New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
[SUGGESTION] French language detection with subs and audio #2275
Comments
Are you saying I think the issue with this is wanting to move away from sub language being used to determine the language, since the audio may still be English and the subs may not be hardcoded. v3 will offer the ability to have multiple languages in a profile so it will be a little more flexible and allow further filtering with restrictions. |
"wanting to move away from sub language being used to determine the language" Well, yes I would like to move away, but currently it's not. Actually VOSTFR = VO+STFR = SUBFRENCH = FASTSUB, etc etc So, I succeed to solve my problem by making a script before Sonarr parse results to "fakely" rename releases on the fly (using Jackett), if it contains one of the following strings (VO+STFR, FASTSUB, SUBFRENCH, MULTI, NF), it adds VOSTFR in the name to be parsed and retained by Sonarr successfully. Of course, as you said it, V3 should solve the problem by allowing French and English and by adding tag to filter results. But, V3 is not here yet... ;) |
you can also add VFF which is the same as TRUEFRENCH |
To sum up (even if not used), FRENCH language could be detected in these cases: audio:FR : FRench subtitles:VOSTFR Exception case:(means French is not explicitly named) At least, in my opinion, audio and subtitles tags should be added. |
would be nice for sonarr to not wronly detect a show as foreign if a language is used in the ep title.. |
@minitoine would you be ok to share your script ? I'd be interested in using it on my server :) |
It is more a trick than a script :) I use Jackett as a source of releases. For exemple, for an definition file (at search function definition): title_normal: And it works. |
FASTSUB is by no mean linked to French language. It just means the subs where translated with Google Translate and corrected only for major translation problems (hence the FAST). Maybe some teams use it for VOSTFR, but then they are wrong, because you can have fastsub for spanish subtitles too. The rule is that you specifiy your language and then specify if PROPER or FASTSUB translation. |
Yeah, I know. But as you mentioned, a lot of teams do that.. :'( |
When consuming data from a non english tracker (like YGG for us french speaking users) I guess we can safely assume that FASTSUB means that the release is subbed in the tracker language. @minitoine snippet is great in this situation. |
I'm facing the completly opposite issue : i want VOSTFR to be caught as being english ( in ygg ). Sonarr, thinking thoose files are french audio, does not take them, but i would like him to understand thoose VOSTFR tags as still "English audio". How could i adapt your script to do that ? |
Yo Irnv, maybe by replacing vostfr by en ?
|
Thanks very much, worked for a while but i still get some error in sonarr's detection. However, someone talked here about the V3 that would implement correct separation between spoken language of the show and subtitles. Do you know where i can get more informations about that ? Is there a alpha version of something ? |
@lrnv To make it work, I had to uncheck within Jackett WebUI on YGGtorrent indexer:
And then within title_normal:
selector: "a.torrent-name"
filters:
- name: replace
args: ["VOSTFR", "en"] |
Hi @minitoine |
@jenkinsddl Well, give me some files to work on or something. I guess your pb has no links with Jackett. |
@minitoine Abn : https://github.com/Jackett/Jackett/blob/master/src/Jackett.Common/Indexers/Abnormal.cs |
@jenkinsddl Try replacing :
to something like
Compile Jackett, install, should be ok GL ;) |
@lrnv About V3, we all don't know. |
Thank you once again minitoine 👍 Got it working for xthor, abn and ygg ! here are the changes I made
ABN:
xThor was a bit different (probably because it's a different api or something ...). You have a section
And of course for xthor and Abn you'll have to rebuild it from source. Don't forget to disable auto update or the changes will be overwritten. |
Hi, I use sonnar but not with torrent but nzb, so i don't use jackett and I cannot use the trick of minitoine (good trick btw). I use NZBHydra who fetches information mostly from my prefered indexer : NZBGeek. It index a lot of French content but mostly with this tag: "multi" (means French language and original langage). I know it won't satisfy everybody but i realy whant in my system that when the parser see "multi" he know it's french content and auto download. I've spend hours trying to modify the parser in the src code and recompile it (installing glup and other npm things on my server) whitout any success I've tryed to find an option in nzbhydra allowing me to rename result before sending to sonnarr, no success neither. So
Thanks' for your help. |
Hi ! If you look at there( src/NzbDrone.Core/Parser/LanguageParser.cs file): Line 14, you have the regex used to detect French. From what you said, you didn't succeed to compile your own Sonarr ? Or you succeed, but your modifications didn't change anything on the french detection ? Keep us in touch. |
Hi, Following your advice i've gone deeper in the compile work. Actually I had one step missing who seems obvious now : Visual Studio. So i download VS, Git and gulp on my windows computer. Open the VS project, change the language parser thing and rebuild the projet. Run the .exe : Great news it work... on windows. I can see Sonarr detecting a "multi" release as french ! After that i :
An I have all those error, i'm completely lost...
