Completed Show Intro Skip Option

[Cheesegeezer 3104]

2 minutes ago, rbjtech said:

Once a 'processed' flag has been set in the dB, those times are effectively locked.  At this point, the FP data is actually removed as it has served its intended purpose.  Any subsequent changes you make to the IntroStart/IntroEnd times are never changed.   I'm not 100% sure how modifications to existing data would reflect in the chapter tasks - @Cheesegeezer would this modification be picked up in subsequent chapter tasks ?

So once the chapter task has run, we check to see if the title sequence has already been added if it has we ignore it and move along. It would require a Refresh Metadata (no image refresh) to reset the chapters after modifying the chapter start and end points. Then rerun the chapter task.

so we could set a flag in the db that states an update has occurred if any changes are noticed. We could then act on that flag to reprocess chapters. 
 

editing the start and end times updates the start and end times immediately in db once saved, the only way to change those times are in the UI. This is proved when refreshing metadata and running the chaptertasks after editing them.

if you delete a show/season then you would need to run FP and detection again

[chef 3810]

7 minutes ago, rbjtech said:

Once a 'processed' flag has been set in the dB, those times are effectively locked.  At this point, the FP data is actually removed as it has served its intended purpose.  Any subsequent changes you make to the IntroStart/IntroEnd times are never changed.   I'm not 100% sure how modifications to existing data would reflect in the chapter tasks - @Cheesegeezer would this modification be picked up in subsequent chapter tasks ?

Not sure if this will help, but... The 'confirmed' bool in the database is not flagged as true until the data has been altered by the user.

So, if you edit the intro times and save the data, then it will be marked as confirmed in the database. 

I suppose that if you wanted to check for that boolean in the chapter task, and double check your chapter times, could be one way  to handle this.

[samuelqwe 317]

Just tried running the chapter insertion task (after first running both FP and Detection tasks), and this is the error I got:

One or more errors occurred. (Index was out of range. Must be non-negative and less than the size of the collection. (Parameter 'index'))
at System.Threading.Tasks.Task.Wait(Int32 millisecondsTimeout, CancellationToken cancellationToken)
at System.Threading.Tasks.Task.Wait(CancellationToken cancellationToken)
at IntroSkip.Chapters.ChapterEditScheduledTask.Execute(CancellationToken cancellationToken, IProgress`1 progress)
at Emby.Server.Implementations.ScheduledTasks.ScheduledTaskWorker.ExecuteInternal(TaskOptions options)

Seems like a scheduled task issue?

EDIT: This is on version 2.0.3.2 of the plugin, but Emby still says 2.0.3.0

Edited September 14, 2021 by samuelqwe

[Cheesegeezer 3104]

4 minutes ago, samuelqwe said:

Just tried running the chapter insertion task (after first running both FP and Detection tasks), and this is the error I got:
 

Seems like a scheduled task issue?

EDIT: This is on version 2.0.3.2 of the plugin, but Emby still says 2.0.3.0

this was what rbj picked up earlier in testing... and it’s fixed now. It was down to chapterInsertion class being instanced as static in the Task class which was fixed.

hmmmm should be 2.0.3.2 in the plugins section

[Cheesegeezer 3104]

11 minutes ago, samuelqwe said:

Just tried running the chapter insertion task (after first running both FP and Detection tasks), and this is the error I got:

One or more errors occurred. (Index was out of range. Must be non-negative and less than the size of the collection. (Parameter 'index'))
at System.Threading.Tasks.Task.Wait(Int32 millisecondsTimeout, CancellationToken cancellationToken)
at System.Threading.Tasks.Task.Wait(CancellationToken cancellationToken)
at IntroSkip.Chapters.ChapterEditScheduledTask.Execute(CancellationToken cancellationToken, IProgress`1 progress)
at Emby.Server.Implementations.ScheduledTasks.ScheduledTaskWorker.ExecuteInternal(TaskOptions options)

Seems like a scheduled task issue?

EDIT: This is on version 2.0.3.2 of the plugin, but Emby still says 2.0.3.0

Is the IntroSkip in the log now instead of APP or Server

also, needs to go in programdata\plugins not the old folder

Edited September 14, 2021 by Cheesegeezer

[samuelqwe 317]

14 minutes ago, Cheesegeezer said:

Is the IntroSkip in the log now instead of APP or Server

also, needs to go in programdata\plugins not the old folder

It does seem to be IntroSkip and not App, and I should have updated the plugin in the right directory.

Perhaps I downloaded the wrong file though? I just took the latest version from this thread.

EDIT: It did add chapters for some episodes before the task failed.

Edited September 14, 2021 by samuelqwe

[samuelqwe 317]

1 hour ago, chef said:

Should it be omitted using paths in the file system, or series to omit by ID?

I can picture it might involve a tab with a list of series, and an add button... or something like that.

