this view could not be built (2 Viewers)

[Lehmden]

Then this has changed sometimes but not to the better. I've tested this as I tried out MP2 in the beginning but never touched this again.

I think it is not useful to throw all kinds of video in a single share (folder). I never would do this. At least a minimum of order inside the media collection can be expected from everyone, no matter how lazy he/she is... I think in general we should not allow both types (or many more in future) of videos in the same share...

 

[breese]

I think it is not useful to throw all kinds of video in a single share (folder).

Agreed BUT....
f:\Videos
f:\Videos\Movies
f:\Videos\Series
Would be valid and the Importer Pointed towards f:\Videos would find everything in every folder from f:\Videos on thru...

Now, what the result of where eveything ends up in MP2 is the BIG question if everything is NOT named / labeled / etc as per the current importer

 

[JSchuricht]

MJGraf, I like you idea of looking for a sub-string in the form ttxxxxxxx. My movie structure was originally setup for MP1 with Moving Pictures which I believe looked for a nfo or txt file with the tt number so this may become useful for other people looking to convert.

I have done some testing with tinyMdiaManager to make new nfo files and files already imported to videos don't reimport to movies when adding a new nfo. Correct me if I am wrong but the only way MP2 is going to reimport everything is by changing the file name or deleting the DB and starting from scratch. I think changing the file names is potentially more dangerous than anything else especially if I have to revert so my plan for now is copying 30TB of movies from the primary server to some scratch space. Run TMM to create new nfo's, may skip this step if the scrapper is modified by then. Run a full import on a MP2 test server, then copy watched flags over with MP2 edit. Then I can check the two known movies with import issues and compare to my current MP2 installation. If all goes well then copy the new nfo files to the primary server and transfer the MP2 DB from the test server to the live MP2 server.

 

[MJGraf]

I think it is not useful to throw all kinds of video in a single share (folder). I never would do this. At least a minimum of order inside the media collection can be expected from everyone, no matter how lazy he/she is... I think in general we should not allow both types (or many more in future) of videos in the same share...

I have to admit, I don't agree on that one. people like you and me and probably most people in this forum love their media collection and spend way too much time in making it "nice" (folder structure, tagging, etc.). But most people out there nowadays don't want to be bothered with that. They just have media files, don't care about where they are on their computer but expect the software they use to take care of this mess and automatically give it a structure they understand. I know a lot of people like that and we should make MP2 take care of them as well.

My movie structure was originally setup for MP1 with Moving Pictures

I didn't know that MovPic is able to handle these kinds of nfo-files. Since it does, we should make MP2 handle them as well...

Correct me if I am wrong but the only way MP2 is going to reimport everything is by changing the file name or deleting the DB and starting from scratch

In general you are right. But you don't have to change the file name. Just use a tool to set the "last modified date" of the files in question to "now" and start a refresh import from Settings. It will re-import all files, the last-modified-date of which is newer than the date of the last import.

my plan for now is copying 30TB of movies from the primary server to some scratch space. Run TMM to create new nfo's, may skip this step if the scrapper is modified by then. Run a full import on a MP2 test server, then copy watched flags over with MP2 edit. Then I can check the two known movies with import issues and compare to my current MP2 installation

If you just want to find out how the two known movies with import issues are doing, this sounds like overkill to me. Just copy the directories with these two movies in it to some scratch space, change whatever you want to test (nfo-file, file-name) and add the scratch space's path as new MediaSource. That seems much quicker to me...

Finally - for the records - some results from analyzing our code:

• We can use SeriesTvDbMatcher.FindAndUpdateSeries(SeriesInfo seriesInfo) with only IMDB-ID in the SeriesInfo parameter to check if the IMDB-ID belongs to a series (then it returns true) or not (e.g. for a movie - then it returns false).
• We can use MovieTheMovieDbMatcher.FindAndUpdateMovie(MovieInfo movieInfo) with only IMDB-ID in the MovieInfo parameter to check if the IMDB-ID belongs to a movie (then it returns true) or not (e.g. for a series - then it returns false).

I wasn't completely sure in particular about the second bullet-point above, because TheMovieDb also has series in their database. But I tested with "Game of Thrones" (tt0944947), which is (as series) in their database (https://www.themoviedb.org/tv/1399-game-of-thrones) and the API method we use for searching for movies did not return a result. Very good...

So the idea would be:

• If the NfoMovieMDE cannot parse a nfo-file as xml-file
  • Treat the nfo-files as string and parse for a valid IMDB-ID. (If multiple are found, we just take the first one and assume the others are the same)
  • If one is found, check via MovieTheMovieDbMatcher whether the IMDB-ID belongs to a movie.
  • If yes, add the IMDB-ID to the MovieAspect and let the MovieMDE do the rest (lookup will then come much faster from cache because NfoMovieMDE already downloaded the metadata)
• If the NfoSeriesMDE cannot parse a nfo-file as xml-file
  • Treat the nfo-files as string and parse for a valid IMDB-ID. (If multiple are found, we just take the first one and assume the others are the same)
  • If one is found, check via TvDbMatcher whether the IMDB-ID belongs to a series.
  • If yes, add the IMDB-ID to the SeriesAspect and let the SeriesMDE do the rest (lookup will then come much faster from cache because NfoSeriesMDE already downloaded the metadata)

This shouldn't affect performance for those who have well-formed xml in their nfo-files, because all the above is only done when parsing xml is not possible.

