Plugin: EmbyCredits, detect end credits and add auto skip.

[Baenwort 124]

I've make sure CPU limit is 100, DMP is on, and I've set what I can figure out to Parallels processing. However, it doesn't seem to be using much resources on my server as I'm only seeing 15-20% resource usage. The server is at very low load at the moment with only ~4% cpu usage from other processes in other jails.

The settings I've adjusted:

 

[yocker 1859]

@Baenwort Looking at number of threads you have available for that CPU 30% usage sounds about right.
The FFmpeg usage is actually rather bursty so any CPU monitoring tool might not report it right either.

I can look into making some more "dedicated machine" settings to really let everything stretch it legs. Like for example increase batch size, looking into getting it to work with multiple tesseracts instances at once.
Atm. it's set rather conservative to make sure it wouldn't bog down peoples servers when running.

For now you can increase the speed a bit by disabling "Optimized Predictive Sampling in the OCR OCR Enhancements tab.

Edited August 22 by yocker

[Baenwort 124]

1 hour ago, yocker said:

@Baenwort Looking at number of threads you have available for that CPU 30% usage sounds about right.
The FFmpeg usage is actually rather bursty so any CPU monitoring tool might not report it right either.

I can look into making some more "dedicated machine" settings to really let everything stretch it legs. Like for example increase batch size, looking into getting it to work with multiple tesseracts instances at once.
Atm. it's set rather conservative to make sure it wouldn't bog down peoples servers when running.

For now you can increase the speed a bit by disabling "Optimized Predictive Sampling in the OCR OCR Enhancements tab.

The CPU used percentage isn't 30% as the FreeBSD htop command doesn't use 100% as the max but rather 100 x number of threads. If you look at the top of the htop window you will see most threads are under 30%. The CPU% column is the percent of a thread that is being used by that process. If a thread is using multiple threads fully then the percentage will show greater than 100.

I've disabled that and we'll see what the new calculation for time remaining levels out at.

[yocker 1859]

16 minutes ago, Baenwort said:

The CPU used percentage isn't 30% as the FreeBSD htop command doesn't use 100% as the max but rather 100 x number of threads. If you look at the top of the htop window you will see most threads are under 30%. The CPU% column is the percent of a thread that is being used by that process. If a thread is using multiple threads fully then the percentage will show greater than 100.

I've disabled that and we'll see what the new calculation for time remaining levels out at.

I meant to write 20% as you mentioned but guess i hit 3 instead of 2.  

Disabling it should speed it up a bit since it's an extra "detection method" used before the actual detection.
I should make it disabled by default tbh.

I've added what we discussed to my todo list and will start work on it next.

[yocker 1859]

@Baenwort did you see an improvement with the feature turned off?

[yocker 1859]

@BaenwortI have added some improvements, they are in the latest beta in the catalog.
Biggest improvement is being able to run multiple instances of local Tesseract detections.

[yocker 1859]

New version v2.7.1.3 in the catalog.

Added:
1) Ability to use multiple local Tesseract instances at once.
2) More control over CPU/GPU to improve performance on bigger machines.

Fixed:
1) Episode process bars should now show the correct info.
 

Edited Friday at 08:17 PM by yocker

[RedNo7 18]

question for understanding.

I have set a schedule for the plugin to run for 18 hours every day (as it needs to get through my whole library which will take a while).

The schedule screen shows the plugin ran correctly, but only for ~40 mins (1st screenshot).

When I run it manually, it states it needs ~450 hrs to complete (2nd screenshot).

Why is the scheduled task not running for the full 18 hours?

 

[yocker 1859]

2 hours ago, RedNo7 said:

question for understanding.

I have set a schedule for the plugin to run for 18 hours every day (as it needs to get through my whole library which will take a while).

The schedule screen shows the plugin ran correctly, but only for ~40 mins (1st screenshot).

When I run it manually, it states it needs ~450 hrs to complete (2nd screenshot).

Why is the scheduled task not running for the full 18 hours?

 

Did you set it to look in the correct folders?

Edited Tuesday at 10:53 AM by yocker

[RedNo7 18]

No - I cannot find those settings anywhere and it's not shown in the instructions. Where can I find them?

 

Sorry, Yes, I have these checked....

 

Edited Tuesday at 03:51 PM by RedNo7

[yocker 1859]

16 minutes ago, RedNo7 said:

No - I cannot find those settings anywhere and it's not shown in the instructions. Where can I find them?

