Jump to content


Photo

Wrong metadata using similar movie names


  • Please log in to reply
26 replies to this topic

#1 antony23 OFFLINE  

antony23

    Member

  • Members
  • 14 posts

Posted 03 January 2020 - 03:10 AM

Hi,
I'm from italy and new to emby.
I have these movie file names:
- Una notte da leoni
- Una notte da leoni 2
- Una notte da leoni 3

Metadata scraping assigns same data to all these movies, ignoring "2" and "3" char.

I don't have (year) in file name as you can see, but why "2" and "3" are ignored? Looking at imdb these titles are all exisiting in italian movie naming. So why not get right infos?

Could be this stuff fixed? I have a lot of movie names in this naming format.

Moreover, i cannot see the original filename using kodi, as the title is "Una notte da leoni" for all. So it is impossibile to me understand which of the three movies i'm going to play.

Thanks

#2 Luke OFFLINE  

Luke

    System Architect

  • Administrators
  • 156956 posts
  • Local time: 01:34 PM

Posted 03 January 2020 - 03:14 AM

Hi, look what happens when you search MovieDb:

https://www.themovie...i 2&language=it

 

This needs to be reported to MovieDb.

 

For best results please put the year in the file name. Please let us know if this helps. Thanks.



#3 antony23 OFFLINE  

antony23

    Member

  • Members
  • 14 posts

Posted 03 January 2020 - 03:55 AM

Hi luke thanks for answer.
I don't know how emby look into TMDB, but searching for "Una notte da leoni 2" (inside double quote) the match is one and correct.

Should not be used exact match before to try other search?

#4 antony23 OFFLINE  

antony23

    Member

  • Members
  • 14 posts

Posted 03 January 2020 - 04:01 AM

But thinking you are right, that should be fixed by tmdb because exact match should appear before other in results. Or maybe they don't want this behaviour?

#5 Luke OFFLINE  

Luke

    System Architect

  • Administrators
  • 156956 posts
  • Local time: 01:34 PM

Posted 03 January 2020 - 03:35 PM

But thinking you are right, that should be fixed by tmdb because exact match should appear before other in results. 

 

This definitely should happen. @Happy2Play can you report it to them? Please note, when running this search from the MovieDb website, you need to change the language to Italian. Thanks.



#6 Luke OFFLINE  

Luke

    System Architect

  • Administrators
  • 156956 posts
  • Local time: 01:34 PM

Posted 03 January 2020 - 03:36 PM

Hi luke thanks for answer.
I don't know how emby look into TMDB, but searching for "Una notte da leoni 2" (inside double quote) the match is one and correct.

Should not be used exact match before to try other search?

 

The answer to this is maybe. We've tried to do this in the past and it has led to users complaining of other problems. I'll have to dig through past notes to recall the examples.



#7 antony23 OFFLINE  

antony23

    Member

  • Members
  • 14 posts

Posted 04 January 2020 - 03:14 PM

Any news on this?

 

I have a lot of movies with this problem, think of Fast & Fourious film series, all my collection of these films have same metadata!

 

I think this is a bug, we should expect right metadata if we have exact film name match... 

 

Do you agree?

 

If TMDB does not fix it, i think emby should try first exact match (with double quote) and then other matchs.



#8 Luke OFFLINE  

Luke

    System Architect

  • Administrators
  • 156956 posts
  • Local time: 01:34 PM

Posted 05 January 2020 - 03:57 PM

 

 

If TMDB does not fix it, i think emby should try first exact match (with double quote) and then other matchs.

Again the problem is that when we've tried to do this in the past, it's actually led to some incorrect matches. I think first we should report this to Tmdb and get guidance from them. @Happy2Play Thanks.



#9 antony23 OFFLINE  

antony23

    Member

  • Members
  • 14 posts

Posted 05 January 2020 - 05:36 PM

I don't think tmdb will change its behaviour, but hope for it. I wrote them by email, but dont recive any answer yet.

If them cannot, i don't understand what could be the problem: try exact match with double quote, if the first tmdb result is egual to the exact match keeps it, else try again as already is currently done.

More,to be sure it doesn't cause quirk matchs, this exact check could be done only when there is no year specified in the movie/serie file name.

Cannot see which problem could cause, maybe i miss something.

