[arnuschky]

Hey,

just upgraded to Gotham and having troubles with my TV show scraper. My shows are organized as follows:

Code:

Showname (year)
   Season 1
      [Showname] S01E01 Episodename
      ...
   Season 2
      ...
Othershowname (year)

According to the wiki [1], this should be ok, but the scraper doesn't recognize the shows.
If I remove the year from the top-level show directory, everything works.
I would prefer keeping the year, as without it some shows are wrongly recognized.

Any hints why this is happening?

PS: I am using thetvdb as a scraper, but I am not sure if this is scraper-specific or general xbmc. Sorry if wrong board.



[1] http://wiki.xbmc.org/index.php?title=Nam...s/TV_shows

[helta]

What specific shows are you having issues with?

[arnuschky]

What specific shows are you having issues with?

Do you mean which shows aren't properly scraped without the year in the directory name?
In that case, I need to look it up, can't remember.

If you mean which shows aren't properly scraped with the year in the directory name, it's easy: all of them.

[helta]

The main thing is, some shows on the TVDB have a year and some don't. You have to name them EXACTLY as they are on theTVDB.

So for example:
Dexter = no year in the title
Doctor Who (2005) = 2005 version of Doctor Who.

[arnuschky]

The main thing is, some shows on the TVDB have a year and some don't. You have to name them EXACTLY as they are on theTVDB.

So for example:
Dexter = no year in the title
Doctor Who (2005) = 2005 version of Doctor Who.

Ah. That's absolutely not clear from the wiki, and, if I might say, quite inconsistent.
If the scraper uses only a subset of the information to map a directory to a database entry, fine, but simply not recognizing an entry due to extra information seems rather unintuitive.

Thanks for your answer, btw. So this is a problem with the scraper and not with the way the base info is pulled from my dir structure?

[helta]

Well I would believe it more or less comes down to disambiguation.

So lets say for example you had the 2005 version of Doctor Who:

Your dir structure is as follows:
Doctor Who (2005)
Which will scrape correctly.

Now lets say you had some over information that said that the 2005 version of that show actually started in 2004 (bad torrent naming or something stupid), and you had your dir structure:
Doctor Who (2004)
Which I believe will not scrape correctly, and for good reason. The scraper can't find the show you are talking about because it assumes YOU know what you are talking about.

What if there is a version of Doctor Who from 2004, but its not added to the TVDB? Would you want the scraper to download all the info from the original Dr Who or the 2005 version?

I agree with that, sometimes its better to have no information rather then wrong information.

[arnuschky]

So, if I understand correctly, I must not use the year in the directory name, unless the theTVDB needs if for some entries, at which point I must put it?

This is 1) unintuitive from the user's point of view, 2) defeats the point of a "hands off" scraper, and 3) inconsistent with the docs.

What you described in your last post is a scenario in which it's required to specify the year, which is fine.
However, if adding extra (but correct) information makes the scraper not recognize directories, something is wrong imho.

Well, anyways, thanks for your help. I'll remove the years from all my dirs and leave it be.
Someone with write access should maybe update http://wiki.xbmc.org/index.php?title=Nam...s/TV_shows

[steelman1991]

The main thing is, some shows on the TVDB have a year and some don't. You have to name them EXACTLY as they are on theTVDB.

So for example:
Dexter = no year in the title
Doctor Who (2005) = 2005 version of Doctor Who.

Ah. That's absolutely not clear from the wiki, and, if I might say, quite inconsistent.

I think you'll find it is crystal clear. Have another read of the wiki, particularly the second and third sentences of 1. Directories - TV Show Naming

[helta]

Yeah, I'm not saying it can't be better either, and I do get your point.

BTW, do some searching of Heimdall in the developlment forum here on the boards. This may address all of your issues in the future, but its not ready yet.

Pretty amazing stuff.

[arnuschky]

Cool, looks really nice. Wow, in dev since 2012 - seems to be a lot of work.

[steelman1991]

No comment on your undeserved criticism of the wiki entry - you seemed vocal enough previously hmmmmmmm.

[un1versal]

It may be unintuitive but its logic, essentially the year is used only to differentiate remakes or shows with same name but made in different year of the same show when they exist, therefore the match is exact.

A good example is Rome vs Rome 2005 where Rome will match but Rome 2005 doesnt exist as a showname therefore wont match.

I agree that smarter regexp can have better fallback but eh, unless you make the sourcecode or the documentation better, moaning wont do you any good, so you have a great opportunity to contribute/improve something here, now that you are familiar with the reasons.

[Ned Scott]

No comment on your undeserved criticism of the wiki entry - you seemed vocal enough previously hmmmmmmm.

It's a lot of information on that page, and much of it could be presented more clearly for users, so I honestly don't blame him for being critical of it. Even if the info is technically there, if it is east to miss then that in itself could be seen as an error on the wiki's side.

The page has improved a lot over the years, but I wonder if a two-tiered approach would be better, with more detailed/technical info on a second page. It's very easy, even for someone like me who works on the wiki and knows XBMC, to get info overload on the existing naming page and miss something under my nose.

Sometimes there is undeserved criticism, and a lot of people don't even try to look up info before asking questions, so I understand your defense of the wiki and thank you for it, but I can see both sides here and see this as good feedback to make improvements on :)

[nickr]

Basically it comes down to knowing what thetvdb.com calls the show. On that website the modern Doctor Who is called Doctor Who (2005). The original series is called Doctor Who.

Dexter is called Dexter, because there is no other programme called Dexter (at least not on that database).

If in doubt name it like thetvdb does, it'll save a world of pain.

[arnuschky]

No comment on your undeserved criticism of the wiki entry - you seemed vocal enough previously hmmmmmmm.

It's a lot of information on that page, and much of it could be presented more clearly for users, so I honestly don't blame him for being critical of it. Even if the info is technically there, if it is east to miss then that in itself could be seen as an error on the wiki's side.

The page has improved a lot over the years, but I wonder if a two-tiered approach would be better, with more detailed/technical info on a second page. It's very easy, even for someone like me who works on the wiki and knows XBMC, to get info overload on the existing naming page and miss something under my nose.

Sometimes there is undeserved criticism, and a lot of people don't even try to look up info before asking questions, so I understand your defense of the wiki and thank you for it, but I can see both sides here and see this as good feedback to make improvements on

Thank you Ned. It was never my intention to appear ungrateful, sorry for that.
I'm a programmer, so I think this shaped how I looked at the documentation.
My mistake was simply that I hadn't understood how theTVdb names its entries, and that this drives everything.