|
Don't you have any target thing before compiling the project ? (Something like a mono target ?) I guess you cannot just copy your build for windows on Linux based computers. You should have, I guess, some configuration to do before, no ? Unfortunately, I cannot help you more in this way, maybe try to contact a Sonarr Dev.. |
Ok, found a workaround: It works with the application : Maybe it's because in the Readme they said
So i have done this in Visual Studio. I should try putting the NzbDrone.Core as Startup in visual studio and try again. Right after that I will write here the all process if someone whants to try to modify the code for his own. I don't know what would happen when an update will be avaiable... |
Clearly, if you update Sonarr, the build will be replaced (If Sonarr updates on its own). But, if you pulled the GitHub project, and modified some files. You still can pull updates and see if it merges your modifications. After this, you should be able to compile again with the update. I guess you should write a shell, or whatever, script to do all this stuff automatically. Good luck. |
Just to let you know, I have the same "problem" and I fixed it by settings my quality set to French language and then adding mandatory words like VOSTFR, SUBFRENCH, etc. From my point of view, the assumption that VO is English is very bad because it can be German (DARK) or Spanish (La Casa de papal) may be we need an "original" language option? |
v3 will have the ability to prefer otherwise equal releases as well as multiple languages per profile (similar to qualities), which should help in some of these cases. Closing this, but we can reconsider after v3 is out if necessary. |
Hi,
The language parsed for file containing VO+STFR is not correct. Here English is found.
Also, FASTSUB or SUBFRENCH are often used on release names and conduct to the same error.
Could you add this to a french detection ?
(?(?:\W|)(?:FR|VOSTFR)(?:\W|)) to
(?(?:\W|)(?:FR|VOSTFR|VO+STFR|SUBFRENCH)(?:\W|))
Or remove FR and VOSTFR detection and set it to English as we can use tags behind to filter VOSTFR releases with English.
As we still need need to specify a langage in a profile, we are blocked in France with release names.
Or let us have profile without Language and let us filter releases with tags
Thanks,
minitoine
Trace:
17-11-7 17:08:52.2|Debug|NzbSearchService|Total of 3 reports were found for [The Shannara Chronicles : S02E04] from 6 indexers
17-11-7 17:08:52.2|Info|DownloadDecisionMaker|Processing 3 releases
17-11-7 17:08:52.2|Debug|Parser|Parsing string 'The Shannara Chronicles S02E04 1080p AMZN WEBRip DDP2 0 x264-ViSUM (VO+STFR)'
17-11-7 17:08:52.2|Debug|Parser|Episode Parsed. The Shannara Chronicles - S02E04
17-11-7 17:08:52.2|Debug|Parser|Language parsed: English
17-11-7 17:08:52.2|Debug|QualityParser|Trying to parse quality for The Shannara Chronicles S02E04 1080p AMZN WEBRip DDP2 0 x264-ViSUM (VO+STFR)
17-11-7 17:08:52.2|Debug|Parser|Quality parsed: WEBDL-1080p v1
17-11-7 17:08:52.2|Debug|Parser|Release Group parsed: ViSUM
17-11-7 17:08:52.3|Debug|AcceptableSizeSpecification|Beginning size check for: The Shannara Chronicles S02E04 1080p AMZN WEBRip DDP2 0 x264-ViSUM (VO+STFR)
17-11-7 17:08:52.3|Debug|AcceptableSizeSpecification|Item: The Shannara Chronicles S02E04 1080p AMZN WEBRip DDP2 0 x264-ViSUM (VO+STFR), meets size constraints.
17-11-7 17:08:52.3|Debug|CutoffSpecification|Comparing file quality with report. Existing file is HDTV-720p v1
17-11-7 17:08:52.3|Debug|LanguageSpecification|Checking if report meets language requirements. English
17-11-7 17:08:52.3|Debug|LanguageSpecification|Report Language: English rejected because it is not wanted, wanted French
17-11-7 17:08:52.3|Debug|ReleaseRestrictionsSpecification|Checking if release meets restrictions: The Shannara Chronicles S02E04 1080p AMZN WEBRip DDP2 0 x264-ViSUM (VO+STFR)
17-11-7 17:08:52.3|Debug|ReleaseRestrictionsSpecification|[The Shannara Chronicles S02E04 1080p AMZN WEBRip DDP2 0 x264-ViSUM (VO+STFR)] No restrictions apply, allowing
17-11-7 17:08:52.3|Debug|QualityAllowedByProfileSpecification|Checking if report meets quality requirements. WEBDL-1080p v1
17-11-7 17:08:52.3|Debug|MinimumAgeSpecification|Not checking minimum age requirement for non-usenet report
17-11-7 17:08:52.3|Debug|RetentionSpecification|Not checking retention requirement for non-usenet report
17-11-7 17:08:52.4|Debug|UpgradeDiskSpecification|Comparing file quality with report. Existing file is HDTV-720p v1
17-11-7 17:08:52.4|Debug|SeriesSpecification|Checking if series matches searched series
17-11-7 17:08:52.4|Debug|DelaySpecification|Ignoring delay for user invoked search
17-11-7 17:08:52.4|Debug|HistorySpecification|Skipping history check during search
17-11-7 17:08:52.4|Debug|MonitoredEpisodeSpecification|Skipping monitored check during search
17-11-7 17:08:52.4|Debug|DownloadDecisionMaker|Release rejected for the following reasons: [Permanent] French is wanted, but found English
The text was updated successfully, but these errors were encountered: