Sure, but since it’s a movie library, it doesn’t matter what TheTVDB says.
Huh, IMDB has it as a TV series, but I don’t know if we would get that from IMDB so checking TheMovieDB, it does have it listed as a movie. So this one should work. 
This raises an interesting question. I have several hundred old silent films. Harold Lloyd, Charlie Chaplin, George Melies, Fatty Arbuckle, and so on.
If this new scanner can’t see this films, what will it do when it seens them? Can it move on to the previous method of scanning? Or will I have to manually fix them? Just curious.
There’s tons of shorts listed as regular movies in The Movie Database etc.; this particular short was only available on IMDb and there it was listed as a different media type (Short vs. Movie). Kind of the oddities users occasionally get with TV Movies that aren’t listed on The Movie Database but in IMDb.
1 Like
Congratulations on this new development. I share my opinion according to my scenario, I hope it is useful.
MY LIBRARY
I use the ViMediaManager program that gets the data from the TMDB and IMDB movies at the same time. The title and plot is taken from TMDB and the other data is taken from IMDB. And when a title does not exist in TMDB it takes all the information from IMDB.
ViMediaManager creates a .nfo file, a .tbn cover image, and a .jpg background image
I rename the folder that contains the files using the command T ( Y), which is the TMDB title and year, except when the title does not exist in TMDB, so the title and year is from IMDB.
All my files are in folders renamed according to the title and year of the film. And there are films that are called the same although they were produced in different years, for example:
Our Very Own (1950)
Our Very Own (2005)
I don’t rename video files, I just rename the folder. So inside each folder my files look like this:
Our Very Own (1950)
Our Very Own 1950.tbn
Our Very Own 1950.nfo
Our Very Own 1950-fanart.jpg
Our Very Own (2005)
very-vh.tbn
very-vh.nfo
very-vh-fanart.jpg
The .nfo file contains the IMDB ID and the TMDB ID like this:
<id moviedb = "IMDb"> tt0042819 </id>
<id moviedb = "TMDB"> 97332 </id>
And it doesn’t include a URL to those sites.
The entire .nfo looks like this:
<?xml version="1.0" encoding="UTF-8"?>
<movie>
<title>Our Very Own</title>
<sorttitle>Our Very Own</sorttitle>
<originaltitle>Our Very Own</originaltitle>
<id moviedb="IMDb">tt0042819</id>
<id moviedb="TMDB">97332</id>
<year>1950</year>
<releasedate>1950-07-27</releasedate>
<rating>7</rating>
<votes>2</votes>
<certification>Argentina:Atp</certification>
<certification>Belgium:KT/EA</certification>
<certification>Finland:S</certification>
<certification>France:Tous publics</certification>
<certification>Norway:A</certification>
<certification>United Kingdom:U</certification>
<certification>United States:Approved</certification>
<certification>United States:Not Rated</certification>
<certification>West Germany:12</certification>
<tagline>SAMUEL GOLDWYN'S unforgettable, compelling entertainment triumph!</tagline>
<taglines>
<tagline>SAMUEL GOLDWYN'S unforgettable, compelling entertainment triumph!</tagline>
</taglines>
<outline>Gail discovers the shocking news that she is adopted during a heated argument with her sister, Joan. With the reluctant support of her adoptive parents and baby sister, Penny, Gail goes in search of her biological mother and true identity...</outline>
<plot>Gail discovers the shocking news that she is adopted during a heated argument with her sister, Joan. With the reluctant support of her adoptive parents and baby sister, Penny, Gail goes in search of her biological mother and true identity.</plot>
<runtime>93</runtime>
<genre>Drama</genre>
<genre>Romance</genre>
<language code="en">English</language>
<studio>Samuel Goldwyn Productions</studio>
<studios>
<studio>Samuel Goldwyn Productions</studio>
</studios>
<country>United States of America</country>
<country>USA</country>
<director>David Miller</director>
<credits>F. Hugh Herbert</credits>
<camera>Lee Garmes</camera>
<sound>Victor Young, Fred Lau</sound>
<actor id="31844">
<thumb>Our Very Own 1950.tbn</thumb>
<fileinfo>
<streamdetails>
<video>
<codec>MPEG-4 Visual</codec>
<width>640</width>
<height>480</height>
<aspect>1.33</aspect>
<duration>93</duration>
<durationinseconds>5562</durationinseconds>
<scantype>Progressive</scantype>
<bitdepth>8</bitdepth>
</video>
<audio>
<codec>MP3</codec>
<channels>1</channels>
<bitrate>105</bitrate>
<bitratemode>VBR</bitratemode>
<size>71819664</size>
</audio>
</streamdetails>
</fileinfo>
</movie>
MATCH
By having each title perfectly identified with an NFO file containing IMDB ID and TMDB ID I would expect Plex to have the ability to accurately and seamlessly identify all of my titles but it is not. Because the match doesn’t do it by getting the IDs inside the .nfo file, but using the title and year of the folder name or, even worse, the shapeless name of the file even if the folder has title and year.