#10 Happy2Play OFFLINE  

Happy2Play

    Trial and Error

  • Moderators
  • 19165 posts
  • Local time: 10:34 AM
  • LocationWashington State

Posted 05 January 2020 - 07:47 PM

Will do some more research but this old topic says

We don't really support any concept of exact matching at this time.

https://www.themovie...9c295564d0926aa

https://www.themovie...2514118ed00d0ba



#11 antony23 OFFLINE  

antony23

    Member

  • Members
  • 14 posts

Posted 05 January 2020 - 08:54 PM

Reading both, they say don't support exact match, but using double quotes exact match works very well, even on transleted titles.

At least on web site, cannot try via api.

Ex. searching for "una notte da leoni 2", "notte da leoni 2", "notte leoni 2" always bring to one exact movie, so on web site exact match works!

Can you try via API?

#12 antony23 OFFLINE  

antony23

    Member

  • Members
  • 14 posts

Posted 05 January 2020 - 09:38 PM

Same happen searching "Fast furious 5" vs no double quote version.

It's seeems that the number is considered only inside double quote.

It seems not anyway an exact match stuff, in fact his example "whistleblower" does not match exactly.

I think finally that using double quotes trigger such a thing that all words inside quotes are considered, versus the no double quote version.

This imho is a good thing/featured for those (like me) not having the year in the fiename.

What do you think?
Without year, this double quote query garantee a best match than without.

#13 antony23 OFFLINE  

antony23

    Member

  • Members
  • 14 posts

Posted 05 January 2020 - 11:45 PM

Looking forward, tmdb have to fix anyway something... Looking at Maleficent... There is no way without year to get exact match. That's with or without double quotes.

Double quotes help a lot in more words filename, but with an only word it fail... Weird no one has underlined this Great bug imho.

#14 scottf007 OFFLINE  

scottf007

    Newbie

  • Members
  • 3 posts
  • Local time: 04:34 AM

Posted 23 March 2020 - 10:39 PM

I thought hash's of files were used for subtitles - can't there be some hash library that is built from the millions of people using this? If they can match a subtitle (which I have not tested), why cant they match a movie name and thumbnail. 

I have this problem a lot with kids movies. 

Incredibles

Incredibles 2

 

Lion King

 

Or an option where I could curate the library and select the movie from a list would also do -> update the IMDB key and it updates everything or something. 


Edited by scottf007, 23 March 2020 - 10:44 PM.


#15 Luke OFFLINE  

Luke

    System Architect

  • Administrators
  • 156956 posts
  • Local time: 01:34 PM

Posted 23 March 2020 - 11:20 PM

I thought hash's of files were used for subtitles - can't there be some hash library that is built from the millions of people using this? If they can match a subtitle (which I have not tested), why cant they match a movie name and thumbnail. 

I have this problem a lot with kids movies. 

Incredibles

Incredibles 2

 

Lion King

 

Or an option where I could curate the library and select the movie from a list would also do -> update the IMDB key and it updates everything or something. 

 

Hi there, can you give an example of file and folder names? Thanks. What exactly is the problem that happened?



#16 scottf007 OFFLINE  

scottf007

    Newbie

  • Members
  • 3 posts
  • Local time: 04:34 AM

Posted 29 March 2020 - 08:41 PM

5e813d156e31a_Capture22.png5e813d1fca54a_Capture22z.png5e813d38ebc56_Capture22b.png



#17 Happy2Play OFFLINE  

Happy2Play

    Trial and Error

  • Moderators
  • 19165 posts
  • Local time: 10:34 AM
  • LocationWashington State

Posted 29 March 2020 - 09:44 PM

TMDB "&query=The+Incredibles+2" returns none relavate item just like there web search and The Lion King query returns the 2019 version first, but there is no year on one to distinguish what it is also, so adding 1994 to the mislabeled Lion King provides proper results. 

 

All same named movies must contain year.

