this view could not be built (1 Viewer)

[MJGraf]

What we could do code-wise (to avoid that you have to Change all your nfo-files) is that if it is not possible to read an nfo-file correctly as XML (which is here in the code), we try to read the whole nfo-file as string and at least try to find a sub-string in the form ttxxxxxxx and treat this as IMDB-ID. For the URL you have in your nfo-files, it would then just filter the IMDB-ID from the URL and use it.
@Lehmden (as master of nfo-files ): Is there any reason we should not do this? (And do you know from the top of your head that the number after "tt" in IMDB-IDs is always 7 Digits? I just can't remember...)
Not sure if I can implement it this Weekend (because it Needs to work for movies and series correctly), but it doesn't look too complicated...

 

[Lehmden]

Hi.

that the number after "tt" in IMDB-IDs is always 7 Digits?

It is. There are leading "0" if the pure number is less that 7 digit. As Example, the IMDB-ID for "Star Wars" is "tt0076759"...

Needs to work for movies and series correctly

Could be problematic as TVDB-ID did not have any leading letters and also some TV shows did have IMDB-ID numbers a lot of them did not. The search for a "tt" Number in Series is not very common or useful...

In general it's not too bad if the importer searches for any text file containing a "tt" number. E.g. the "scene" downloads often contains an "nfo" file that is no valid XML .nfo as we know it from Kodi and co. Those .nfo are mostly advertising for the "scene group" that has released this video and it often contains some ASCII artwork and most of the time a link to IMDB with the "tt" number. So even those ".nfo" are useful to identify the movie then...

 

[breese]

While I do a little more testing... It is starting to look like there has been an undocumented change on the www.themoviedb.org site.
A search for Sweeney Todd 2007 shows No Results Found
A search for Sweeney Todd shows 7 movies and 3 TV shows...
Also, under each find the Date (undocumented I believe) is now shown as Month/Day/Year
I would swear just a few days ago it only listed Year

IMDB search for Sweeney Todd 2007 shows a number of results and within those results show
Name (Year)
and or Name (Year) (type TV Movie, Short, etc)

Also, As I have discussed in the Dev Area, (here comes the broken record) defaulting to Videos when something is not found is making it very hard on everyone envolved. The new user trying to import, the testers and others wanting to help the new user, and the developers in time away from actually writing / finishing the code to make this all work....

 

[MJGraf]

Could be problematic as TVDB-ID did not have any leading letters and also some TV shows did have IMDB-ID numbers a lot of them did not. The search for a "tt" Number in Series is not very common or useful...

Yep, we won't be able to parse TVDB-IDs because they are not distinguishable enough.
For series we would also search for the IMDB-number. Even if it is not common or not useful, it doesn't hurt... (but I need to have a look (a) whether we have IMDB-ID for series at all in our MIAs [EDIT: Confirmed it is in the SeriesAspect] and (b) if so, how we treat series vs. Episode IMDB-ID - IIRC I already had to do with this question and we decided only to store the series-IMDB-ID. Need to look into the code...)

But the most difficult thing is to distinguish between Series and Movie. When there is no <movie> tag and no <tvshow> or <episodedetails> tag in the file, there is no way to find out locally whether the IMDB-ID belongs to a series or a movie (at least if a MediaSource is added with both categories: movies and series so that both, the NfoMovieMDE and the NfoSeriesMDE are applied). But we need this Information to find out whether we store the IMDB-ID in a SeriesAspect or a MovieAspect. This means, once we have found an IMDB-ID, we need to look up online at IMDB to find out what it is. And since an online look-up take a lot of time, this behavior should only kick-in, when we are not able to read the data locally in XML-form.

Let me think about it...

 

[breese]

I am going to go thru this again because I still do not understand and maybe it will help solve the issues.

If a lookup onto www.themoviedb.org for a movie name fails, it shows up in Videos - OK - Got it.