Plex relies too heavily on TMDB for having a free api and that causes the problems. When a title does not exist in TMDB the scanner does not know how to recognize it. And associates it wrongly even if the NFO has the IMDB ID.
Having almost all my titles in folders with the title and the year according to TMDB, the Plex scanner works almost fine, but the problem comes when it must recognize titles that do not exist in TMDB. There the association fails.
So the only option to solve the problem has been to create the titles in TMDB that do not exist adding the IMDB ID to achieve the identification, which is not always possible. The correct thing would be for the Plex scanner to associate the IMDB ID without relying on TMDB using the information contained in the .nfo file
SPEED
In my experience the speed of identification of the titles is good. On average, it identifies 28 titles per minute, which means that the scanner has an average speed of 1680 titles per hour or 40,320 titles per day per user.
POSTERS
There is a problem with the posters. And it is again due to dependence on TMDB. In TMDB many movies do not have posters, because the site moderators delete, without an objective criterion, all the posters that they do not like because according to their subjective criteria “it is better that a film does not have a poster than it has an ugly poster”.
In Letterboxd, which like Plex depends on the TMDB api, a cached copy of the posters that are deleted in TMDB is generated. So Letterboxd is a more reliable source for finding posters that no longer exist in TMDB. Maybe the scanner could run a script to take the poster from the Letterboxd cache when it can’t find it in TMDB and thus solve the problem (the TMDB ID is associated with each title in Letterboxd: http://letterboxd.com/tmdb/TMDB_ID/ and thus the identification can be obtained).
In addition to adding the option to search the poster at Cinematerial.com.
Although the definitive solution to the problem of posters would be for Plex to create its own database of posters and for each user to be able to upload their posters from their library from the Plex app to that independent database in order to always synchronize them, which is what is done in DVDpedia with the Doghouse site (https://doghouse.bruji.com).
YEAR
There is a problem with the year. The scanner uses the year the film was released in the US and not the year of production. So a movie like Ana and the wolves that was produced in 1973 appears with the date 18 May 2018 because it is the date it was released in the US (45 years later). The correct thing would be for the year to be the year of production or the year of release in the country where they were produced.
I hope these comments are helpful and if you need additional information please let me know.
2 Likes
Thank you @robot2xl for your detailed feedback. I will wait for @drzoidberg33
to confirm, but we should already support reading the IMDB ID out of the NFO file in this new agent. We have a regex in PMS that looks for tt and then 7 to 11 numbers after it. Do you have an example of something that should have matched based on that IMDB GUID but didn’t?
I have raised an issue internally to discuss this. I have also seen user requests here on the forums to integrate with https://theposterdb.com/ when they have an API which I have also put forward for discussion.
We have some smarts to look at the filename and the foldername to help us determine the correct title and year. if you have some examples where you don’t have an NFO file and the match was incorrect, I would love to know so we can investigate and improve our filename parsing and matching logic.
We have a backend that merges both IMDB and TMDB metadata for a movie so we can match if a movie exists on either of those databases.
These are some examples of movies that had a bad match for not existing in TMDB, even when they have IMDB ID.