2020-03-29 18:21:37.720 Info App: MovieDbProvider: Finding id for item: The Incredibles
2020-03-29 18:21:37.720 Info HttpClient: GET https://api.themoviedb.org/3/search/movie?api_key=f6bd687ffa63cd282b6ff2c6877f2669&query=The+Incredibles&language=en
2020-03-29 18:21:38.025 Info HttpClient: GET https://api.themoviedb.org/3/movie/9806?api_key=f6bd687ffa63cd282b6ff2c6877f2669&append_to_response=casts,releases,images,keywords,trailers&language=en&include_image_language=en,null
2020-03-29 18:21:38.361 Info HttpClient: GET https://api.themoviedb.org/3/movie/9806?api_key=f6bd687ffa63cd282b6ff2c6877f2669&append_to_response=casts,releases,images,keywords,trailers
2020-03-29 18:21:38.538 Info HttpClient: GET https://image.tmdb.org/t/p/original/2LqaLgk4Z226KkgPJuiOQ58wvrm.jpg
2020-03-29 18:21:38.834 Info HttpClient: GET https://image.tmdb.org/t/p/original/wiDGnsn9RtNglgKQy4J1jZQBG5v.jpg
2020-03-29 18:21:39.066 Info App: MovieDbProvider: Finding id for item: The Incredibles 2
2020-03-29 18:21:39.066 Info HttpClient: GET https://api.themoviedb.org/3/search/movie?api_key=f6bd687ffa63cd282b6ff2c6877f2669&query=The+Incredibles+2&language=en
2020-03-29 18:21:39.653 Info HttpClient: GET https://image.tmdb.org/t/p/original/2LqaLgk4Z226KkgPJuiOQ58wvrm.jpg
2020-03-29 18:21:39.877 Info HttpClient: GET https://image.tmdb.org/t/p/original/wiDGnsn9RtNglgKQy4J1jZQBG5v.jpg
2020-03-29 18:21:40.046 Info App: MovieDbProvider: Finding id for item: The Lion King
2020-03-29 18:21:40.046 Info HttpClient: GET https://api.themoviedb.org/3/search/movie?api_key=f6bd687ffa63cd282b6ff2c6877f2669&query=The+Lion+King&language=en
2020-03-29 18:21:40.356 Info HttpClient: GET https://api.themoviedb.org/3/movie/420818?api_key=f6bd687ffa63cd282b6ff2c6877f2669&append_to_response=casts,releases,images,keywords,trailers&language=en&include_image_language=en,null
2020-03-29 18:21:40.671 Info HttpClient: GET https://api.themoviedb.org/3/movie/420818?api_key=f6bd687ffa63cd282b6ff2c6877f2669&append_to_response=casts,releases,images,keywords,trailers
2020-03-29 18:21:40.788 Info HttpClient: GET https://image.tmdb.org/t/p/original/pSfwXxP94xktZKn3UaeVe6VdFZl.jpg
2020-03-29 18:21:41.202 Info HttpClient: GET https://image.tmdb.org/t/p/original/nRXO2SnOA75OsWhNhXstHB8ZmI3.jpg
2020-03-29 18:21:41.482 Info App: MovieDbProvider: Finding id for item: The Lion King
2020-03-29 18:21:41.482 Info HttpClient: GET https://api.themoviedb.org/3/search/movie?api_key=f6bd687ffa63cd282b6ff2c6877f2669&query=The+Lion+King&language=en

@Luke

Can't remember if we discussed this before but quoted searches returns one result.

https://www.themovie...&language=en-US

https://api.themovie... 2"&language=en



#18 scottf007 OFFLINE  

scottf007

    Newbie

  • Members
  • 3 posts
  • Local time: 04:34 AM

Posted 30 March 2020 - 09:26 AM

I read at the top they should contain a year - however if they subtitles are matched to a movie using a hash (according to the library settings), cant you use this to identify movies more accurately?



#19 Luke OFFLINE  

Luke

    System Architect

  • Administrators
  • 156956 posts
  • Local time: 01:34 PM

Posted 31 March 2020 - 12:57 PM

I read at the top they should contain a year - however if they subtitles are matched to a movie using a hash (according to the library settings), cant you use this to identify movies more accurately?

 

When downloading subtitles, Open Subtitles supports searching using a hash of the movie file. But that will not help with movie metadata fetching.



#20 Riggs OFFLINE  

Riggs

    Advanced Member

  • Members
  • 1005 posts
  • Local time: 12:34 PM

Posted 31 March 2020 - 08:09 PM

Is there a change in the naming convention?

 

I always used into the folder movie

 

fanart

poster

cdart

banner

etc.

 

Now Emby do the scratch and use the complete name of the movie

Star Wars Episode I - The Phantom Menace (1999)-poster

 

I just change the 1080p for 4k. Same name in .mkv and directory

 

Normally Emby take the artwork already in the directory, but now is pulling new content not needed, because it was there already.

 

Any reason of this?

 

 

My movies are curated 100%

 

One folder by movie

 

Directory capture below.

 

5e83dbc260308_dir.jpg


Edited by Handbrake, 31 March 2020 - 08:25 PM.





0 user(s) are reading this topic

0 members, 0 guests, 0 anonymous users