Movie matching with included tagline

[joshinils 13]

I generate movie stubs from lists i scrape from  moviepilot.de

the movies on that (german) site sometimes contain a tagline, separated by " - ".
for example: "Alles ist erleuchtet - Nichts ist normal" is the title of https://www.themoviedb.org/movie/340-everything-is-illuminated which of course is a german variant.

however themoviedb does not include the tagline with the title. the log shows that is searches themoviedb for 4 variants:

"Alles ist erleuchtet - Nichts ist normal&language=de&year=2005"
"Alles ist erleuchtet   Nichts ist normal&language=de&year=2005"
"Alles ist erleuchtet - Nichts ist normal&language=en&year=2005"
"Alles ist erleuchtet   Nichts ist normal&language=en&year=2005"

none of which result in any match, since the title listed on themoviedb is only "Alles ist erleuchtet" without the tagline (which I now added to the german translation part of the movie on themoviedb)

maybe searching for the movie without the tagline (i.e. cut off after the dash) could help, since themoviedb does not "care" about the tagline and wont find a movie if you include it.


another example: "Cloud Atlas - Alles ist verbunden"  https://www.themoviedb.org/movie/83542-cloud-atlas?language=de-DE
for that i have added an alternate title for the german translation which includes the tagline separated via the dash. that (of course) now gets found since it is a "viable" title, after manually triggering the update metadata on the stub entry.

i see this issue mainly with themoviedb and taglines being excluded from their search.

[joshinils 13]

I could just go ahead and add the tagline-version for all moves it does not match, however that feels wrong to me and like a lot of (potentially unnecessary) work.

[Luke 42078]

Hi there, how are your files named and organized?

[Happy2Play 9780]

Also not it takes time for the API to show website updates.  Or at the same time Emby cached language metadata in \cache\tmdb-movies2\340.

But at the same time not every movie has a translated tagline per language.

As for second issue that is just TMDB rules as there is no true translation for German release title, which cause them to Lock the Title.

Quote

"The translated title should be the first official translation. In other words, the title of the first—theatrical, physical, digital or TV—release." and "The translated title field should be left blank if there is no official translation. Please do not add unofficial translations.".

 

[joshinils 13]

12 hours ago, Luke said:

Hi there, how are your files named and organized?

i have various libraries, each for different purposes.
one folder structure containing the stubs:

├── emby_stubs_jo
│   ├── blacklisted
│   ├── rated
│   │   ├── moviepilot
│   │   └── streams
│   └── watchlisted

where i generate the stub-files from moviepilot. each folder containing the stubs files, so for example:

emby_stubs_jo/rated/moviepilot/12 Monkeys (1995) - moviepilot 8.0 jo_rated.disc
emby_stubs_pe/rated/moviepilot/Alles ist erleuchtet - Nichts ist normal (2005) - moviepilot 2.5 pe.disc

is such a generated filename. (the content seems to be irrelevant, but I put the link for moviepilot in it or to netflix for example if it is available there)

I get the movie names and release years from moviepilot (which seems like my main problem) since i have an account there for rating the movies i have seen and they generate predictions for movies i have not seen.

One of my libraries contains my rated movies, so i can look at only those and filter by not seen and mark all movies i have seen as seen.
(/home/nials/settings/moviepilot/emby_stubs_jo/rated)

I also scrape their information on who is streaming a movie, so i can have a stub for movies which are available for streaming and on my watchlist.
I then have a library with all of these combined, so that they get merged into a single movie and i can see where i can watch a movie.

My actual movie files are structured like this:

├── filme
│   ├── 1910-1919
│   ├── 1930-1939
│   ├── 1940-1949
│   ├── 1950-1959
│   ├── 1960-1969
│   ├── 1970-1979
│   ├── 1980-1989
│   ├── 1990-1999
│   ├── 2000-2009
│   ├── 2010-2019
│   ├── 2020-2029

where each decade contains folders for each move, like this:

Logan - The Wolverine_2017
├── extras
│   ├── Logan - The Wolverine_2017 - bonus 1.mp4
│   ├── Logan - The Wolverine_2017 - bonus 2.mp4
│   └── Logan - The Wolverine_2017 - bonus 3.mp4
└── Logan - The Wolverine_2017.mp4

but most of them only contain the single media file of the movie, a few have multiple versions.

[joshinils 13]

12 hours ago, Happy2Play said:

Also not it takes time for the API to show website updates.  Or at the same time Emby cached language metadata in \cache\tmdb-movies2\340.

Of course, but for the movies where i have added the alternative title they have all been found within minutes, deleting the cache entry 340 and waiting till now did nothing.
manually searching for https://www.themoviedb.org/search?query=Alles+ist+erleuchtet+-+Nichts+ist+normal also does not find a movie, they don't care for the tagline in their search

12 hours ago, Happy2Play said:

But at the same time not every movie has a translated tagline per language.

that does not seem to matter for taglines, or extended titles (which germany just loves for some reason... meh) so the only thing i could realistically do is add the "taglined" version as an alternative title, which has worked great so far.

12 hours ago, Happy2Play said:

As for second issue that is just TMDB rules as there is no true translation for German release title, which cause them to Lock the Title.

yeah, but i can always add alternative titles, which get included in the search.

 

a problem with moviepilot is also that they have the wrong release year, i.e. one year too early, and neither wikipedia or imdb or tmdb has any source for that.
also some tv-series are listed as a movie.
sometimes imdb shows an earlier release at a festival and i can add that to tmdb which fixes the "wrong" release year on tmdb, so then it matches with the release year on moviepilot.

both of which is a data problem with moviepilot, where i can only submit a database problem and hope anyone cares about it.
but so far i have to live with these empty entries of series listed wrongly listed as a movie on moviepilot where i can only add the imdb id within emby, since they are unique across movies and tv-series.

[Happy2Play 9780]

Okay now if understand, I don't see how a dev is suppose to know how to parse thst TITLE - TAGLINE naming scheme?  How are they to even know that is not just a TITLE?

3 hours ago, joshinils said:

a problem with moviepilot is also that they have the wrong release year, i.e. one year too early, and neither wikipedia or imdb or tmdb has any source for that.

Common issue as every country could have different release date so it truly creates chaos as Cloud Atlas is 2012 or 2013.

[joshinils 13]

Replying to mark a solution.

When this happens, the easiest option is to add the differing title with the tagline to themoviedb.org, which then means that it will get matched since it is a valid title.