this view could not be built (4 Viewers)

[breese]

I've traced one of the non-matches down: "Armageddon (1998)".
The issue is that even with the year, there is no unique match.

Interesting find here...
includes a search method for Movie name + Year and got Armageddon 1998 the first time.
http://www.omdbapi.com/

 

[MJGraf]

Interesting find here...
includes a search method for Movie name + Year and got Armageddon 1998 the first time.

Right, because "Armageddon Boulevard" ist registered on IMDB with the year 1999 (and IIRC omdbapi is just the API-Website of IMDB). So the question in this case is: Was "Armageddon Boulevard" published in 1998 or 1999? But apart from that: we do not use IMDB (or omdbapi), just tmdb...

The server service crashed twice

While I was re-reading your post, I noticed this. I'd be more than interested to find out why the service crashed - which should definitely not be the case... There is nothing in your logs that points to a crash. How did you determine that the server service crashed? I'm asking because sometimes there is no activity in the logs for quite some time and it may look like it crashed, but it is just waiting for something. During your movie-import e.g. we are downloading lots of data from tmdb and they have a request rate limit, which MP2-Server needs to obey, so it might be the case that it didn't actually crash, just waiting for something. Or did you actually see in the task manager that the process had ended?

 

[Lehmden]

Hi.

and IIRC omdbapi is just the API-Website of IMDB

Also you are right most of the time, here you are not right, I fear.
Excerpt from the website:

API by Brian Fritz.

All content licensed under CC-BY 4.0.

This site is not endorsed by or affiliated with IMDb.com.

 

[breese]

So the question in this case is: Was "Armageddon Boulevard" published in 1998 or 1999?

What does it matter if this movie was made in 1998 or 1999??? Its not the movie i am trying to find.

Also, I am sorry but there has to be Something in the importer of search method that is using IMBD.
While I know this might make sense to you, the results of the importer make No Logical Sense to me...
Including an unsupported query (name + tt# ) to themoviedb.org and getting exact hit results makes no sense.

If I include the [tt####] in the name of the movies, the importer works but
If you go up to tmdb and do a search for the movie name and the tt#####
It Always shows Unable to find any items


Matter of fact... if you go up to the
type in armageddon and it gives you a list of possibilities
now add a Space 1 - Shorter List But the Movie we are looking for is Not Shown.
Now add a 9 - the movie we are looking for not on the list
Now add a 9 - the movie we are looking for not on the list
Now add an 8 - Nothing is found.

So how is it that the importer (and not accessing IMDB) can locate a movie the First Time, Every Time, with the tt### in the name, yet a Year addition on the TMDB site cannot locate the Same Movie?

I really do not want to do another network trace of MP2....

 

[Lehmden]

Hi.

If I include the [tt####] in the name of the movies, the importer works but
If you go up to tmdb and do a search for the movie name and the tt#####
It Always shows Unable to find any items

So how is it that the importer (and not accessing IMDB) can locate a movie the First Time, Every Time, with the tt### in the name, yet a Year addition on the TMDB site cannot locate the Same Movie?

If you do a web- search on themoviedb.org (means browse to the website and enter something into the search field) the tt numbers are not found. Simply becaue no human being is familar with (lots of) such machine- IDs. In a special case you may know the tt number, but in general a web- user is searching for a movie title or an actor or so.

But MP2 don't do a web search at all. We are calling the API from themoviedb.org as this is the counterpart of a websearch for machines. And this API supports the IMDB-ID for sure as a computer is very familiar with computer id's So it's absolutely logical that the tt number will result in a perfect match on the API call and also never will get a hit on the web- search...

 

[Lehmden]

If you want to find out more about the TMDB- API you may have a look at this link:
https://www.themoviedb.org/documentation/api

 

[breese]

Well it seems that while TMDB allows for external finds... Using their own search methods do not work to a caliber required to get reliable hits without jumping thru hoops.

Even their own Tips for searching with a Year cannot produce reliable results.
Tip: You can use the 'y:' filter to narrow your results by year. Example: 'star wars y:1977'.

Using that Tip on their web site:
armageddon y:1998
Produces 2 hits

 

[Lehmden]

Using that Tip on their web site:

The issue is that all foreign language names, possible country release dates and so on are being found (and in general this is a good thing!). It's very common that a move is premièred later (maybe many years) in other countries and/or languages...
Another example... If you are searching for "Avatar" it gives a result of 25 movies... Even if you add y:2009 you will get 4 results. The movie itself, two documentaries about the poroduction of the "Blockbuster" Avatar and a french movie from 2007 originally named "Chrysalis" this was released in 2009 in Germany finally. A human being easily can see which one is the "right" one. A computer can't see this. That's why unique iD's are that important.

 

[JSchuricht]

So I have some not so great news. With all the nfo's renamed to movie.nfo the importer seems to be locking up. I started the initial import yesterday, checked on it about 14 hours later and found 955 movies imported but the server sitting idle. Another 8 hours after that it was still at 955 movies. Restarting the server service seems to import a few movies then stall. After restarting the service 6 or so times I got it up to 971 movies and 999 videos. Somehow I get the feeling that some random ASCII art in some nfo's is making the system choke.

I won't have much time to work on it today but just for kicks, I will setup a script to restart the service every 10 minutes to see what the count is at tomorrow.

 

[breese]

Every one of the ServerError logs stop at the exact same place
[2015-12-14 15:51:56,645] [22906 ] [82 ] [ERROR] - MediaLibrary: Error adding or updating media item(s) in path '{03dd2da6-4da8-4d3e-9e55-80e3165729a3}:////NUMBERSIX/movies/E/Eight Men Out (1988)/Eight.Men.Out.1988.1080p.BluRay.x264-SiNNERS.mkv'
System.Data.SQLite.SQLiteException (0x80004005): constraint failed
FOREIGN KEY constraint failed
at System.Data.SQLite.SQLite3.Reset(SQLiteStatement stmt)
at System.Data.SQLite.SQLite3.Step(SQLiteStatement stmt)
at System.Data.SQLite.SQLiteDataReader.NextResult()
at System.Data.SQLite.SQLiteDataReader..ctor(SQLiteCommand cmd, CommandBehavior behave)
at System.Data.SQLite.SQLiteCommand.ExecuteReader(CommandBehavior behavior)
at System.Data.SQLite.SQLiteCommand.ExecuteNonQuery(CommandBehavior behavior)
at System.Data.SQLite.SQLiteCommand.ExecuteNonQuery()
at MediaPortal.Backend.Services.MediaLibrary.MIA_Management.InsertOrUpdateManyToManyMIAAttributeValue(ITransaction transaction, AttributeSpecification spec, Guid mediaItemId, Object value)
at MediaPortal.Backend.Services.MediaLibrary.MIA_Management.InsertOrUpdateManyToManyMIAAttributeValues(ITransaction transaction, AttributeSpecification spec, Guid mediaItemId, IEnumerable values, Boolean insert)
at MediaPortal.Backend.Services.MediaLibrary.MIA_Management.AddOrUpdateMIA(ITransaction transaction, Guid mediaItemId, MediaItemAspect mia, Boolean add)
at MediaPortal.Backend.Services.MediaLibrary.MediaLibrary.AddOrUpdateMediaItem(Guid parentDirectoryId, String systemId, ResourcePath path, IEnumerable`1 mediaItemAspects)