 

[Lehmden]

we should make MP2 take care of them as well.

I'm sure at some point this simply will be impossible. Imagine we will later have MDE for musicvideos, documentary, anime, home videos, adult videos, sports, animated picture books, classic cartoons,... How should the importer handle all this without any pre- ordering by the user???

Imho MP2 (or any other HTPC software) is not meant for those "don't want to be bothered with that - people". Such people aren't even be interested to install and run such a complex tool at all. They simply attach their USB sticks or external HDD to the TV and watch the videos...

 

[Lehmden]

We can use MovieTheMovieDbMatcher.FindAndUpdateMovie(MovieInfo movieInfo) with only IMDB-ID in the MovieInfo parameter to check if the IMDB-ID belongs to a movie (then it returns true) or not (e.g. for a series - then it returns false).

Are you sure with that? In meantime TheMovieDB has a rapidly growing series section too... So I would not count on this...
By the way, are there any plans to improve the SeriesMDE by using TheMovieDB as primary source and TVDB only as fall-back? TVDB has lots of performance issues and they are removing a lot of content or locking series so no localized descriptions can be added and so on. In general TVDB is a great pain in the a... but before TheMovieDb was supporting TV series there was no other choice. But now we can use TheMovieDB for series too as we do for movies already...

 

[MJGraf]

I'm sure at some point this simply will be impossible. Imagine we will later have MDE for musicvideos, documentary, anime, home videos, adult videos, sports, animated picture books, classic cartoons,... How should the importer handle all this without any pre- ordering by the user???

Well, that's just a question how smart our MDEs are. Sure, you will get the best results if you pre-sort your media files. But it can work differently:
Imagine that we have a video file for which the MusicMDE finds information on MusicBraniz - must be a music video...
I think as long as people like we are able to pre-sort the media files and get a 100% match, there should still be the possibility for those who don't and make the best out of it. I even think that this will be THE feature that in the future distinguishes a successful media application from an unsuccessful one. I have a music file collection and Amazon Prime (now with music). I sometimes just don't care where the music comes from. Someone is sitting in my living room and telling me a about this great song with name xyz. In this situation I don't care where the music comes from - I want to hear it. And if I can just use MP2 and say "play that song", it should just play that song if it can get it from any source that is accessible. If I have it as FLAC from my own CD, it should play that one because of the quality. But if it isn't available locally, it should try to get it from Amazon Prime as mp3-stream and if that doesn't work, try to play a music video from youtube with that song. That's at least in my opinion the future of a media application.

Are you sure with that?

Yep. TheMovieDb does provide series as well. But we use the "/movie" method of the API to get our information and this method only returns results if the IMDB-ID provided is for a movie. If you want to get information on series, you need to use the "/tv" method, which we currently don't use.

By the way, are there any plans to improve the SeriesMDE by using TheMovieDB as primary source and TVDB only as fall-back?

I don't have plans, yet (mostly due to lack of time). But in any case, this would require extensive changes to OnlineLibraries, which I would postpone to after the MIA rework...

As to the topic above: Thinking and experimenting a bit more, parsing just for IMDB-ID will likely only work for movies. The reason is that there are two "types" of IMDB-IDs: There is an IMDB-ID for the whole series (e.g. tt0944947 for "Game of Thrones") and one IMDB-ID for each episode of that series (e.g. tt1480055 for season 1, episode 1 of "Game of Thrones").
TheTvDb can handle the series-IMDB-ID - but then we don't have a season or episode number, so this one alone doesn't help us.
But TheTvDb cannot handle the episode-IMDB-ID - so this one won't help us either...

So for now I'll only implement the movie part of it...

 

[MJGraf]

Created Jira-Issue for this...

 

[Lehmden]

Imagine that we have a video file for which the MusicMDE finds information on MusicBraniz - must be a music video...

As example this video is named "Abba - Mama Mia.mp4". This definitely will have a match at MusicBrainz so we assume it is a videoclip. But it also could be a movie, a documentary, a musical and even a Youtube clip with a hauling dog. So I want to see the Musical "Mama Mia" and got a hauling dog instead. How smart this would be...

 

[breese]

You all are aware this is the Public forum.....

Sure, you will get the best results if you pre-sort your media files.

Always have.... Still does not get it right without work.

Seeing we are on this topic, it is possible as I explained prior to Choose any or all Types to be imported and then choose a folder.
I did another test of this yesterday and its interesting that there is only 1 importer running.
[2015-12-05 12:13:57,373] [104311 ] [31 ] [INFO ] - ImporterWorker: Import for path '{e88e64a8-0233-4fdf-ba27-0b44c6a39ae9}:///F:/MPImport/', MediaCategories: 'Audio,Video,Movie,Series', including subdirectories requested...
[2015-12-05 12:13:57,389] [104327 ] [4 ] [INFO ] - ImporterWorker: Scheduled ImportJob(#2) (Activated) (Path ='{e88e64a8-0233-4fdf-ba27-0b44c6a39ae9}:///F:/MPImport/', ImportJobType='Import', IncludeSubdirectories='True')
[2015-12-05 12:13:58,217] [105155 ] [4 ] [INFO ] - ImporterWorker.ImportJob(#2): Activated

I would have thought there would be one for Each Type of Media....