YAMJ / yamj-v2

Yet Another Movie Jukebox (YAMJ) v2
GNU General Public License v3.0
28 stars 11 forks source link

Episode title only extracted if "-" comes right after S01E01 #582

Closed Omertron closed 9 years ago

Omertron commented 9 years ago

Original issue 583 created by Omertron on 2009-03-04T13:50:42.000Z:

What steps will reproduce the problem?

  1. File name Friends.S01E01.DVDRip - The One Where Monica Gets a Roommate.avi
  2. The episode title "The One Where Monica Gets a Roommate" not extracted from the file name. Episode marked "Episode 1".

What is the expected output? What do you see instead? "The One Where Monica Gets a Roommate"

Please use labels and text to provide additional information. r790

Omertron commented 9 years ago

Comment #1 originally posted by Omertron on 2009-03-05T00:09:00.000Z:

Are you looking at the output of a skin (which one), or the actual xml file generated?

If the former, that's probably a skin issue. If wrong in the xml, then I agree with you that the '-' should be permitted anywhere after the season info.

Omertron commented 9 years ago

Comment #2 originally posted by Omertron on 2009-03-05T02:49:39.000Z:

That was an done intentionally. See the discussion around http:// www.networkedmediatank.com/showthread.php?tid=15877&pid=134918#pid134918

Omertron commented 9 years ago

Comment #3 originally posted by Omertron on 2009-03-05T03:37:53.000Z:

What was intentional? The use of '-' being mandatory, or it has to follow the season/ep info.

The use of '-' is fine, but it should be allowed anywhere after season/ep and only the part that follows the '-' is the ep title.

Omertron commented 9 years ago

Comment #4 originally posted by Omertron on 2009-03-05T07:47:15.000Z:

Currently there are two places in the code where file title is being read from file name:

In both cases different parsing algorithm is used. After Part01 - the file title is any text between last dash (after tags) and file extension. After S01E01 dash MUST come right after the S01E01. I guess the algorithm is to be unified for both cases.

I do not see any problem to allow title recognition in the given example: "Friends.S01E01.DVDRip - The One Where Monica Gets a Roommate.avi"

Omertron commented 9 years ago

Comment #5 originally posted by Omertron on 2009-03-05T13:13:36.000Z:

The choice to limit the dash to immediately following the season and episode.

The problem is with file names like these:

House.S03E14.HDTV.XviD-XOR.avi Dead Like Me S02E12 - Forget Me Not.HDTV.XviD-vfua.avi

Omertron commented 9 years ago

Comment #6 originally posted by Omertron on 2009-03-05T13:54:04.000Z:

Thanks for examples.

House.S03E14.HDTV.XviD-XOR.avi

This I would skip by using the "ignore list". We already have something. Or another option would be to use "dash with spaces" to determine the title.

Dead Like Me S02E12 - Forget Me Not.HDTV.XviD-vfua.avi

This is more interesting. I will take in into account for sure.

Omertron commented 9 years ago

Comment #7 originally posted by Omertron on 2009-03-05T20:10:54.000Z:

Now I remember more what the problem was with those examples. Yes, I also hated the scene info showing up as title even though I am one to often keep the original filenames. My suggested solution is to find the first metadata keyword and ignore until end of file or beginning of [] for parts and trailers.

metadata keywords would include: PDTV, HDTV, DVD, NTSC, PAL, 720p, 1080p, BluRay, HDDVD, DTS, AC3, x264, xvid, etc Once the first occurrence is found, use everything from '-' to this first keyword, beginning of [], or end of file (up to extension '.'), whichever occurs first.

Perhaps when I have some time I'll take a look at the code to see how the current metadata functions could be used to help with my idea.

Omertron commented 9 years ago

Comment #8 originally posted by Omertron on 2009-03-05T20:17:17.000Z:

One more point I forgot, those metadata keywords have to be surrounded by valid delimiters like ( ,-.) so that '.PAL.' would not be confused with APPALOOSA movie.

As well, because there are so many scene groups, using the moviejukebox.properties to exclude keywords will not be a convenient solution. However, most scene tags are preceded with 'xvid-', 'x264-', etc so ignoring that '-' may be possible but not very clean. As for scene groups that start the file name with group-obscure.file.etc.avi, I say screw them and let people rename the file :D

Omertron commented 9 years ago

Comment #9 originally posted by Omertron on 2009-03-05T20:23:20.000Z:

Thanks for detailed description.

Omertron commented 9 years ago

Comment #10 originally posted by Omertron on 2009-03-06T07:59:31.000Z:

<empty>

Omertron commented 9 years ago

Comment #11 originally posted by Omertron on 2009-03-06T16:43:40.000Z:

Here's a more detailed list I created inside an AutoIt3 script I use to rename filenames by stripping out metadata keywords.

Dim $listExtra = "1080p|720p|BluRay|Blu-ray|HDTV|PDTV|HDDVD|HD.DVD| DVD|NTSC| PAL|x264|" & "AC3|DTS|AAC|DD5.1|DD51|DD2.0|DD20|1080i|480p|576p|SDTV| OAR| FS| WS|" & "Xvid|divx|dxva|L41|L4.1|REPACK|PROPER|" & _ "Unrated|DirCut|Extended|Uncut| UR| DC| EE| UC| SE| TC| XXX|"

In my case I ignore whitespace unless I included it specifically. Most of those are already known by existing functions in the Java code, and would probably serve well enough in 99% of the cases.

Omertron commented 9 years ago

Comment #12 originally posted by Omertron on 2009-03-07T18:56:22.000Z:

Fixed in r798.

Omertron commented 9 years ago

Comment #13 originally posted by Omertron on 2009-03-07T21:59:30.000Z:

Also helpful to state here what was actually done, and how it affect the rules for '-'. I believe it's important to let those that starred an issue, developers, and those answering support questions to know what has been changed.

Since many various discussions occur in an issues' thread, it's not always obvious what 'fixed' means.

Omertron commented 9 years ago

Comment #14 originally posted by Omertron on 2009-03-08T11:55:27.000Z:

I'm going to describe the rules in Wiki. Just had no time to do it. I'have used a lot of examples from this an similar issues in the test. There are probably still problems but now it will be easier to fix.