[programmkino]

Hi once again,

Kodi doesn't show the movie's thumbnail picture even though it has the same information and syntax like from the IMDB scraper.
Is it because it doesn't have a file extension? If so, is there any way around?

This is an excerpt from the nfo file:

xml:


<?xml version="1.0" encoding="UTF-8" standalone="yes" ?>

<movie>

    <title>Shoplifters - Familienbande</title>

    <originaltitle>MANBIKI KAZOKU</originaltitle>

    <userrating>0</userrating>

    <top250>0</top250>

    <outline>x</outline>

    <plot>Eine am Rande von Tokio wohnende Gruppe von Menschen hält sich mit kargen Einkünften und kleinen Diebstählen über Wasser. </plot>

    <tagline>x</tagline>

    <runtime>16</runtime>

    <thumb aspect="poster" preview="https://www.filmdienst.de/bild/filmdb/154701">https://www.filmdienst.de/bild/filmdb/154701</thumb>

    <mpaa></mpaa>

    <playcount>0</playcount>

[programmkino]

As a workaround, I'm thinking about retrieving the thumbs outside of Kodi via a shell script or similar to putting them in the place where Kodi expects them.
I've noticed that the thumb files are put into userdata/Thumbnails directory.
Their names are changed to 8-digit hex values e.g. 0a2dbced.jpg or 0d784a2c.jpg.
I guess some hash algorithm like MD5 is used for that, but the details aren't obvious to me.
Does someone have details on that?

[Klojum]

Kodi doesn't show the movie's thumbnail picture even though it has the same information and syntax like from the IMDB scraper.

I don't know what IMDB uses as a poster syntax, but Kodi uses the "Movie name-poster.jpg" syntax.

The 8-digit hex value filenames are calculated indeed, and then processed via the local Textures13.db SQLite database.
The thumbnail, poster, fanart, etc will be stored in the Thumbnails folder. So creating the thumbs yourself, even if you have the correct hash, is not good enough.

Fanart should be picked up when scraping for local information, provided all image files have the correct naming format.

[programmkino]

Hi Klojum

Thanks for the reply. Then I don't go for putting the thumbs there by external scripts.
But once more to my original issue.

If I import that nfo file, Kodi downloads, stores and shows the thumb image.

xml:
<?xml version="1.0" encoding="UTF-8" standalone="yes" ?>

<movie>

    <title>Shoplifters - Familienbande</title>

    <originaltitle>MANBIKI KAZOKU</originaltitle>

    <userrating>0</userrating>

    <top250>0</top250>

    <outline>x</outline>

    <plot>Eine am Rande von Tokio wohnende Gruppe von Menschen hält sich mit kargen Einkünften und kleinen Diebstählen über Wasser.plot>

    <tagline>x</tagline>

    <runtime>16</runtime>

    <thumb>http://image.tmdb.org/t/p/original/7ZCWWFLazB7QiKHHLBWe8OuDMbY.jpg</thumb>

    <mpaa></mpaa>

However, if the thumb image does not have an file extention (it's a normal jpg) Kodi doesn't download, store and show it.

xml:
<?xml version="1.0" encoding="UTF-8" standalone="yes" ?>

<movie>

    <title>Shoplifters - Familienbande</title>

    <originaltitle>MANBIKI KAZOKU</originaltitle>

    <userrating>0</userrating>

    <top250>0</top250>

    <outline>x</outline>

    <plot>Eine am Rande von Tokio wohnende Gruppe von Menschen hält sich mit kargen Einkünften und kleinen Diebstählen über Wasser.plot>

    <tagline>x</tagline>

    <runtime>16</runtime>

    <thumb>https://www.filmdienst.de/bild/filmdb/154701</thumb>

    <mpaa></mpaa>

So it's not related to scraper anyhow. It's the behaviour of Kodi itself.
Do you think this could be changed?

[programmkino]

Hi again,
Kodi works like it should do.

I noticed the log entry:

Code:

18:31:34.612 T:3372   DEBUG: CTextureCacheJob::GetImageHash - unable to stat url https://www.filmdienst.de/bild/filmdb/142054

Some more testing showed that Kodi sends HTTP HEAD requests for the thumb pictures (which is valid).
However, the website filmdienst.de returns the pictures for HTTP GET requests (200, success) but not for HTTP HEAD requests (404, file not found).
So it's the behaviour of that website.

I'll set up a liltte http proxy to change HTTP HEAD into HTTP GET requests. Still, it's ugly.

Thanks everyone. Feel free to close the thread.