If the Movie Name has the tt## included, the lookup switches to IMDB.com? I say that because most of mine have tt##
Now here is the thing I Do Not Understand and never have...

Max [tt3369806].mp4 goes thur the importer and lookup.
The importer then creates 2 folders (sometimes 3) within C:\ProgramData\Team MediaPortal\MP2-Server\TheMovieDB

For the movie Max
C:\ProgramData\Team MediaPortal\MP2-Server\TheMovieDB\tt3369806
With only a jason file within it

C:\ProgramData\Team MediaPortal\MP2-Server\TheMovieDB\272878
With the Posters, Backdrops, and jason file

How did the importer jump from using the tt## and locating all apsects of the movie to Using TheMovieDB info?
Is it possible this same thing is happening for other imports such as movie names AND folders with (Year) in them?
Is it possible because these movie names AND folders have (Year) in them that the .nfo is not being looked at?
Then there is the '-' in a file name......

Then there is additionally an outer processing which tries to split titles by "-" (which is quite common for translated titles, where the original one is followed by a localized title):

So the movie halcyon-swingers as a last case try gets its name split for a lookup...
Is one or both parts of the name searched?
Does one of the 2 parts get the '-' iincluded in the search?
What happens if there is a folder with a '-' in the name?

 

[MJGraf]

How did the importer jump from using the tt## and locating all apsects of the movie to Using TheMovieDB info?

That's an easy one. We use the IMDB-ID, but we don't use IMDB TheMovieDb has the possibility to search for IMDB-IDs and then Returns TheMovieDb data, which we use.

So the movie halcyon-swingers as a last case try gets its name split for a lookup...

IIRC we first search for the Name as it is ("halcyon-swingers"). And if this doesn't work, we split it by "-" and then use only the substring before the "-" for a search ("halcyon"). That's without having looked into the code, just from what I remember...

 

[MJGraf]

Oh, and this is not entirely true:

If a lookup onto www.themoviedb.org for a movie name fails, it shows up in Videos - OK - Got it.

If there is a (valid) nfo-file for a movie, the movie doesn't have to be found on themoviedb.org and will nevertheless Show up as movie. We just Need "anything" that identifies the file as a movie. Anything currently means (a) a valid nfo-file with a <movie> section in it; OR (b) a unique match on themoviedb.org.

 

[breese]

OR (b) a unique match on themoviedb.org.

Seems to be a loosing battle...

Confusing is the fact that there are only 2 exact hits Yet the Lookup seems to Ignore the Year UNLESS it is inclued in the search with a y: filter...
"Tip: You can use the 'y:' filter to narrow your results by year. Example: 'star wars y:1977'."

Full Name - multiple hits because its a Remake

With the (Year) added to the name

With the Year added to the name


Now, I renamed a movie and did some testing.
The only import that correctly imported the movie into Movies (2 went to Videos with no other info assigned) was the one with the tt##
One of the tests I added TheMovieDB ID# to the end of the name
MovieTheMovieDbMatcher: No unique match found for "Sweeney Todd 13885"

 

[Lehmden]

Hi.

But the most difficult thing is to distinguish between Series and Movie.

Don't think so as we can leave this to the user. Unless the behaviour will change in (near) future the user only can select one option, movies or series, not both at the same time while adding a media source...

For series it would be good if we can use the TVDB-ID as (part of the) name for the series folder so all episodes inside this specific folder are belonging automatically to the series with ID xxxxx... With this user that don't want to use .nfo and/or matroska tags still have a 100% save method to identify their media... tt number as (part of the) video file- name for movies and TVDB-ID in the series- folder name combined with season and episodes number as (part of) the episodes file- names always is enough for a 100% save identification....

 

[breese]

the user only can select one option, movies or series, not both at the same time while adding a media source...

Not True...
I tested this a Very Long time ago and it still works

On import, everything is Sorted to Either Movie or Series (and of course to Video ARG.... )