Where should we keep the omitted series IDs? Xml? Or db?

I think you should be able to pick which librairies get scanned for intros. That’s probably a good start.

Then you can probably add excluding certain series and/or folders.

Is it not possible to save this information with the plugin’s configuration file somehow?

[Cheesegeezer 3104]

1 minute ago, samuelqwe said:

I think you should be able to pick which librairies get scanned for intros. That’s probably a good start.

Then you can probably add excluding certain series and/or folders.

Is it not possible to save this information with the plugin’s configuration file somehow?

Yup agreed, we can store all this in the database, only need to access one db for all info. Set a flag to ignore and use the SeriesId or FolderId to store and keep it in a nice list 

[Cheesegeezer 3104]

19 minutes ago, samuelqwe said:

It does seem to be IntroSkip and not App, and I should have updated the plugin in the right directory.

Perhaps I downloaded the wrong file though? I just took the latest version from this thread.

EDIT: It did add chapters for some episodes before the task failed.

Ok i will double check later, need to feed the livestock and get some dinner on.

[jad3675 26]

I'm running the latest release against my library, performing the initial finger print. I have process monitoring working through datadog, and you can see here that a 30 minute show took 20 seconds to run through and didn't consume much in the way of proc or memory. Granted, that's a simple 720p OTA recording done over an rclone mount, but it shows efficient it can be.

 

Good job guys. I'm impressed by everything you've done in this thread.

 

[rbjtech 5284]

13 hours ago, rbjtech said:

The time has come - I'm running this on my main system (Release) with the ffmpeg from the Beta - 550 shows / 30K episodes ..

It'll take a while (i've set it to just 2 sessions) - but CPU is about 4% per ffmpeg process, memory is under 200-210Mbyte.  So it's running in the 'background' happily without impacting other services.

I'll report back when it's finished the FP - it's the Detect where I had big issues before (I'm memory limited), but I think the memory changes to the SQL lookup should have solved the issue.  Either way, the FP needs to be done at some stage, so may as well do it now ..

.. so now at 40% complete  .. titlesequence.db is approaching 1 GByte - so estimate for completion is 2~2.5 Gbyte - this should test out the SQL query speed .. 

[Cheesegeezer 3104]

12 hours ago, samuelqwe said:

It does seem to be IntroSkip and not App, and I should have updated the plugin in the right directory.

Perhaps I downloaded the wrong file though? I just took the latest version from this thread.

EDIT: It did add chapters for some episodes before the task failed.

Hello mate, apologies for not getting back last night.

So i had a look thru and it would appear the file version was correct, however the Assembly version was stuck at 2.0.3.0.  I will post the correct dll in a sec.  Please use the new dll for the following tests.

Can you do the following for me please.

1. turn on debug logging in emby. 
2. do a library refresh for all TV Shows (dont refresh images, not required).
3. 
4. Restart emby (create a nice new log for us)
5. Check that the IntroSkip Advanced setting sliders are both on
6. 
7. Then run the chapter Insertion Task under scheduled tasks.
8. Once completed (or failed) PM me the full log if you would like to keep things private.

I will have a look

 

Edited September 15, 2021 by Cheesegeezer

[Cheesegeezer 3104]

New DLL

1. Update Version correctly
2. Logging mostly moved to debug now - for full logs you will need to enable debug logging in emby
3. Tidied up logging prefixes
4. double checked ChapterTasks methods and logics

IntroSkip 2.0.3.3.zip

[neik 873]

Just saw that writing the intros into an external file has been rejected and am wondering about the consequences of it.
Does this mean, every time one sets up a new Emby instance that this needs to be redone - just like the thumbnail extraction for example?

[Cheesegeezer 3104]

27 minutes ago, neik said:

Just saw that writing the intros into an external file has been rejected and am wondering about the consequences of it.
Does this mean, every time one sets up a new Emby instance that this needs to be redone - just like the thumbnail extraction for example?

No, because all the required information is stored in the TitleSequence.db so this can be backed up using the backup configuration utility and transferred easily - we don't want any other apps just using our data that's not inside the Emby Environment.

Edited September 15, 2021 by Cheesegeezer

[Micael456 45]

@chef,

Gosh, look away for a day and loads of progress. Looks amazing.

I was running on version 2.0.2.7. Want me to grab the latest and try again (main system doesn't get used too much during the working week), or are you looking to make some new improvements first? I see rbjtech is testing on his main now as well.

[Micael456 45]

26 minutes ago, Cheesegeezer said:

No, because all the required information is stored in the TitleSequence.db so this can be backed up using the backup configuration utility and transferred easily - we don't want any other apps just using our data that's not inside the Emby Environment.

@neik, it's also worth noting that uninstalling the plugin does *not* remove those files.

