Refresh/forget/identify content and scan again

[xe` 44]

As far as I can tell based on my testing there is no "in GUI" way to truly refresh content as if it is just being seen new.

 

Currently if something goes wrong with the first tagging process you cannot "start again" without renaming or otherwise altering the actual filesystem data to cause Emby to see the old copy as gone and new content arriving.

 

Given that there will always be edge case bugs with identification, metadata extraction and tagging there is significant value in adding a native option to truly forget a video and immediately scan it again as if it is new content.

[Luke 40111]

The identify feature is pretty close to treating as new.

[xe` 44]

Perhaps I am missing something but the identify process requires that you must manually add new information. You should not "have" to manually enter information to get a "pretty close" refresh.

 

I can cite specific examples of how the tagging process has gone wrong and simply starting it again will fix the issue but I am not sure that specific examples add value...  since it is without question that tagging bugs have happened and will keep happening (and be fixed) it should be implicit that being able to recover from them in-gui should be possible?

 

Lets call it a "refresh of last resort". Failing that just being able to delete the Emby entry (without touching source data) letting it be found again next scan achieves the same end goal and is possibly an easier request to dimension.

[Luke 40111]

Correct, we don't have a clear metadata function, but if you need to change it to something else then identify will do the trick.

[xe` 44]

OK great. I think given the discussion above this is proven to be a valid feature request and I wont labor the point. Please give it some serious consideration.

[xe` 44]

I want to touch on this again. During the 4.0 cycle a large amount of my content was misidentified by two bugs that have since been closed in 4.2.

 

The problem I have is that I cannot find a way to fix the left over garbage.

 

Each movie has a correct "URL only NFO" but there is simply no way to instruct Emby to treat the misidentified movie as new and scan again (fixing the issue since the bugs are now closed).

 

Manually identifying the movies will be a massive chore.

 

Can anyone suggest any means to "trick" Emby into scanning a movie again ideally without having to rename every movie folder at source.

[Luke 40111]

I suppose one way is temporarily moving the folders out of your library, running a full scan to clear them out of the database, and then adding them back.

[xe` 44]

I can certainly do this and it would be pretty easy to pull off (albeit slow) what I have never understood properly though is watched status.

 

Would this procedure also lose my historical "watched" stats?

[Luke 40111]

I can certainly do this and it would be pretty easy to pull off (albeit slow) what I have never understood properly though is watched status.

 

Would this procedure also lose my historical "watched" stats?

 

Depends. For movies, we store watch data based on moviedb/imdb id's. It sounds like for some titles you want to change those values so that means you're going to lose watch data anyway.

 

You shouldn't lose watch data for other titles that aren't changing, but the safest thing to do would be to only do this procedure on the individual movie folders that need changing.

[xe` 44]

I just rediscovered myself posting about this issue whilst debugging and looking for an answer to this issue.

 

Summary,

Neither the GUI or API can rerun the initial movie identification automation, they can only rerun the metadata scraper.

Request an Emby option to allow content to be forgotten so that the next update can pick it up as new.

 

 

 

 

[GrimReaper 4157]

Out of curiosity, wouldn't that just make a movie be mis-identified once again, landing you in the same spot of having to identify it manually? 

[xe` 44]

I dont think so, certainly not all the time.

Example, if Emby reads a URL only nfo with wrong TTID the match is permanent even if the nfo is subsequently corrected. You simply cannot natively make Emby read this new corrected value without tricking it into thinking the old movie has been deleted first. i.e. renaming folders or moving folders out of your library, full re-scan and then a full metadata scan again.

For context this is not always nfo error, on a couple of occasions bugs in Emby have dropped support for some matches and even though they are fixed quickly there is no easy recovery.

 

[GrimReaper 4157]

11 minutes ago, xe` said:

if Emby reads a URL only nfo with wrong TTID the match is permanent even if the nfo is subsequently corrected.

Than that is the thing that should be corrected, rather than implementing any new feature, IMHO. Don't have URL-only NFOs, but on regular Kodi-style ones, any timestamp change triggers re-reading/re-importing changed data upon scan, don't know why that shouldn't apply to URL-only NFOs, @Luke? 

[Luke 40111]

3 hours ago, GrimReaper said:

Than that is the thing that should be corrected, rather than implementing any new feature, IMHO. Don't have URL-only NFOs, but on regular Kodi-style ones, any timestamp change triggers re-reading/re-importing changed data upon scan, don't know why that shouldn't apply to URL-only NFOs, @Luke? 

It does. This is just a special case that needs some additional handling added.

[xe` 44]

Do I need to produce anything else to get this request accepted (the timestamp special case handling)? Happy to do any user side testing/legwork but it seems from above we might be there now?

[Luke 40111]

1 hour ago, xe` said:

Do I need to produce anything else to get this request accepted (the timestamp special case handling)? Happy to do any user side testing/legwork but it seems from above we might be there now?

No, we have enough. Thanks.

[xe` 44]

What is the correct way to track this fix.

Since I reported the issue I have been doing my duty and diligently reviewing release notes and testing each release.

Now that I have done this for three releases it feels like testing every release is a sub-optimal way of going about this.

[Luke 40111]

Hi, best thing to do is follow the release notes for new releases. Thanks.

[xe` 44]

Is there anything I can do to help make this fix happen?

I find myself 3 years 8 months and 7 days since my first post still fighting the exact same problems, doing all the laborious manual fixes, just for the lack of Emby detecting nfo timestamp changes.

I dread to to even think how many hours I have spent on this.

Save me please.

[Luke 40111]

If you clear the metadata and refresh it again, aren't you just going to end up with the same result? Seems like you need to clear, edit and lock.

[xe` 44]

Thanks for replying...

The source of truth for TV and Movie identification on my setup are URL only nfos (now referred to as parsing NFO file defined by this standard).

I have 100% coverage for all content which allows for all tools (not just Emby) to identify with absolute certainty for all time what the show/movie is.

I do not use metadata nfos as my goal is to just define the id once, "anchoring" the match, whilst still allowing any client to scrape metadata when and how they want.

 

If the URL only nfo is set initially incorrectly I will fix it at source. Emby never notices this fix and has no native tools to tell it to refresh the content. Currently I have to manually enter into the identify setting and add the tvdb/imdb.

 

I am requesting that Emby notice when a URL only nfo changes and do this for me.

[Luke 40111]

OK yes for url only nfo's, if the url is changed then currently the server doesn't react to that. It's something that needs to be looked at. Thanks.

[xe` 44]

Can I ask for an update on this.

Emby displaying an incorrect movie until I manually fix it by copying info from an nfo into the webgui is a laborious task that should be automatic.

Please please fix this, honestly it is soul destroying having to do this manually for over four years now when you could probably fix this in no time at all.

[Luke 40111]

3 hours ago, xe` said:

Can I ask for an update on this.

Emby displaying an incorrect movie until I manually fix it by copying info from an nfo into the webgui is a laborious task that should be automatic.

Please please fix this, honestly it is soul destroying having to do this manually for over four years now when you could probably fix this in no time at all.

I tested this a couple months ago on the 4.8 beta channel and didn’t see a problem , so hopefully it is resolved now.

[xe` 44]

I was hopeful this was fixed but the betas never seem to make it to stable now. Is there any chance of a backport or an ETA?