 

Sorry, Yes, I have these checked....

 

And it still doesn't run detection for all videos with missing end credits?



Be sure this setting is disabled or it will only process new files since it last ran.

[RedNo7 18]

no, that's not enabled.

 

 

[yocker 1859]

23 minutes ago, RedNo7 said:

no, that's not enabled.

 

 

That's strange.
There is nothing that should stop it then other than some settings in Emby.



This is the schedule task in Emby.
See if the time limit is set as that will stop it.

Edit: I mean if it is set incorrectly.  

Also are these enabled in the plugin?

Edited Tuesday at 08:35 PM by yocker

[yocker 1859]

I'm considering making an automatic resume if a failure happens.. Problem will be having it working well together with Embys time limit setting.

[RedNo7 18]

Yes,

I have those checked.

But, interestingly, the schedule shows as this today:

I haven't knowingly changed any setting since reporting in this thread.

[yocker 1859]

I recommend keeping at least "skip videos with existing markers" on as other wise it will run detection on everything, even files that has already been processed.

Going to check later today if I forgot to add those settings to the data manager or else they get ignored and reset when changed.

Edited yesterday at 09:27 AM by yocker

[RedNo7 18]

this one?

 

[yocker 1859]

1 minute ago, RedNo7 said:

this one?

 

Yes

[RedNo7 18]

thanks!

[RedNo7 18]

back to as before this morning.  I've deleted it and created a new one (and started manual detection).

I'll check again tomorrow.

[yocker 1859]

1 hour ago, RedNo7 said:

back to as before this morning.  I've deleted it and created a new one (and started manual detection).

I'll check again tomorrow.

Strange. Pretty sure Emby doesn't have a built in time limit so might indeed be something wrong i just can't think about what i might be.

Do you by chance have log file i can see? The EmbyCredits settings file in the plugin folder would also be nice to see.
I'm working on a version that auto resumes if a fault happens but if the process gets actively canceled (by Emby) or some setting is wrong then it won't work.  

[RedNo7 18]

40 minutes ago, yocker said:

Do you by chance have log file i can see? The EmbyCredits settings file in the plugin folder would also be nice to see.

Sure - which log file are you referring to?  ...have added embyserer.txt in case that is what you are referring to.  "Save Debig logs to File" is not currently enabled in EmbyCredits - would you like me to turn it on?

EmbyCredits.xml attached

EmbyCredits.xml embyserver.txt

[yocker 1859]

30 minutes ago, RedNo7 said:

Sure - which log file are you referring to?  ...have added embyserer.txt in case that is what you are referring to.  "Save Debig logs to File" is not currently enabled in EmbyCredits - would you like me to turn it on?

EmbyCredits.xml attached

EmbyCredits.xml 12.86 kB · 0 downloads embyserver.txt 3.38 MB · 0 downloads

Saving debugs would be preferred but i will look in that log for now. Should be enough.
I've forgotten to say, the scheduled task is meant more for any new files and not for whole library detections. It should still work though.

I will look into it.

[yocker 1859]

@RedNo7Looking in the logs there is nothing wrong. It ran and stopped as it should.

2026-09-03 11:16:28.891 Info Plugin: Processing complete: 46 succeeded, 0 failed
It skipped a lot of files that already has end credits in them hence why it might have been so fast.

Have you enabled the schedule task to run in all the libraries that you want detections in?
I saw from your screenshot that you have 2 libraries and only one is enabled.

Edit: I have just uploaded a new version to the catalog as i found and fixed a small bug that could cause the schedule task to run in dry run.
In short i had set a line to be enabled and not disabled by mistake.
I don't think this is the problem you have but worth a try.

Edited 1 hour ago by yocker

[RedNo7 18]

50 minutes ago, yocker said:

Looking in the logs there is nothing wrong. It ran and stopped as it should.

Thanks, but I have 17,824 shows - has it done them all (according to the original ETA it was going to take 300 hrs to complete).

The manual processing says ~6,400 episodes skipped which surely means it's only ~1/3rd done?

 

50 minutes ago, yocker said:

Edit: I have just uploaded a new version to the catalog as i found and fixed a small bug that could cause the schedule task to run in dry run.
In short i had set a line to be enabled and not disabled by mistake.
I don't think this is the problem you have but worth a try.

I'll try it anyway

Edited 31 minutes ago by RedNo7