Incorrect identification: not using year

[computerprep 115]

I very frequently have misidentified movies. And I've always just corrected them manually. And it's always stumped me that when I manually identify the movie, with the correct title and year in their correct fields, it almost never fails to return the correct match as the first choice.

So today I finally looked into my server logs for the most recent misidentified movie, "Go (2020).mp4" Here's what I found.

2020-03-13 09:22:33.724 Info App: MovieDbProvider: Finding id for item: Go (2020)
2020-03-13 09:22:33.724 Info HttpClient: GET https://api.themoviedb.org/3/search/movie?api_key=????&query=Go+(2020)&language=en
2020-03-13 09:22:34.031 Info HttpClient: GET https://api.themoviedb.org/3/search/movie?api_key=????&query=Go&language=en
2020-03-13 09:22:34.106 Debug App: Running OmdbItemProvider for \\????\Share\Movies\Go (2020).mp4
2020-03-13 09:22:34.107 Debug App: Running BaseGenreCleaner for \\????\Share\Movies\Go (2020).mp4
2020-03-13 09:22:34.107 Debug App: Running MovieDbImageProvider for \\????\Share\Movies\Go (2020).mp4
2020-03-13 09:22:34.108 Info HttpClient: GET https://image.tmdb.org/t/p/original/kr1nz11EILLPdRmn51jtvvBQrWL.jpg

It appears that the first API call includes the year inside the title query field. When it doesn't find a match, it removes the year entirely, which then matches to the movie "Go (1999)"

But TMDBAPI allows the GET to include a year=integer. With a properly formatted filename, is there a reason the second GET request isn't the following instead?

GET https://api.themoviedb.org/3/search/movie?api_key=????&query=Go&year=2020&language=en

Adding the year produces perfect results in all of my tests with other mistaken identity files.

[computerprep 115]

Lol, just to highlight how often this type of thing happens, another movie was misidentified only 1 hour before Go (2020). This time Redemption (2020).mp4 was identified as The Shawshank Redemption (1994). Because Shawshank's popularity is so much higher, it overrides Redemption as the first hit unless 2020 is properly formatted into the year spot of the API query string.

[Luke 25393]

We handle the year ourselves to allow for a one year tolerance. So the filter is still used.

[computerprep 115]

Both of these examples are far greater than a single year tolerance.

[Luke 25393]

Complete log file? I don't see any evidence that it was matched as the 1994 title.

 

@@Happy2Play

@@arrbee99

 

Can you reproduce with this example?

[arrbee99 715]

Assuming I've done the right thing - restarted server and found a mp4 file, renamed it to Redemption (2020) and added it to Movies - it seems to have become Shawshank...

 

 

Using Widows, latest beta.

embyserver.txt

[computerprep 115]

Complete log file? I don't see any evidence that it was matched as the 1994 title.

I just rotated the log file, added a fake file with the filename "Redemption (2020).mkv" and scanned the library before retrieving the following lines from the log

 

2020-03-19 02:59:29.644 Info App: MovieDbProvider: Finding id for item: Redemption (2020)
2020-03-19 02:59:29.644 Info HttpClient: GET https://api.themoviedb.org/3/search/movie?api_key=f6bd687ffa63cd282b6ff2c6877f2669&query=Redemption+(2020)&language=en
2020-03-19 02:59:29.832 Debug App: Throttling Tmdb by 112 ms
2020-03-19 02:59:29.951 Info HttpClient: GET https://api.themoviedb.org/3/search/movie?api_key=f6bd687ffa63cd282b6ff2c6877f2669&query=Redemption&language=en
2020-03-19 02:59:30.093 Debug App: Throttling Tmdb by 157 ms
2020-03-19 02:59:30.257 Info HttpClient: GET https://api.themoviedb.org/3/movie/278?api_key=f6bd687ffa63cd282b6ff2c6877f2669&append_to_response=casts,releases,images,keywords,trailers&language=en&include_image_language=en,null
2020-03-19 02:59:30.318 Debug App: Running OmdbItemProvider for \\????\Share\Movies\Redemption (2020).mkv
2020-03-19 02:59:30.318 Info HttpClient: GET https://private.omdbapi.com?apikey=fe53f97e&i=tt0111161&plot=short&tomatoes=true&r=json
2020-03-19 02:59:30.469 Debug App: Running BaseGenreCleaner for \\????\Share\Movies\Redemption (2020).mkv
2020-03-19 02:59:30.469 Debug App: Running MovieDbImageProvider for \\????\Share\Movies\Redemption (2020).mkv
2020-03-19 02:59:30.469 Debug App: Throttling Tmdb by 87 ms
2020-03-19 02:59:30.563 Info HttpClient: GET https://api.themoviedb.org/3/movie/278?api_key=f6bd687ffa63cd282b6ff2c6877f2669&append_to_response=casts,releases,images,keywords,trailers
2020-03-19 02:59:30.606 Info HttpClient: GET https://image.tmdb.org/t/p/original/zGINvGjdlO6TJRu9wESQvWlOKVT.jpg
2020-03-19 02:59:31.738 Debug ProviderManager: Saving image to C:\Emby\Metadata\metadata\library\f0\f0db287d378623f2976d9460175ddbfe\poster.jpg
2020-03-19 02:59:31.739 Info HttpClient: GET https://image.tmdb.org/t/p/original/65nAfGikljrOkfIrcfmrNTJTAXF.jpg
2020-03-19 02:59:31.797 Debug ProviderManager: Saving image to C:\Emby\Metadata\metadata\library\f0\f0db287d378623f2976d9460175ddbfe\backdrop.jpg
2020-03-19 02:59:31.798 Debug App: Running FanartMovieImageProvider for \\????\Share\Movies\Redemption (2020).mkv
2020-03-19 02:59:31.799 Info HttpClient: GET https://webservice.fanart.tv/v3/movies/278?api_key=5c6b04c68e904cfed1e6cbc9a9e683d4
2020-03-19 02:59:32.663 Info HttpClient: GET https://assets.fanart.tv/fanart/movies/278/hdmovielogo/the-shawshank-redemption-503e5a463d35c.png
2020-03-19 02:59:33.835 Debug ProviderManager: Saving image to C:\Emby\Metadata\metadata\library\f0\f0db287d378623f2976d9460175ddbfe\logo.png 

Edited March 19 by computerprep

[Luke 25393]

Please attach the complete log file as that is only a small snippet. thanks.

[computerprep 115]

server log.txt

[Luke 25393]

Ok thanks this is only happening on year 2020+.

[computerprep 115]

This has always happened on various years, I just didn't have the time or energy to pursue it before.

This literally just happened in the past 30 minutes to a 2015 movie. Here's a log file.

Filename: Bad Blood (2015)

Identified as: Batman Bad Blood (2016)

embyserver-63739092364.txt

[Luke 25393]

Yea it's due to the one year tolerance. It's something that can be improved in future updates. thanks.

[computerprep 115]

It's been a while since I've played with TheMovieDb's API, but an exact match on title with a 1 year tolerance on release date would be helpful.

Possibly with a switch in library advanced settings to control how strict the identification matching should be.

Personally, I would prefer no match to an incorrect match. No match would at least prompt me to manually identify or rename the file with correct information.