this view could not be built (5 Viewers)

[MJGraf]

Thanks, breese, very helpful!
Logs are looking good.
Reason for (at least some of) the movies above not being recognized is that there is no unique match on tmdb. You can either put the imdb-id into the file name or the year in (brackets). So if Armageddon is the one with Bruce Willis from 1998, either rename to "Armageddon (tt0120591).mkv" or "Armageddon (1998).mkv". Both should (hopefully) work...

 

[breese]

I normally have the TT#'s included....
To help with your Testing, I removed them so as to see the % that can or cannot be found by name only.
I will look into the (Year) idea and see what happens

 

[breese]

On just the ones it could not find I added the (Year) to the file name
Next I did a Reimport.
Not so good...
Now Movies see's 28 of them...

 

[breese]

Cleared everything out and did a Fresh import
Still only see 28 Movies out of 47
As I said, I added the (Year) to the end of the name on Only the ones the importer did not find.
Interesting to note: the importer does not allways see the (Year) in the name
I am also including a text file of the exact directory of Movies

 

[JSchuricht]

I'm a bit behind. I did have the dll but had imdb in the wrong spot in the nfo config. Still missed a bunch of movies. After 3 imports, 2297 in movies and 3488 in videos.

Running a new import now with the files from MDEImprovements.7z applied. The server service crashed about 30 seconds into the first import. If it stays running, I should have some logs in the next few hours.

 

[MJGraf]

Thanks for testing, JSchuricht!

I did have the dll

I dare to disagree - the logs you posted above cannot have been produced by the new dll. To make it better visible, I pushed the version number of the NfoMetadataExtractors plugin from 0.10 to 0.11.
In the logs you attached to the post right above this one, it still says v0.10 - and again the NfoMovieMetadataExtractorDebug.log file contains entries that cannot stem from any binary posted in this thread.
Before you wait another couple of hours: Can you post where exactly you put the dlls? Just to make sure we are on the same page...

 

[JSchuricht]

The dll went under C:\Program Files (x86)\Team MediaPortal\MP2-Server\Plugins\NfoMetadataExtractors\
The only other place I found the old dll is under MP2-Client\Plugins\NfoMetadataExtractors. Perhaps that one is getting used instead?

Anyways, here's the details from the dll, new on the right and old on the left. It kinda looks like the new one is v 10 and the old one v 11.

This is the current version from the files in MDEImprovements.7z which is currently in C:\Program Files (x86)\Team MediaPortal\MP2-Server\Plugins\NfoMetadataExtractors\

 

[MJGraf]

Thanks, JSchuricht. Cannot reproduce the upper two, but the lower one is certainly the right one.

@breese: I can reproduce your result. I get 29 matches instead of 28, but in any case not >40.
I've traced one of the non-matches down: "Armageddon (1998)".
The issue is that even with the year, there is no unique match. Tmdb returns the right movie "Armageddon", but also "Armageddon Boulevard" also dating 1998. That's why we don't get a unique match. What I could imagine as an improvement at this place is that if we have a 100% match of title and year (i.e. TitleFromTmdb == TitleFromFileName), we prefer it to a "contains-match" (i.e. TitleFromTmdb contains TitleFromFileName). That way we would have a unique match in that case.
What I generally found out (but I already feared this), is that our tmdb-wrapper is not really ideal. What we e.g. do is that even if there is a year, we only search for "&query=Armageddon"; as a result we get 20 results (with a paging possibility to get another 26 results) and locally filter for those with the right year. We could also just search for "&query=Armageddon&year=2008" and we would only get two results (as described above). That would save us a lot of unnecessary internet traffic and, hence, time...
Anyway, at some point we will have to write our own wrappers at least for the most important databases such as tmdb. The ones available via nuget are altogether not ideal...

 

[JSchuricht]

Now we are getting somewhere. The server service crashed twice but 3270 movies and 3488 in videos.

Since I have gotten a small glimpse into how the importer works, I think part of my issue may be from nfo files that don't match the movie name and aren't named movie or imdb. Working on a batch file to rename *.nfo to movie.nfo on the extra copy I made of everything. Hopefully that will pickup the last 218 movies.[DOUBLEPOST=1450048615][/DOUBLEPOST]So renaming all the nfo to movie.nfo was easy. Doubt this import will finish before I have to leave so results will have to wait till tomorrow. I will probably convert all the imdb.txt to movie.txt next to eliminate the extra config mods.

If anyone is interested, just a batch file in the root directory and may cause issues so backup first.

@ECHO OFF
PUSHD .
FOR /R %%d IN (.) DO (
    cd "%%d"
    IF EXIST *.nfo (
       REN *.nfo movie.nfo
    )
)
POPD

 

[Lehmden]

Hi.
I've done a first quick test... Prepared two movies, one that has no IMDB-ID but a TMDB-ID, the other one was the same as the example above (Avatar). Both got nonsense names and both were tagged only with the TMDB-ID, nothing else.... Then I've added a movie without any metatags, but with a striped down .nfo (but I keep the valid xml structure) to contain only the TMDB-ID. And I also prepared two series episodes, also both with nonsense names. One series has a valid TVDB-ID and the other is from a series that was deleted from TVDB recently. This now only can be found on TMDB (what we don't use for series up to now, knock, knock ).... No .nfo files (excerpt .nfo for the movie without any tags), no imdb.txt nothing...
The result is as expected. All three movies and the "valid" series episode are recognised properly. the other series episode is not recognised at all (as I had expected it).