Yuppp!!! Lol
Bitofinger:
The prospect of allowing a “smart” program to configure this information is scary, to say the least. I’ll test it in small doses.
LOL !! You ain’t kidding about that. I spend more time checking the work of these renamer programs than the renaming process itself. It takes just seconds to scan/match(find) online/and rename tens of thousands of files and a headache to check the work.
I’m not really worried about the 2-20kb posters taking up space when im already sitting here with 24 4TB drives of stuff… the whole of the metadata folder for my server takes up less than 4GB of a 96TB server.
As I’ve displayed with the test library, its doing exactly what you want – listing all the files individually by file name. I’m not sure what else I can do to change it, other than just say dont name things identically. It didn’t download any metadata, didn’t rename anything, didn’t merge anything… I’m sorely confused by how or why yours is.
I am.
There is no need for this on top of the space requirements.
there kind of is… Since plex can’t discern whether you have the DVD box set, Bluray box set, special edition platinum edition, etc etc etc, it grabs them all. and the reason it grabs ones from different languages is because they’re not differentiated in the host - they’re just listed like:
Neon Genesis Evangelion Poster 1
Neon Genesis Evangelion Poster 2
Neon Genesis Evangelion Poster 3
not
Neon Genesis Evangelion Poster DVD ENUS
Neon Genesis Evangelion Poster DVD ENEU
Neon Genesis Evangelion Poster BluRay ENUS Platinum Edition
so yes, some do have it that way, but that’s on the sites home page. the actual API used to fetch them doesn’t distinguish them, and not all hosts are the same – some don’t even do it internally and they’re forced to adhere by a standard set that also doesn’t serve the region.
When you fetch them by the API, all you can do is say “i want the posters for show xxxxxx” and the API returns an array of files with no other information.
basically like this: https://www.thetvdb.com/series/jojos-bizarre-adventure-2012/artwork/poster
So yes, PLEX could theoretically scrape each and every page and find out which one is which, but this would take a LOT of resources from both your machine AND the moviedb servers… OR… it could just download an extra couple 15kb images and say bugger it, use the default one and if its not right, they can hit edit and pick the right one.
BTW: scraping sites for info like that is a legal grey area, and they can technically sue you for undue bandwidth, which is also one of the reasons you’re HIGHLY RECOMMENDED to just use the public API they set up instead.
edit: gimme a sec, explaining further…
When you add a show to plex, it asks which “Agent” to use (which databases API).
You select the API you want, either themoviedb or plextv or whatever.
That API sends a search request for “I have a file called Deadpool, what movie ID is this” and the API returns an ordered list of movies, sorted by relevance. For example it might return Deadpool AND Deadpool 2 AND Dead Pool AND Pool of the Dead. Plex then picks the top one (most relevant according to the API) and stores that movies “ID Tag” that the API host gave so it can then move on tot he next step
The API then sends a Fetch Metadata request to the api for the movie by the tag the search function gave it, that way it knows its getting the right movies information.
The API then returns ALL the metadata for the movie, in ALL languages, and Plex then stores the ones from the language you wanted.
The API then sends a Fetch Images request, fetching all the posters/banners/backgrounds for the movie, and the API returns ALL of them as an array. ALL the posters as one array, ALL the banners as another array, ALL the background as another array.
This is not something PLEX can do anything about, this is all about a standard set up by the IEC (International Electrotechnical Commission) for sharing media metadata in a STANDARDIZED way that all software can accomodate with as little knowledge storable by the host as possible – this means you fetch the data about a movie – the host cannot discern what language or region you are looking for, and store that for later use by selling it to advertisers for example (a common problem in the early days of metadata hosts, being able to discern which music was more popular in which regions by who was polling for which languages metadata)
So I’m sorry your server needs an extra gig or two of harddrive space, but you’re already working with software that regularly requires a multi-bay NAS to hold double-digit terabytes of data. If you really don’t want the extra posters, go into the metadata folder and delete the unused ones. Once it fetches them once, it won’t again unless you delete all of them and it has no posters at all, or you manually tell it to refetch them.
Nope. It definitely does not do absolutely nothing. I HAVE chosen “Personal Media”. That’s why I’m so upset. I think it should do absolutely nothing for personal media, unless I ask it to. I’ve explicitely asked it to do absolutely nothing for my personal media. Did you miss that part? I’ve done everything you say. It still merges my files (and no I don’t mean it changes the files on disk. I mean it shows different files as being the same thing and then I can’t sort and view them separately, which is INFURIATING.)
Why can’t you just listen to the countless requests to add an option to DO NOTHING? At least to turn of the merging?
And why not fix your bug that does do what YOU say it should not be doing for “Personal Media”? Or can you at least be consistent about what it is intended to do?
Finally, yes I’m pissed that this continues to be a problem and the devs are being too defensive in their obstinance rather than implementing a simple option. Please, consider that. But until then, it sounds like I need to put sufficient dummy fields into the filename such that I consume all the “smart fields” it wants to interpret?
For reference…
Can you confirm that this is my problem?
YYYYMMDD-HH:MM:SS-pandas-156.avi
is interpreted as
Series - Season - Episode(?) - Title?
OR
Series - Season - Episode - Part?
OR
Series - Episode - Title - Part?
OR
Series - Episode - ?
If multiple of these interpretations are valid, what happens? Randomness? Seems to be what I have.
Can “HH:MM:SS” be interpreted as integer? Does that get split at all? Do spaces matter, or is it alwasy splitting on “-”?
