Scanner mistakes with movies

[gquiring]

There seems to be a trend here, if the first name is the same on these movies the scanner is not looking at the date.  It thinks they are all the earlier version.  For example Ghostbusters 2016 is shown as the 1984 version.  Blade Runner 2017 it thinks it is the 1982 version.  


Blade.Runner.1982.4K.2160p.UHD.BluRay.HDR.x265.mkv
Blade.Runner.2049.2017.UHD.BluRay.2160p.TrueHD.Atmos.7.1.HEVC.mkv

Ghostbusters.1984.2160p.uhd.x265.mkv
Ghostbusters.2016.Extended.UHD.BluRay.2160p.DTS-HD.MA.5.1.HEVC.mkv

Hellboy.2004.DC.2160p.UHD.BluRay.X265.mkv
Hellboy.2019.UHD.BluRay.2160p.TrueHD.Atmos.7.1.HEVC.mkv

Halloween.1978.UHD.2160p.BluRay.REMUX.HDR10.HEVC.TrueHD.7.1.mkv
Halloween.2018.REMUX.2160p.(10bit).BluRay.UHD.HDR.HEVC.DTS-X.7.1.mkv

[Venson]

https://jellyfin.org/docs/general/server/media/movies

[gquiring]

I'm really surprised the Jellyfin scanner can't handle the dates with periods. I have been using Kodi for 20 or so years and am trying out Plex and Jellyfin. Plex and Kodi work perfectly with those formats. If I have to rename every file and structure it like that document has I'm not sure it's worth the effort.

[Venson]

JF is quite strict when it comes to filenames and media structure. Its been long on our todo list to rewrite the parser but that will not happen overnight.
If you are more inclined to use plex because it fits your purpose, by all means please do so. If you however want to give it a try, there are plenty of tools out there that refactor the naming scheme so jellyfin is able to read them. Otherwise you could just go with .nfo files so jellyfin will more or less ignore the filename for any data collection.

[tmsrxzar]

from what i see your file names need attention any way, they're not even consistent

Halloween.1978.UHD.2160p.BluRay.REMUX.HDR10.HEVC.TrueHD.7.1.mkv
Halloween.2018.REMUX.2160p.(10bit).BluRay.UHD.HDR.HEVC.DTS-X.7.1.mkv

on 1 you went date.UHD.2160p.BluRay.REMUX then change to date.REMUX.2160p.(10bit).BluRay.UHD

then on another you abandoned everything and just went date.2160p.uhd.x265.mkv

for my own collection, dates and periods work completely fine BUT you should be more consistent and accept that in some cases the file may need more information in the filename

go with this format Title.Year.Version.Resolution.ExtraCrap.Codecs.ext
example: Ghostbusters.1984.2160p.UHD.BluRay.REMUX.TrueHD.Atmos.DV.HEVC.mkv
example: Ghostbusters.1984.DC.2160p.UHD.BluRay.REMUX.TrueHD.Atmos.DV.HEVC.mkv

bladerunner 2049 is already an issue for most scrapers because it indicates to the scraper that the movie year is "2049" so putting it in a folder of the same name with an added [imdbid-] will be needed

ghostbusters 2016 is probably the same

and since youve been doing this for 20 years you should already know that scrapers basically use a regex to isolate title+year then literally search tmdb with that exact information, very little logic is applied to the results unfortunately so you will have to provide your own logic


to Venson and the rest of the jellyfin team, it would indeed be helpful to filter the returned results and only use ones that match an identified year from the filename, if available
if the search was for "ghostbusters"+"&year=2016" then results returned with a release year of 1984 are obviously incorrect

[tmsrxzar]

small follow up to that

you can see tmdb respects &year= in the case of blade runner 2049, when the title is identified correctly as such
https://www.themoviedb.org/search?query=...&year=2017

this is where identifying and differentiating 2049.2017 would be needed

but in the case of ghostbusters it does not
https://www.themoviedb.org/search?query=...&year=2016

this is where filtering the returned results would be needed

[Venson]

> to Venson and the rest of the jellyfin team, it would indeed be helpful to filter the returned results and only use ones that match an identified year from the filename, if available
> if the search was for "ghostbusters"+"&year=2016" then results returned with a release year of 1984 are obviously incorrect

We discussed several times to improve the parser but came to the conclusion that its pointless as its a hot mess of spagetti code and no amount of "fixing" will make anything really better as the possibility to introduce new bugs is just too high.

[tmsrxzar]

> to Venson and the rest of the jellyfin team, it would indeed be helpful to filter the returned results and only use ones that match an identified year from the filename, if available
> if the search was for "ghostbusters"+"&year=2016" then results returned with a release year of 1984 are obviously incorrect

We discussed several times to improve the parser but came to the conclusion that its pointless as its a hot mess of spagetti code and no amount of "fixing" will make anything really better as the possibility to introduce new bugs is just too high.

that's rather discouraging, at least imdbid's work for edge cases since nobody wants to fix the code

[Venson]

its not we dont _want_ it, just the other 10000000 todos we have^^ if we fix that, the people who want intigrated jellyscrub will be disapointed and if we dont do that its intro detection and if not that its the EFCore refactoring and not that the latest section bugs and not that ...