Author Topic: File Scanner - concatenation  (Read 6980 times)

0 Members and 1 Guest are viewing this topic.

Offline goddert

  • User
  • ***
  • Posts: 39
    • View Profile
File Scanner - concatenation
« on: April 03, 2012, 04:30:44 pm »
Hi,

I have a question regarding the File Scanner. I did a forum search but didn't find an answer (or didn't use the right terms)

These are my regular expression and they are working very well for most of my movies
^.*\\(?P<origtitle>.*) - S(?P<season>\d\d)E(?P<episode>\d\d) - (?P<eptitle>.*) \[\d+p\d+\]
^.*\\(?P<origtitle>.*) - (?P<title>.*) \[\d+p\d+\]
^.*\\(?P<origtitle>.*) \[\d+p\d+\]

But a lot of my movies exist in two versions (two languages).
Origtitle - Title1 [1080p23].mkv
Origtitle - Title2 [1080p23].mkv

In PVD: Same original title (origtitle) but different titles (every one in its own language). I think this is a very consistent and canonical way.
When I now start a scan the File Scanner tries to concatenate the two movies he finds as one movie (the first it finds) and set this as the new path for the first movie. (see screenshot)

I didn't find an option to switch off this behaviour.
Any ideas?

Thanks  :)


Offline rick.ca

  • Global Moderator
  • *****
  • Posts: 3241
  • "I'm willing to shoot you!"
    • View Profile
Re: File Scanner - concatenation
« Reply #1 on: April 03, 2012, 09:33:21 pm »
Quote
In PVD: Same original title (origtitle) but different titles (every one in its own language).

So the scanner should be using <title>, not <origtitle>.

But why maintain multiple records for the same movie, when all that's different is the language? And why have multiple files, when one video container can include multiple audio streams and any number of subtitles? Wouldn't it be easier just to select the desired language and/or subtitles in the player?

Quote
When I now start a scan the File Scanner tries to concatenate the two movies he finds as one movie (the first it finds) and set this as the new path for the first movie.

This is a real problem where there's a need to intentionally record two or more paths in one record (e.g., a movie and it's trailer). When one is matched, the scanner is unable to add the other.

Offline goddert

  • User
  • ***
  • Posts: 39
    • View Profile
Re: File Scanner - concatenation
« Reply #2 on: April 04, 2012, 09:25:22 am »
Quote
So the scanner should be using <title>, not <origtitle>.

Hmm, so you're saying currently the scanner evaluates only the named capture <origtitle> to compare the filename with the movies in PVD? To tell the truth I expected the scanner to evaluate every one of the named captures (<origtitle>, <title>, <year>, <season>, <episode>, <eptitle>) contained in a regular expression. Then in case that their combination in a filename matches an existing record in PVD it would offer to change the path (if different). Only if they don't match it would offer other guesses it made in base of filename, records etc.
It does so at least for the named captures <episode>, <season> and <eptitle>. There it does detect whether <eptitle> already exists in PVD or not and it doesn't concatenate.

From my point of view for the file scanner a filename (movie/episode) which matches exactly a regular expression and matches exactly a movie in PVD (id est all three, filename, regular expression and record, are intentionally managed by the user) should always have the highest priority over the suppositions which makes the software.

Quote
But why maintain multiple records for the same movie, when all that's different is the language?

Yes, that would be my ideal world :). Most of my movies are done this way (and a lot of work to synchronize :(. Unfortunately this can't be done for all movies ... First there are different versions for both languages like an extended cut and a normal version. (those have different original title) Then we have different versions which seems to be the same movie. But for some reason the distributors for one or the other country decided to cut out different scenes of the movie (I.g The Queen, Angels & Demons and many others). So in the end you have different video tracks (and consequently different audio tracks). Although there would be hardly a saving in diskspace I could put them all in one container (2 video tracks and 2 audio tracks) but unfortunately my streaming box doesn't support choosing between different video tracks in the same container, only between audio tracks and subtitles. So in the end I have to maintain two different movie containers id est files.

Quote
This is a real problem where there's a need to intentionally record two or more paths in one record (e.g., a movie and it's trailer). When one is matched, the scanner is unable to add the other.

As I said above for <episode>, <season> and <eptitle> it doesn't. Trailers for episodes are also very common ;) And it doesn't concatenate. This doesn't seem to be a consistent way to treat the filenames. I think the treatment should be consistent through all kind of filenames. One way or the other.

I didn't consider the fact trailers (because more than one file for the same movie (parts) could be resolved very easily with regular expressions)
I could immagine a solution to this problem would be an option to switch of this concatenation for all kind of filenames (a similar option exists for multiple disc detection). But the best way, IMHO, would be the logic mentioned above.

Thanks
« Last Edit: April 04, 2012, 09:27:27 am by goddert »

Offline rick.ca

  • Global Moderator
  • *****
  • Posts: 3241
  • "I'm willing to shoot you!"
    • View Profile
Re: File Scanner - concatenation
« Reply #3 on: April 05, 2012, 02:59:23 am »
Quote
Hmm, so you're saying currently the scanner evaluates only the named capture <origtitle> to compare the filename with the movies in PVD?

No, I mean I didn't read your post carefully enough to see your second regex would capture <title>. :-[

Yes, I too would expect it should be able to make a correct match when both <origtitle> and <title> have been captured. But I can also vaguely see why the logic might cater to the 'normal' use of the corresponding fields (where [Title] is an alternate name for [OrigTitle], not a different movie). It's been a long time since he wrote it, but maybe nostra can recall whether this is something he might easily modify.

Quote
I could immagine a solution to this problem would be an option to switch of this concatenation for all kind of filenames (a similar option exists for multiple disc detection). But the best way, IMHO, would be the logic mentioned above.

I'm not sure what you mean by this. My trailers are intentionally named so they match the same regex that matches the movie. So I understand the program cannot know whether such a match (when [File Path] is not empty) is something to be added (i.e., concatenated), a change to the existing value, or is for a new unrelated movie. But there should be the possibility to in the results dialog to change the action from 'Change' to 'Add'. I suppose it might be handy to have buttons for changing all of one type of recommendation (i.e., 'New movie/episode', 'Change path' or 'Add file') to another—especially when scanning after adding a batch of movies, episodes or trailers, or moving folders.