Metadata not being picked up for large part of movie library

[steve1977 67]

I am coming back to Emby from using Plex for a while. I have a fairly large library that was 100% detected by Plex. I have now tried to add the same library to Emby, but facing some issues with movies not being detected or metadata not being added correctly. We are talking about a few dozen movies that are not or wrongly identified. Let me start with what I have seen as root causes:

 

1) IMDB vs TMDB movie name

Example file name: /The Last Man on the Moon (2014) [WEBDL-480p]/The Last Man on the Moon (2014) [WEBDL-480p].mkv

IMDB movie name: The Last Man on the Moon (2014), tt3219604

TMDB movie name:The Last Man on the Moon (2016), 276537

Emby result: movie not identified

How Plex does it: allows to identify movies by their IMDB ID and then use what metadata from what sources to scrape

 

2) Foreign language movies

Example file name: /The Beauty Inside 2015 (2015) [bluray-1080p]/The Beauty Inside (2015) [bluray-1080p].mkv

TMDB movie name: The Beauty Inside (2015), 338729

TMDB original movie name: 뷰티 인사이드

Result: movie not identified

How Plex does it: scans through all languages of the movie title

 

3) Wrongly picks up "in development movies"

Example file name: /The Brothers Lionheart (1977) [DVD]/The Brothers Lionheart (1977) [DVD].avi

TMDB movie name: The Brothers Lionheart (1977), 11291

TMBD movie name (work-in-progress): The Brothers Lionheart, 254318

Result: wrongly picks up "work-in-progress" movie

How Plex does it: filters out work-in-progress movies

 

4) Movies with wrong NFO files

Some of my movies include outdated NFO files

Result: lots of info included from NFO, but some info missing

How Plex does it: allows to prioritize order of accessing info (first NFO or later)

 

5) Poster not being picked up

Some movies are correctly identified, but don't have poster art

Example: IMDB tt0053268 / TMDB 12552

Result: all info correctly reflected, but no poster being picked up

 

 

Thanks for looking into this.

Edited October 7, 2018 by steve1977

[Luke 42080]

Hi, for #1 we already support that as well, but where is the imdb id? You would have to have it in an nfo file or in the file name.

 

We do not query imdb to find the imdb id. It is not a free and open data service.

[steve1977 67]

Thanks for your insanely quick reply, much appreciated!

 

re 1 - I thought Plex somehow queries some database that has IMDB titles (not just takes it from the NFO files)? Does Plex pay for it or do I understand this wrong how it is done? The example above (The Last Man on the Moon) does not have an NFO file, but it is being picked up by Plex (but not by Emby).

 

When you have a chance, would be good to hear your thoughts on 2-5 (I manually edited and added a #5). Thanks!

[Luke 42080]

It's possible they are paying for an imdb data license. It's very expensive. Maybe we get there at some point. We need to grow our user base a bit first.

[Luke 42080]

#2 should already work so I suspect something else is at play. You can verify on your own using the MovieDb website. Set the language to english, search for The Beauty Inside, and it will come up.

[Luke 42080]

For #3 can you find out how to determine this using Moviedb api data? When using the search api for the brothers lionheart, this is the response:

{"page":1,"total_results":2,"total_pages":1,"results":[{"vote_count":0,"id":254318,"video":false,"vote_average":0,"title":"The Brothers Lionheart","popularity":0.6,"poster_path":"\/us4HARgUkkFluMIi7rOklKB0CJ5.jpg","original_language":"en","original_title":"Bröderna Lejonhjärta","genre_ids":[],"backdrop_path":null,"adult":false,"overview":"A pair of brothers who have both died, meet up in an afterlife world known as Nangijala.","release_date":""},{"vote_count":32,"id":11291,"video":false,"vote_average":6.9,"title":"The Brothers Lionheart","popularity":3.581,"poster_path":"\/1hmscSYcc2T177eXdRK0MPvGMtK.jpg","original_language":"sv","original_title":"Bröderna Lejonhjärta","genre_ids":[12,14,10751],"backdrop_path":"\/fIRnxV5zNnqTro1e6bEAyBPgHkj.jpg","adult":false,"overview":"9-year old Karl Lejon, 'Skorpan' is suffering from tuberculosis and knows that he will die soon. But his older brother Jonathan tries to comfort him by promising that they one day will meet in Nangijala, a magic country beyond the stars. Jonathan dies soon after in a house fire and Skorpan also passes away not long after. They meet in Nangijala, a country divided into the Cherry Blossom Valley and the Briar-Bush Valley. The people in the later are oppressed by the black knight Tengil and his dragon Katla. The two brothers, who now call themselves Jonathan and Karl Lion Heart, decide to help the freedom fighters against the oppression.","release_date":"1977-09-23"}]}

