Kodi Community Forum

I am migrating from YAMJ (https://github.com/yamj/yamj-v2) to KODI. However, I am a bit confused about what settings to use in the scraper. The directory structure I have for movies is firstLetter/movieName/movie.mkv. So for examplemb://ipaddr/Movies/S/Serenity (2005)/Serenity (2005) [tt0379786].720p BluRay DD5.1 x264-RightSiZE.mkv. I am using the TMDB scraper.

I am confused about whether to check the: Movies are in separate folders that match the movie title, the Scan recursively button or both. I tried both and KODI seemed to get confused. It produced a lot of folders with no name and posters for only about 1/10 of the movies.

The movies are on a SMB share and was originally set up to be indexed using YAMJ. Almost all movie folders have a .nfo and a .jpg (poster file). The .nfo is in XMBC format so it should be ok. The .nfo and .jpg files match the movie filename (or the filename of the first part of the movie if multipart). The .jpg is not named folder.jpg. It is also unclear if the folder name needs to match exactly the movie filename (what about multipart movies then?)

This is on a RasberryPi 2 running OpenElec 6.0.0

Any help appreciated. Thanks.

Set the Movies folder as the source, with scan recursively on but separate folders off.

If you check the next level down (the individual letters one, "S" in your example above) then both should be set to on as each film within the letter folder is at that level in its individual named folder (as per your example above again).

See the wiki page on naming video files (wiki) for more details

Thanks, that seems to be working. However as soon as it encounters a .nfo file it seems to use that only and doesn't get any other info online. That is a problem because the .nfo only contains supplemental info.

I wrote a program to rename all the .nfo files to another extension and rescanned the directories. It now seems to be populating the database correctly. After the scan is complete, I plan to rename the .nfo files back (YAMJ still uses them) and I am in transition. TMDB could not figure out some of the movies, so I will use the local .nfo to refresh those. I assume renaming the .nfo file back should be ok unless I tell KODI to rebuild the database.

Under the change content menu (for those movies that it couldn't figure out), it asks if I want to change the scraper. I assume that is only associated with that particular directory (the movie directory) and not global for the source. Is that correct?

Yes, that's how Kodi works. NFO files (wiki) override network fetched data.

Not sure about your final question, maybe someone else can assist there.

Thanks, This is very helpful. I am not complaining, I am just trying to understand how KODI works. I was used to the YAMJ setup where most of the info was gotten from online, but the .nfo was used to override some of the data (unless you specifically set a flag in the .nfo telling it not to scrape). Kodi seems to take an either or approach, but that is ok as long as I know what to expect.

The last question was about how to fix the movies that it couldn't figure out in TMDB. If you go to one of those movies (I am using the confluence skin), select it and right click then it brings up a context menu. One of the options is to change content. This brings up the set content menu which allows you to set the scraper. I assume this is only for this particular folder (the movie folder), so I can set it to local scraping only without disturbing all the other movie scrapings.

For the last question in the above post, yes that is correct.

You can also guide and tweak the scraper search string there too if you do it on a file rather than a folder, using the movie information context menu entry.