the interim I am now using pastebin for uploading this config file as it allows for a permanent home and direct download of the file (which is easier for new users). Eventually I may have a direct download permanent home, who knows

See it here: http://pastebin.com/f62eee83b

I am also looking for feedback about how this is working for you.

I recently started using this version (didn't include all you personal preferences, but all the regex). Works like a charm. Maybe some of these should be pushed into the main branch further down the road...

Thanks a lot for writing it up!

xexe

2009-11-17, 12:25

No problems.

Unfortunately these will likely never be included since they have the potential of false positives which XBMC cant include.

Glas to see it works for you and thanks for the feedback

outleradam

2009-11-20, 02:22

xexe: is it possible to use regex to identify something like this:
/showname-(episode title)-YYYY-MM-DD-X.mpg

The YYYY-MM-DD = recording date, the X is to identify parts of the same recording. This is the default for BeyondTV (http://www.snapstream.com/).