That work in progress title is the first one in the list. It's not obvious how to filter it out.

[Luke 42080]

#4 is possible for the future.

 

For #5, what is your metadata language setting? MovieDb only has one poster, and it is german:

https://www.themoviedb.org/movie/12552-serengeti-darf-nicht-sterben/images/posters

[steve1977 67]

Thanks. Makes a lot of sense and shouldn't be a priority.

 

You are right that #2 is not an issue, but an issue in how I entered the data (folder included an additional 2015 as seen before). I had others movies that I also thought struggled with #2, but in reality this was #3. Both an in-progress and proper movie existed and it picked up the in-progress one instead.

[steve1977 67]

Got in on #5, yes issue on my end as I am scraping an English movie, but no English art available.

[steve1977 67]

For #3 can you find out how to determine this using Moviedb api data? When using the search api for the brothers lionheart, this is the response:

{"page":1,"total_results":2,"total_pages":1,"results":[{"vote_count":0,"id":254318,"video":false,"vote_average":0,"title":"The Brothers Lionheart","popularity":0.6,"poster_path":"\/us4HARgUkkFluMIi7rOklKB0CJ5.jpg","original_language":"en","original_title":"Bröderna Lejonhjärta","genre_ids":[],"backdrop_path":null,"adult":false,"overview":"A pair of brothers who have both died, meet up in an afterlife world known as Nangijala.","release_date":""},{"vote_count":32,"id":11291,"video":false,"vote_average":6.9,"title":"The Brothers Lionheart","popularity":3.581,"poster_path":"\/1hmscSYcc2T177eXdRK0MPvGMtK.jpg","original_language":"sv","original_title":"Bröderna Lejonhjärta","genre_ids":[12,14,10751],"backdrop_path":"\/fIRnxV5zNnqTro1e6bEAyBPgHkj.jpg","adult":false,"overview":"9-year old Karl Lejon, 'Skorpan' is suffering from tuberculosis and knows that he will die soon. But his older brother Jonathan tries to comfort him by promising that they one day will meet in Nangijala, a magic country beyond the stars. Jonathan dies soon after in a house fire and Skorpan also passes away not long after. They meet in Nangijala, a country divided into the Cherry Blossom Valley and the Briar-Bush Valley. The people in the later are oppressed by the black knight Tengil and his dragon Katla. The two brothers, who now call themselves Jonathan and Karl Lion Heart, decide to help the freedom fighters against the oppression.","release_date":"1977-09-23"}]}

That work in progress title is the first one in the list. It's not obvious how to filter it out.

 

My best guess is that I didn't run into this issue with Plex as IMDB may handle / prioritize them differently.

 

I don't know how to code, so no idea whether feasible, but one idea could be to exclude movies from scraping that don't have a release date? There should not be any user that has movies that are work-in-progress (i.e., no release date).

[Luke 42080]

My best guess is that I didn't run into this issue with Plex as IMDB may handle / prioritize them differently.

 

I don't know how to code, so no idea whether feasible, but one idea could be to exclude movies from scraping that don't have a release date? There should not be any user that has movies that are work-in-progress (i.e., no release date).

 

The missing release date is the only way to differentiate, but is that really safe? I'm not so sure. It could be missing because it was never released, or missing because it hasn't been entered into Tmdb yet.

