Problem with adding films containing Russian letters in the file name from FTP
#1
Hey.

Kodi 18.6 RaspberryPi

Recently I had to redo access to my warehouse of video files. Initially, access was via the SMB protocol, now the same warehouse has become available under the FTP protocol.
I deleted the old media library, cleaned it, and deleted the old source.
Then he added a new source using the FTP protocol.
The source was scanned using the standard add-on The Movie Database.
Films appeared in the Movies section of the main menu. With descriptions, posters, etc.

But there are some very unpleasant problems:

- if you go to the video -> files -> "source name" section, then for all files containing Russian letters in the title, information about the movie is not displayed when highlighted. For example, at the place where the plot of the film should be, the inscription "Information is not available" appears.
At the same time, for those files in the name of which only numbers or letters of the English alphabet are used, this problem is not traced.
If you select a file and call the context menu, then there will not be a button "Details" or "Get information".

- if in the Movies section select one of them and press "confirm", then instead of showing information about the movie, its playback starts. Despite the fact that it is configured to "show information".

- if you select a file in the Movies section and select the "Information" section using the context menu, then the information will be shown, but in the "Actors" section there will be an inscription "Information is not available"
And also, the MOST PLEASANT: there is no "Refresh" button, that is, if the information about the film was not downloaded correctly, then it will not work in any way to fix it.

For a long time I tried to deal with this problem and it seems to me that I found its root.
If you look at the log file, then there we can see, among other things, the following lines:

2020-05-16 21: 13: 03.931 T: 1650365312 DEBUG: VideoInfoScanner: Adding new item to movies: http://ftp.USERNAME:[email protected].0.1: 21/Movies/4.3.2.1 (2010).mkv

2020-05-16 21: 13: 05.385 T: 1650365312 DEBUG: VideoInfoScanner: Adding new item to movies: http://ftp.USERNAME:[email protected].0.1: 21/Movies/16 кварталов (2006).mkv|utf8 = 0

As you can see, when adding a file with Russian letters to its name, the subscript "|utf8=0" is made.
There is a suspicion that this leads to some kind of inconsistency in the structure of the database, which in turn leads to all the problems described above.

I'm not so cool as to say exactly the problems of this scraper or FTP client, but that the problem is a fact.

I would really appreciate it if you could help me solve it.
I still did not understand how to attach the log file completely.

P / S / There is no such problem when using SMB sources.
I apologize for my English through Google translator)
Reply
#2
This is just a guess.  For FTP, the FTP client expects the FTP server to send utf-8 encoded folder and filenames.  It looks like if utf-8 is not used it will try to use a charset (from locale maybe?).  So if you can confirm what encoding the FTP server uses that would be a start.

scott s.
.
maintainer of skin  Aeon MQ5 mods for post-Gotham Kodi releases:
Matrix
Leia see: 304472
Reply
#3
(2020-05-17, 23:02)scott967 Wrote: This is just a guess.  For FTP, the FTP client expects the FTP server to send utf-8 encoded folder and filenames.  It looks like if utf-8 is not used it will try to use a charset (from locale maybe?).  So if you can confirm what encoding the FTP server uses that would be a start.

scott s.
.
FTP server (Filezilla) based on Windows and its internal encoding by itself is cp1251. FTP server is correctly execute the utf8=0 parameter. Otherwise, problems arise when processing file names containing the Russian letter "я". But there is no problem with this, the file names are displayed and processed correctly.
My assumption is that the links in the video database between file names and metadata do not work correctly, because postfixes "|utf8=0" are added to the file names.
I do not have the opportunity to examine the contents of the Kodi database directly, because I am not aware of how to do this. Therefore, I can’t say anything for sure.
As a result, as you rightly said, I expressed only an assumption.

At the moment, I have set up an HTTP server based on Apache, there are no such problems with this configuration with the same source. That is, this is exactly the problem with three components:
- Russian letters in the names of the files;
- FTP client Kodi and FTP server with support utf8=0;
- a scraper - in fact, it doesn’t matter which one, I tried two more, besides, the result is similar.
Reply
#4
P.S. In the Cody settings, the encoding "Cyrillic (Windows)" is selected.
Reply
#5
First, the language setting in Kodi is just the UI language.  Encoding uses the system locale.

I don't have any experience, but I installed filezilla server and set up Kodi FTP user (not SFTP) on my localhost just to test, and didn't have any problems getting files with non-latin1 unicode characters added to the database.  I use nfo file not scraper, but I don't think your problem is in the scraper.

Need a log with debug level logging uploaded to a pastebin site, might be able to understand better what is going on.

scott s.
.
maintainer of skin  Aeon MQ5 mods for post-Gotham Kodi releases:
Matrix
Leia see: 304472
Reply
#6
(2020-05-18, 20:02)scott967 Wrote: First, the language setting in Kodi is just the UI language.  Encoding uses the system locale.

I don't have any experience, but I installed filezilla server and set up Kodi FTP user (not SFTP) on my localhost just to test, and didn't have any problems getting files with non-latin1 unicode characters added to the database.  I use nfo file not scraper, but I don't think your problem is in the scraper.

Need a log with debug level logging uploaded to a pastebin site, might be able to understand better what is going on.

scott s.
.
As for the encoding - you know better, of course, but until I changed the encoding from "default" to "Cyrillic (Windows)", the file names were not displayed at all when entering the source. I mean names with the Cyrillic alphabet.
Although before that I used the SMB source and did not perform any manipulations with the choice of encoding, because names were displayed correctly already.

I just think that when using .nfo files there should be no problems. Because information is read directly from there.
And the problem is not that the scraper cannot download information. Scrapper downloads information, but it is displayed incorrectly in the interface.

Link to the full log:
https://yadi.sk/d/S8AkTNxgcibFlw
Reply
 
Thread Rating:
  • 0 Vote(s) - 0 Average



Logout Mark Read Team Forum Stats Members Help
Problem with adding films containing Russian letters in the file name from FTP00
This forum uses Lukasz Tkacz MyBB addons.