What does minidlna do with nfo files

[Josch]

Does minidlna really use nfo files and import the content into its database?
If yes, which parts of the nfo files are used?

[Shrimpkin]

Yes, minidlna supports metadata import from nfo (xml) files. Programs are available to generate them. Although, they are not difficult to make. Currently supported tags: title, episodetitle, plot, capturedata, genre, mime. Ex:

<?xml version='1.0' encoding='UTF-8'?>
<title>The Matrix</title>
<plot>A Sci-Fi movie.</plot>
<capturedate>1999</capturedate>
<genre>Sci-Fi</genre>

[Josch]

Thanks

[Shrimpkin]

Not to encourage invalid xml but I only posted what minidlna needed. The xml library from miniupnp doesn't seem to care about the invalid xml above. Just add root tags to make valid. Ex:

<?xml version='1.0' encoding='UTF-8'?>
<movie>
  <title>The Matrix</title>
  <plot>A computer hacker...</plot>
  <capturedate>1999</capturedate>
  <genre>Sci-Fi</genre>
</movie>

[Ruben Denolf]

Sorry for commenting on this old post, but I am trying to make use of the nfo-file with my minidlna-server. I have nfo-files generated by FileBot, does the file need to start with:

<?xml version='1.0' encoding='UTF-8'?>

Or would the generated file from filebot suffice?

<tvshow>
  <title>Blindspot</title>
  <sorttitle>Blindspot :: 2015-09-21</sorttitle>
  <year>2015</year>
  <rating>7.9</rating>
  <votes>46</votes>
  <plot>A vast international plot explodes when a beautiful Jane Doe is discovered naked in Times Square, completely covered in mysterious, intricate tattoos with no memory of who she is or how she got there. But there's one tattoo that is impossible to miss: the name of FBI agent Kurt Weller, emblazoned across her back. "Jane," Agent Weller and the rest of the FBI quickly realize that each mark on her body is a crime to solve, leading them closer to the truth about her identity and the mysteries to be revealed.</plot>
  <runtime>45</runtime>
  <mpaa>TV-14</mpaa>
  <id>295647</id>
  <genre>Action</genre>
  <genre>Crime</genre>
  <genre>Drama</genre>
  <genre>Mystery</genre>
  <genre>Thriller</genre>
  <thumb>http://thetvdb.com/banners/graphical/295647-g.jpg</thumb>
  <premiered>2015-09-21</premiered>
  <status>Continuing</status>
  <studio>NBC</studio>
  <actor>
    <name>Jaimie Alexander</name>
  </actor>
  <actor>
    <name>Sullivan Stapleton</name>
  </actor>
  <actor>
    <name>Ashley Johnson</name>
  </actor>
  <actor>
    <name>Francois Arnaud</name>
  </actor>
  <actor>
    <name>Ukweli Roach</name>
  </actor>
  <actor>
    <name>Marianne Jean-Baptiste</name>
  </actor>
  <actor>
    <name>Rob Brown</name>
  </actor>
  <actor>
    <name>Audrey Esparza</name>
  </actor>
  <tvdb id='295647'>http://www.thetvdb.com/?tab=series&amp;id=295647</tvdb>
</tvshow>

[Shrimpkin]

The XML declaration is not needed by the XML parser.

May be useful for other applications and is recommended. For example, the file(1) program will show the nfo as html without it.

[Phantom of Krankor]

Does this still work? I have tried this with 1.1.4 and 1.2.0, and no luck. I even copied Shrimkin's example, and still nothing.

I do see this in the dlna debug output:

[2017/07/26 18:56:10] metadata.c:175: debug: Parsing .nfo file: /storage/movies/dir1/dir2/Fname.nfo

I have Fname.mp4, Fname.nfo, and Fname.jpg

I have tried using foobar2000 and BubbleUPnP, and neither of them are picking up the data from the nfo file. Bubble is picking up the jpg as a thumbnail.

Any idea what I'm missing?

Thanks.

[Shrimpkin]

Version 1.1.4 (up to 1.1.6) should work. Version 1.2.0 nfo support appears to be broken. I will post a patch in Tickets > Patches.

For working versions, only supported tags: title, episodetitle, plot, capturedate, genre, mime.

[Phantom of Krankor]

Thanks! Looking forward to the update.

What would go in the "mime" entry?

Also, I managed to get 1.1.4 working. What I found is that the info is being cached, so updating the xml file doesn't do any good until a rescan is forced. That's someone cumbersome, because of how long it takes to rescan my library.

[Shrimpkin]

There is a fix in git.

The mime content type (usually not needed as minidlna can figure it out). https://en.wikipedia.org/wiki/MIME

minidlna doesn't actually scan for nfo's but rather the media file. You can "toggle" the media file to get minidlna to (re)parse the nfo. Ex: videoname.{nfo,mp4}

touch videoname.mp4

Or to keep the file's modtime, try using mv. Ex:

mv vidoename.mp4 .tmp_hidden && mv .tmp_hidden videoname.mp4

[Konrad Brunner]

I compiled the latest version of minidlan got by git. (1.2.0).
There are no nfo informations in my kodi or Samsung tv. (also not shown with version 1.1.2)
There is no albumart shown on Samsung TV (ES 6300). (is shown with version 1.1.2)
What must be done to show information (e.g. title) from the nfo file?

[Phantom of Krankor]

Try getting the source via git clone. It's newer than the 1.2.0 package.

Between the fixes and the advice about touching the video file, I'm able to get this working. (Although my TiVo can't see any videos, but that's a different issue, which I'll open up another thread for.)

Is having the file extension be "nfo" part of a standard? My Windows brain clashes with that, and even my Linux brain says that it should be "xml" I'm good with whatever, I'm just curious.

[Shrimpkin]

Current git/version now has nfo support using inotify. You no longer need to use the "toggle" method.

Note: There is no support for nfo's using the new rescan feature. Make any nfo changes while minidlna is running or rebuild database.

[tony]

Can the same be done for music?

The reason I ask is because I can't change the metadata on my actively seeding torrents. My only option is to make corrections through a nfo file.