Just a small note. Infuse doesn’t use TVDB for any metadata anymore it uses TMDB for both movies and TV shows.
You example is kind of rare where the exact same name is used for two releases in the same year. As long as you have iCloud sync turned on all copies of Infuse on different clients will sync the correction to the right movie. You don’t have to do it on each.
The thing is that all the clients are not on the same iCloud ID, so can’t count on iCloud sync, have to be “perfect match” at the first time.
For this example, I ended up making a custom art file & xml, but it’s frustrating that being able to point out by the ID is not possible (yet ? )
Always a possibility! In the meantime, do you have any other titles that are causing this issue? I’ve run literally thousands of movies through the metadata scan and haven’t run across more than a couple of these cases. Maybe we can figure out a name that will work the first time for any others you may have.
I have the same issues with Tom and Jerry I’m setting up for my son. I’ve got all the classic shorts from 1940-1960 but Infuse can’t find the correct series, although they are all present with ID 72860 on TVDB (Tom and Jerry - TheTVDB.com). I tried naming the files all according to the Metadata 101, matching it with the TVDB (the call it series 1940, S1950 and so on), I’ve written XML files to override but that doesn’t work either after the last update. It’s stuck on a Hanna-Barbera “Tom and Jerry Show” from 1975 (separate issue I guess).
But a feature to search for the ID, or enter the ID in the XML files would be highly valued.
You’ll find that most of the group of Tom and Jerry shorts are classified as movies. For example the first one from 1940 that shows as S1940E01 on TVDB is actually the movie “Puss Gets the Boot 1940” and if you name that file “Puss Gets the Boot 1940.ext” it will find it on TMDB where all of the metadata for Infuse is now gathered from.
Changing to this format will gather the correct metadata from TMDB
The “problem” is I don’t agree with the classification and want them listed as TV series. I guess XML is my only way out then. I don’t know why they are not read properly. I select the XML file, correctly listed as an option for embedded/local metadata, instead of this incorrect “Tom and Jerry Show” but it sort of defaults back to it whatever I do. I can’t seem to clear it. I’m still working on the XML files, so maybe I need one for each episode before it “understands”?
Since Infuse no longer gathers metadata from TVDB the classification as a tv show isn’t correct. TVDB has several shows that were incorrectly classified as TV shows when they were actually movies/shorts like these.
Instead of fighting with xml files and adding artwork, changing the file name will gather the correct info with less work.
I understand that your intent was to keep these as a TV show but currently Infuse only works with xml files for movies not tv series.
There is a currently running suggestion for this here.
First let me give a big +1 here as well. For bigger collections it’s not accuarate enough and many tiltes like 1984 (1984) are not properly discovered.
I would like to through in a different approach or way to store IDs properly to the files.
In case you have a pure MKV based collection, I could imagine to have that IDs just “burned in” to the MKV file. It’s anyway a container and you can (meanwhile) easily add meta data into it. Unfortunately the schema for that has been rather “poorly” described by MKV devs and is quite vage in some details. Here you can finde the schema: https://www.matroska.org/technical/tagging.html
But why not checking for a global tag named like: “TMDB” or “IMDB” and fill it with that values. So the ID is all time “glued” to the movie. Everything else can be added in ext files or scrapping to DB or or or.
What do you thing about that proposal? At least for the sake of having MKV…
Always check the titles at TMDB (and possibly alternate titles if you don’t like what they went with) if Infuse scraping doesn’t match your files correctly.
This would be really helpful for troublesome titles that fail to scrape correctly even when your filename’s title and year are an exact match to your target at TMDB.
The thread also suggested adding support to the .nfo file, which would be much cleaner (if one wanted to use .nfo files). I think the hold-up there is Infuse doesn’t seem to read the .nfo files until after the related title is scraped from TMDB and a match is made. Files that don’t match never make the library (except as “Other”) and metadata is not overwritten by the local .nfo. Infuse doesn’t seem to have a process to either interrupt a scan of TMDB on a per-title basis due to the presence of an .nfo file (in order to look for a TMDB ID), nor to re-scan a title based on finding a TMDB ID tag inside the .nfo after a file is already matched.
I’d love to see these features added, but considering Infuse currently has a post-scan method to correct improper matches built into the UI, and those corrections are saved both in Infuse’s on-device database and (if enabled) uploaded to the cloud, it seems they’ve figured they’ve done enough.
But yes, it is frustrating for those of us who might relocate or rename their files relatively often, or can’t use iCloud, or need to delete their iCloud data periodically when making big changes to their library (or for other reasons). And frustrating in that sometimes you have no choice but to name a file by an alternate title (sometimes in a foreign language) because no other reasonable file name will scrape correctly — as is too often the case with movies with very short single word titles.
If the entire world would be only US English, maybe yes but there are other languages and movie titles around the world for the same movie.
Anyway, in my language it’s 1984 (1984) officially from TMDB, nothing wrong about it. I guess we can all agree, there is no accurate way of scrapping titles. Only identifiers could solve that issue once and for all, right?
I’m not arguing your point; I’m just saying if you spell it out, you won’t have problems with it scraping incorrectly. Infuse/TMDB scraping is most inaccurate when titles are one word long, and shorter words all the more so. Titles that are solely numbers are absolutely the worst.
Identifiers would indeed be nice. But they aren’t yet supported; so I was trying to help you work around the issue in the meantime. Do what you will.
[Edit to add]
As seen in the second post to this thread, Infuse support for identifiers in filenames is (potentially) being considered. [I suppose the fact this thread is marked ‘planned’ also indicates the same … lol]