When there is no TMDB the NFO is created with the code:
<id moviedb="IMDb">tt1718816</id>
<uniqueid type="IMDb" default="true">tt1718816</uniqueid>
IMDB title:
Folder:
Prostitución y sida - Mujeres de la calle (1993)
File:
Prostitución.y.sida.-.Mujeres.de.la.calle.(1993).720p.H264.AAC.mkv
NFO File:
<title>Prostitución y sida - Mujeres de la calle</title>
<sorttitle>Prostitución y sida - Mujeres de la calle</sorttitle>
<originaltitle>Prostitución y sida - Mujeres de la calle</originaltitle>
<id moviedb="IMDb">tt3107214</id>
<uniqueid type="IMDb" default="true">tt3107214</uniqueid>
<year>1993</year>
<genre>Drama</genre>
<language code="">Spanish</language>
<studio>Lo Nuestro en Video SA</studio>
<studios>
<studio>Lo Nuestro en Video SA</studio>
</studios>
Result obtained wrong: Reservoir Dogs 1992
IMDB title:
Folder:
La prisión del tiempo (1991)
File:
La.prisión.del.tiempo.(1991).720p.H264.AAC.mkv
NFO File:
<title>La prisión del tiempo</title>
<sorttitle>La prisión del tiempo</sorttitle>
<originaltitle>La prisión del tiempo</originaltitle>
<id moviedb="IMDb">tt11443630</id>
<uniqueid type="IMDb" default="true">tt11443630</uniqueid>
<year>1991</year>
<certification>Mexico:B</certification>
<genre>Thriller</genre>
<language code="">Spanish</language>
Result obtained wrong: The Time Game 1992
IMDB title:
Folder:
La ley de Dios (2005)
File:
La.ley.de.Dios.(2005).720p.H264.AAC.mkv
NFO File:
<movie>
<title>La ley de Dios</title>
<sorttitle>La ley de Dios</sorttitle>
<originaltitle>La ley de Dios</originaltitle>
<id moviedb="IMDb">tt1718816</id>
<uniqueid type="IMDb" default="true">tt1718816</uniqueid>
<year>2005</year>
<genre>Action</genre>
<language code="">Spanish</language>
<studio>La Raza Mex Films</studio>
Result obtained wrong: Dirty 2005
IMDB title:
Folder:
Si callarás a la prensa (2005)
File:
Si.callarás.a.la.prensa.(2005).720p.AAC.mkv
NFO File:
<title>Si callarás a la prensa</title>
<sorttitle>Si callarás a la prensa</sorttitle>
<originaltitle>Si callarás a la prensa</originaltitle>
<id moviedb="IMDb">tt1773044</id>
<uniqueid type="IMDb" default="true">tt1773044</uniqueid>
<year>2005</year>
<genre>Action</genre>
<language code="">Spanish</language>
<studio>La Raza Mex Films</studio>
Result obtained wrong: Turistas 2006
IMDB title:
Folder:
Cobrador de la Mafia (2013)
File:
Cobrador.de.la.Mafia.(2013).720p.H264.AAC.mkv
NFO File:
<movie>
<title>Cobrador de la Mafia</title>
<sorttitle>Cobrador de la Mafia</sorttitle>
<originaltitle>Cobrador de la Mafia</originaltitle>
<id moviedb="IMDb">tt12198136</id>
<uniqueid type="IMDb" default="true">tt12198136</uniqueid>
<year>2013</year>
<genre>Drama</genre>
<language code="">Spanish</language>
Result obtained wrong: The Corridor 2013
All my files always have a .nfo file.
PROBLEM WHEN CREATING LIBRERY
An additional problem is that the new library cannot be created using app.plex.tv, since it does not allow “save” (Your changes could not be saved). It can only be created using 127.0.0.1:32400 in the URL.
i don’t care so much about properly matching this short movie. my issue is that it got matched/tagged with Get Out. That should not be happening.