Though maybe plugin installation should check for existing DBs and prompt users on what to do? i.e. import or delete and start afresh. What do you think @Cheesegeezer, @chef?

Edited September 15, 2021 by Micael456

[CBers 7450]

2 hours ago, Cheesegeezer said:

Hello mate, apologies for not getting back last night.

So i had a look thru and it would appear the file version was correct, however the Assembly version was stuck at 2.0.3.0.  I will post the correct dll in a sec.  Please use the new dll for the following tests.

Can you do the following for me please.

1. turn on debug logging in emby. 
2. do a library refresh for all TV Shows (dont refresh images, not required).
3. 
4. Restart emby (create a nice new log for us)
5. Check that the IntroSkip Advanced setting sliders are both on
6. 
7. Then run the chapter Insertion Task under scheduled tasks.
8. Once completed (or failed) PM me the full log if you would like to keep things private.

I will have a look

 

I assume these are the initial steps required?

When I open the plugin config page, it defaults to a TV show that has no seasons currently.



What is the function of the Series field? Shouldn't the first option in the list be "All Shows"?

Also, shouldn't the field be "Show Name", rather than "Series"?

The default "Maximum parallel series to process at once" value is 2, but the text below the field says default is 4.



I've deleted the schedule tasks triggers, as suggested earlier in this thread - once all completed, I will put them back.

So what order do the tasks need to run in, as is on the Scheduled Tasks page?



Thanks.
 

[neik 873]

51 minutes ago, Cheesegeezer said:

No, because all the required information is stored in the TitleSequence.db so this can be backed up using the backup configuration utility and transferred easily - we don't want any other apps just using our data that's not inside the Emby Environment.

Gotcha, thanks for clarifying.

Giving the latest release a go on my main Emby instance, will probably take some time. Will report back once I've got anything to report.

Thanks for all the effort guys!

Edit: iirc then recordings are also being included (someone was talking about sport events I think), wouldn't it make sense to give the admins the opportunity to select or exclude some libraries? 

Edited September 15, 2021 by neik

[Cheesegeezer 3104]

25 minutes ago, Micael456 said:

@chef,

Gosh, look away for a day and loads of progress. Looks amazing.

I was running on version 2.0.2.7. Want me to grab the latest and try again (main system doesn't get used too much during the working week), or are you looking to make some new improvements first? I see rbjtech is testing on his main now as well.

Hey, did you do the scans on 2.0.2.7?  The only difference is some work done on the detection for lesser powered NAS's so it doesn't clog the memory.  If you have already ran FP and detection the only improvements are to the Chapter Insertion Tasks.  I have found another bug when doing some testing on my main system, for some reason I have a few episodes with only 1 chapter in them, not sure what's happened there, so I have written some more code to check for this and Flag in the log.

[Cheesegeezer 3104]

7 minutes ago, CBers said:

I assume these are the initial steps required?

When I open the plugin config page, it defaults to a TV show that has no seasons currently.



What is the function of the Series field? Shouldn't the first option in the list be "All Shows"?

Also, shouldn't the field be "Show Name", rather than "Series"?

The default "Maximum parallel series to process at once" value is 2, but the text below the field says default is 4.



I've deleted the schedule tasks triggers, as suggested earlier in this thread - once all completed, I will put them back.

So what order do the tasks need to run in, as is on the Scheduled Tasks page?



Thanks.
 

Ning me ole Corn Beef!!  I have been busy this morning

 

Edited September 15, 2021 by Cheesegeezer

[Cheesegeezer 3104]

6 minutes ago, neik said:

Edit: iirc then recordings are also being included (someone was talking about sport events I think), wouldn't it make sense to give the admins the opportunity to select or exclude some libraries? 

it's on the road map and will be implemented.

[Micael456 45]

11 minutes ago, Cheesegeezer said:

Hey, did you do the scans on 2.0.2.7?  The only difference is some work done on the detection for lesser powered NAS's so it doesn't clog the memory.  If you have already ran FP and detection the only improvements are to the Chapter Insertion Tasks.  I have found another bug when doing some testing on my main system, for some reason I have a few episodes with only 1 chapter in them, not sure what's happened there, so I have written some more code to check for this and Flag in the log.

Fingerprinting completed succesfully. Not sure where I am on detection as this is what kept silently killing my Emby instance.

[Cheesegeezer 3104]

9 minutes ago, Micael456 said:

Fingerprinting completed succesfully. Not sure where I am on detection as this is what kept silently killing my Emby instance.

So yes please update to 2.0.3.3 and run the detection process, hopefully the fix that chef has implemented will solve your issues.

Very interested to get all the information you have during this process if you don't mind.

Thanks

[CBers 7450]

25 minutes ago, Cheesegeezer said:

Ning me ole Corn Beef!!  I have been busy this morning

Cheers mate