Can you confirm that this scheme will do what I hope and solve my problem for now?
Series - Season - Episode number - Title - Part?
YYYY - MMDD - HHMMSS - Location - Index
If I have that format, and it still thinks “Index” is a part number (it is an integer) is that still going to confuse it? If I have consecutive part numbers, but they are on different episode numbers, will it still merge? Do part numbers need to be consecutive or starting at 1?
Do you begin to see why this is confusing and frustrating? I think it works well, I really do, when used against “Properly named” files that are actually movies or TV shows.
It sucks balls when it’s being applied to personal videos! I just want to turn it off for that library. PLEASE.
Alternatively I could do
Series - Season - Episode - Title
YYYY - MMDD - HH:MM:SS:Index - Location
If you think that would work more predictably?
again… im not sure how you and Bitofinger are having this merging issue…
i set up a Other Video → Agent: Personal Media library and it just does this…
The only time i had it “merge” two of them was if I used the same file and just renamed it, because the bitmask hash was the same so it knew both files were the same thing. Using different base videos for each and just renaming them… nothing went weird…
You wrote alot there and I’m not sure what to quote but, without saying your wrong.
Here is a good quote…
PLEX can and does see the language.
Which will also explain why you, me, and everyone else gets the right poster when we add content. If PLEX didn’t then every user out there would be getting random posters not of their language.
ehhh… kinda.
When it gets the metadata from the host, like thetvdb, there’s a “most popular picture” tag, which basically comes in as an index number.
“We sent you 8 pictures, most people voted index 4 as most popular”. However, it doesn’t tell plex which language that picture is, or what version it is. It still has to download all of them to have the correct indexing, and again, if you have the platinum edition, some people want to have the box art matching their actual box art, and not the box art from the korean bootleg version with a naked lady on it because 12 people decided they like that one best because nakeds.
It’s a matter of IEC standards, ease of use, and practicality. Nobody really cares about an extra 100kb of data on a multi-terabyte media server, and the information sent just isn’t there for plex to know if its correct or not – it’s just assuming based on the information sent by the API that it has the right thing – thats why it LETS you select different posters, and LETS you say “fix match…” because it occasionally gets it wrong.
When I first imported Record of Grancrest War, it came up with the japanese poster instead of the english one. I had to change it. and when I imported Rouroni Kenshin, it came up with the Samurai X poster. While yes they’re fromt he same studio, use a lot of the same footage out of lazyness and expense, they’re not the same show. (It’s more akin to Fullmetal Alchemist vs Brotherhood… same story for most of it, but one follows the manga more)
edit:
Okay, seems thetvdb and themoviedb both have their APIs set to push extra language info – but that’s not standard IEC information, so I can’t say whether plex is setup to look for that info or not, since none of the other hosts seem to push language IDs. Either way, it still needs to download them all to get the indexing correct, and it doesn’t help with the rest of the problems (box arts for different versions of the show/movie)
Here you go. Prime example not only of the inconsistency users have to overcome, but also of the stupidity.
---- Liestenn plex@discoursemail.com wrote:
The second one looks fine, the first one looks absolutely *----*'d.
Mind mousing over one of them and clicking the pen icon, then going to Info, and taking a screenshot?
I just wanna see how they’re named –
example:
After reading through this thread again I have to say that although I cannot say what it is many people are clearly doing something wrong in the way they are setting up libraries or agents or other plugins.
I have never had even one file, and I have thousands, that are set as “Other videos” using “Personal media” mismatch or combine even once.
Basically “Other videos” using “Personal media” never even tries to match anything so there is no way such a library could have any “combined” files.
My guess is that the libraries involved are a movie type library or has something other than “personal media” as an agent or had such at one time (changing agents after the fact does not fix previous matches) or there are plugins that are interfering in some way with the matching process.
Once again a library of “Other videos” with “Personal media” as an agent will never combine videos as the mechanism is just not there.
I disagree with this. Many post are about this very thing. And I truly believe we could minimize this issue if we could save the metadata info on outside storage(shares, NAS, etc) instead of our SSD’s that have a very limited amount of space.
EDIT, the space for that movie is 1.87MB and not 100kb.
It would HAVE to if you think about it. There is no way around it. There is no way PLEX is that lucky in picking the right language poster every time. 23 languages and up to 23 possible language specific posters. But I guess it’s all moot at this point.
Anyway, I do agree with @Elijah_Baley. I believe the library type is not OTHER VIDEOS
I see they are combining the videos together. That does suck.
Unfortunately I can’t. I was so fed up last night with that database that I deleted it and started again. It appears to have straightened itself out in the rescanning.
what is this “bitmask hash” where does it come from and what does it do? I think I might suspect that as the source of my woes.
can you give me any info on that?
what happens if a file is first created and then grows? is plex maybe glomming onto the empty file and keeping that even after the file becomes not empty?
the private video section appears to be using the file contents to match identical files. so it you have a ton of 0-bit size files, they’ll become “this is another version” … just split them. click the […] button and press split. itll as “do you REALLY wanna do this?” … click yes.
A bitmask hash is those hashes you see on download files like github to make sure the file is what it says it is, regardless of file name. If plex does a quick bitmask check on the file to see what it is, and the file returns Bitmask:[000000000000] because there’s no data, oh look, all those files have the same hash and therefor must be the same file, just renamed.
I don’t know WHY it does this, but just set plex to rescan once it detects a change in the folder and it should sort itself out after a while (set it to full library scan, not quick scan, or it wont touch already known about files that still exist)