It is not a big deal but it is annoying. I just wanted to know why it got matched up with the wrong movie and how to prevent it in the future
Is there any mitigation that i can do on my front to address this mismatched movie. about all i see is the ability to delete the file which i don’t want to do. For multiple matches to a movie, there does not seem to be a means to select a specific file and do that whole unmatch process.
so
a) just wanted to flag this for the mismatching to see if you guys could confirm and then address with fixes to the movie matching code.
b) see what i might do on my end to help out the current code base
i guess i could try to rename the short and go back to that period separator method for naming things in this case and see if that helps. that is a shot in the dark for me.
i had this occur myself, but then I realized the library NAME was a duplicate (the error ‘your changes could not be saved’ is not very helpful).
I changed the library name, ie movies-test, and then I was able to save, I use only app.plex.tv (unless internet is out of course).
We definitely use the IMDB id from the nfo file and therefor those should match perfectly.
Based on the examples you posted it just looks like we don’t have those in our database and therefor are falling back to searching via filename. We’re busy discussing this and whether it makes more sense to not match at all in these scenarios.
Thanks for your detailed info on the mismatches this is super helpful!
You can create ones in the hosted app by setting the scanner to Plex Movie and the agent to Plex Movie (preview) in the advanced tab. It is however highly recommended to use the bundled version when testing this build.
I seriously wonder why people think of re-installing the server to be the one-size-fits-all solution for almost every problem 
Re-installations might help if something is really messed up – but that’s about it.
1 Like
this didn’t seem to work either, neither 2001 or 2004 matched as a movie with the new agent.
/data/media/movies/New Crouching Tiger, Hidden Dragon (2001)/New Crouching Tiger, Hidden Dragon (2001).mkv
/data/media/movies/New Crouching Tiger, Hidden Dragon (2004)/New Crouching Tiger, Hidden Dragon (2004)-part1.mkv
/data/media/movies/New Crouching Tiger, Hidden Dragon (2004)/New Crouching Tiger, Hidden Dragon (2004)-part2.mkv
its a minor issue for a rare situation.
after a night worth of pulling metadata, it appeared to finish up with an other than the above exceptions 100% match rate, but that I what I expected given all my movies are named proper with movies/moviename (year)/moviename (year) [extrastuff].ext
As far as I can see/remember, my only issue is with how the matching/metadata process BLOCKS all other libraries from being rescanned, thus preventing any other new content from being added during long matching/metadata retrieval sessions.
the same thing happens with large music library scans.
the scan itself, not that bad, but the matching/metadata takes forever (days), during which plex will not add any other new media.
yup, I got the movie, which is awesome btw, it matches normally.
its just the awkward situation of the tv series
We have a regex in PMS that looks for
ttand then 7 to 11 numbers after it.
As there is no native export support of NFO files in Plex, I download XML files from Plex into every movie and tv folder. These XML files do contain IMDB IDs as well. Will Plex recognize it’s own XML files? And If not, will it work If I do rename your XML files to *.NFO?
Just to make it clear: These XML files are fetched thru official Plex API and copied unmodified into the movie/tv folder.
Currently:
Movie (2020)/Movie (2020).mkv
Movie (2020)/Movie (2020).xml
Movie (2020)/folder.jpg
Movie (2020)/fanart.jpg
No. Those XML files are actually generated from the database. These are used so the database doesn’t have to be read every time. There is no native way to go backwards and populate the database with these files.
Just a scan for the ttnnnnnnn pattern, nothing else.
Not from the XML file. The scanner isn’t expecting an XML file. Only an nfo file will be scanned for the imdb id. Might be an interesting feature request to have the scanner also look at xml files.
2 Likes
Hello, i have a bad match on the following movie: Zoo (2018) [tt3688342]/Zoo (2018) WEBDL-1080p.mkv
I use Filebot & Radarr for renaming which use TMDB. Even using the IMDB id in the folder name, the match has been made on the following id: tt7624934
We don’t currently use any IDs in the filename only if an IMDB id is stored inside an nfo file in the movie’s folder.
Just on the subject of .nfo’s
Can i ask regarding the naming of them… Will both having the bare unnamed .nfo there AND also naming them to match the filename work? If that makes sense.
As a similar example, posters can be either
poster.jpg or
A Bug’s Life (1998) - Remux-1080p [AVC DTS 6.1].jpg
So is it the same with with .nfo’s? ie
A Bug’s Life (1998) - Remux-1080p [AVC DTS 6.1].nfo