 

The Tmdb search engine is pretty good. I think it is best to respect their search results rather than try to be smart about them, which is something we've tried to do in the past and it has caused problems. In my opinion I think it would be best to request Tmdb to add a search filter to exclude these in-progress titles from the results, and then we can utilize that as part of our search.

[steve1977 67]

Make sense and agree that this may cause other problems. It seems that we are not talking about a lot of issues caused by this. I can well manually change them.

[Luke 42080]

What's probably happening is that if they do have an Imdb search, then the Imdb database will be cleaner as it is professionally maintained with a strict set of rules. It won't even have those in-progress things so it would never be an issue in the first place.

[steve1977 67]

Agree. Thanks for your replies. It seems all solved and makes sense. No big deal to make some manual changes. And most folks won't have such large libraries, so agree that this is nothing to prioritize at this stage.

 

What indeed seems missing though is a feature that allows me to pick up what movies are lacking posters, fanart or TMDB ID? I have placed a feature request for this.

[CharleyVarrick 283]

For #3 can you find out how to determine this using Moviedb api data? When using the search api for the brothers lionheart, this is the response:

{"page":1,"total_results":2,"total_pages":1,"results":[{"vote_count":0,"id":254318,"video":false,"vote_average":0,"title":"The Brothers Lionheart","popularity":0.6,"poster_path":"\/us4HARgUkkFluMIi7rOklKB0CJ5.jpg","original_language":"en","original_title":"Bröderna Lejonhjärta","genre_ids":[],"backdrop_path":null,"adult":false,"overview":"A pair of brothers who have both died, meet up in an afterlife world known as Nangijala.","release_date":""},{"vote_count":32,"id":11291,"video":false,"vote_average":6.9,"title":"The Brothers Lionheart","popularity":3.581,"poster_path":"\/1hmscSYcc2T177eXdRK0MPvGMtK.jpg","original_language":"sv","original_title":"Bröderna Lejonhjärta","genre_ids":[12,14,10751],"backdrop_path":"\/fIRnxV5zNnqTro1e6bEAyBPgHkj.jpg","adult":false,"overview":"9-year old Karl Lejon, 'Skorpan' is suffering from tuberculosis and knows that he will die soon. But his older brother Jonathan tries to comfort him by promising that they one day will meet in Nangijala, a magic country beyond the stars. Jonathan dies soon after in a house fire and Skorpan also passes away not long after. They meet in Nangijala, a country divided into the Cherry Blossom Valley and the Briar-Bush Valley. The people in the later are oppressed by the black knight Tengil and his dragon Katla. The two brothers, who now call themselves Jonathan and Karl Lion Heart, decide to help the freedom fighters against the oppression.","release_date":"1977-09-23"}]}

That work in progress title is the first one in the list. It's not obvious how to filter it out.

When doing manual searches on themoviedb, title followed by year in parenthesis (1977) yields no hit, hence you need to search movie title only which get lots of false positive. They offer a workaround to fine tune search with the suffix" y:1977", ie: "The Brothers Lionheart y:1977".

 

Could Emby, when searching a proper movie file name (title and year in parenthesis), do the fine tune search as well?

If I understand correctly, for now it ignores the year, hence some wacky results once in a while.

Edited October 7, 2018 by jlr19

[Luke 42080]

We already do incorporate the year into the search.

[Happy2Play 9782]

We already do incorporate the year into the search.

 

Testing with identify doesn't search with year.  I entered in name and year in year input boxes.  I get the same results with or without year, at least with this example.

2018-10-07 14:55:49.429 Info HttpClient: GET https://api.themoviedb.org/3/search/movie?api_key=f6bd687ffa63cd282b6ff2c6877f2669&query=The+Brothers+Lionheart&language=en
2018-10-07 14:55:49.910 Info HttpClient: GET https://private.omdbapi.com?apikey=fe53f97e&plot=full&r=json&s=The+Brothers+Lionheart&type=movie

[Luke 42080]

It doesn't go into the url, but it is still incorporated.