Jump to content

Cast & Crew Images Don't Display As They Should


freddyfinn
Go to solution Solved by freddyfinn,

Recommended Posts

levander
On 10/24/2020 at 9:01 AM, PenkethBoy said:

I dont have this issue but....

i modified the sql a bit so it's more specific to the problem rather than nuking all images from all items!


UPDATE MediaItems
SET Images = NULL
WHERE type=23

type 23 are people

so this removes the references for all image for just people

running the scheduled task - refresh people - makes emby look again and pulls the images from the cache i believe

@PenkethBoy @cayars  Is this actually safe to try to do for large libraries?

 

On 10/24/2020 at 9:01 AM, PenkethBoy said:

 

Now just doing that may not be enough to fix the problem - but you can add in the other Set X - Null if they are really needed

 

@PenkethBoy what do you mean by that this not be enough exactly?

Link to comment
Share on other sites

freddyfinn
52 minutes ago, PenkethBoy said:

Wow this escalated quickly

There are two things going on here that need to be separated

1. getting rid of incorrect image references from the db

2. Missing images for people

They are not the same issue - although to an end user they appear the same

for 1 the nulling of the image links with the sql i uploaded with the where clause fixes those images as it makes emby go an look in its cache and reload them.

for 2 the nulling by just image type is not enough as emby never had the images in its cache so they come back blank for those people - i suspect ,as have not tested this, but one of the other null'd fields kicks emby to go and have another look for the images with the refresh people task or a metadata refresh of the movie/tv show.

this is very likely PresentationUniqueKey = NULL or less likely ProviderIds = NULL

Somebody would have to test systematically to confirm my assumptions.

 

However, none of these work arounds are for the average user - thats on @Luke to add extra functionality to address both issues which have been outstanding for sometime. But if people are careful and to some extent adventurous then an enhanced version of the where sql will work - as i would not advise the whole sale nuking of 5 columns of data - as we dont know in detail, as not tested or supported what other implications it would have. If there are other issues that crop up then you are faced with using a backup or starting from scratch each has its obvious downsides.

Correct as I have said before removing the links on their own doesn't work which some here don't understand as English seems to be a second language. I have loaded this sql across five servers at the moment six including mine. All have db backups. Everyone happy so far. If anything happens I'll be first to know. My father hates things that he pays for not working and believe me do I hear about it. I'm not an sql writer don't profess to be and I've tried several combinations by manually filtering the data to get a result.

As I said before good idea @PenkethBoy however this needs to be crafted and tested and nobody has stepped up just bleeted like old goats and repeated themselves without a solution. It's also very frustrating when end users are pointed to a plugin that doesn't work or are told to reload everything. This doesn't wash with me or others that pay good money and end up trying to find a solution on their own while the issue sits in the too hard basket. 

If you can come up with something I will restore my old db and run the SQL to see if it kicks emby into updating the links and displaying them correctly. Should be a good test I've got 62000 lines. 

Link to comment
Share on other sites

PenkethBoy

Not going to do anymore on this as dont have the problem so i dont have the need was just pointing you to get a better targeted solution - you could actually test it yourself to see what is actually need to make it work specifically rather than the nuke it approach you have now - which is bad advice that others may follow and have problems - much better to have a specific targeted approach.

Link to comment
Share on other sites

PenkethBoy
1 hour ago, levander said:

Is this actually safe to try to do for large libraries?

 

 what do you mean by that this not be enough exactly?

yes - library size is irrelevant - just make sure you have a backup db file

you may need other fields set to null - depends on somebody (nobody has yet) actually testing further and which problem you are trying to fix.

  • Like 1
Link to comment
Share on other sites

Happy2Play

For anyone else here is a quick fix.

UPDATE MediaItems
SET Images = NULL
WHERE type=23
AND Images LIKE 'https://image.tmdb.org%';

This removes good and bad TMDB urls.

Then do a Refresh-Search for Missing or Refresh all if you like to repopulate image urls.

Don't have a system with bad urls, but modifying a couple to bad urls seemed to work.

 

  • Like 3
Link to comment
Share on other sites

levander
7 hours ago, Happy2Play said:

For anyone else here is a quick fix.


UPDATE MediaItems
SET Images = NULL
WHERE type=23
AND Images LIKE 'https://image.tmdb.org%';

This removes good and bad TMDB urls.

Then do a Refresh-Search for Missing or Refresh all if you like to repopulate image urls.

Don't have a system with bad urls, but modifying a couple to bad urls seemed to work.

 

Do you think that removing only bad tmdb urls (AND Images LIKE ‘https....?) would fix issue explained in this post?:

Or do both like you wrote?

 

Link to comment
Share on other sites

Happy2Play
1 hour ago, levander said:

Do you think that removing only bad tmdb urls (AND Images LIKE ‘https....?) would fix issue explained in this post?:

Or do both like you wrote?

 

All you should need is the update I posted and refresh/scan for missing on the library.  As all of those blank placeholders are bad TMDB url (all placeholders that do not show a name on placeholder)  if there is a name on placeholder that already means there is no image in database.

Not sure what you mean by both though.  It is one command filtering by type 23 and all Images that contain a TMDB url, and setting only them to null.

Link to comment
Share on other sites

levander
52 minutes ago, Happy2Play said:

 

Not sure what you mean by both though.  It is one command filtering by type 23 and all Images that contain a TMDB url, and setting only them to null.

Because you said that it removes GOOD and BAD tmdb urls. So I assumed that GOOD was defined with part that sets all image links to null (type=23) and BAD urls was defined with that after ‘AND’ part (images LIKE...). That is what I I meant with both.

Link to comment
Share on other sites

Happy2Play
26 minutes ago, levander said:

Because you said that it removes GOOD and BAD tmdb urls. So I assumed that GOOD was defined with part that sets all image links to null (type=23) and BAD urls was defined with that after ‘AND’ part (images LIKE...). That is what I I meant with both.

There is no way to distinguish good or bad so we have to clear all urls specific to TMDB.

Link to comment
Share on other sites

levander
On 10/26/2020 at 3:59 AM, Happy2Play said:

For anyone else here is a quick fix.


UPDATE MediaItems
SET Images = NULL
WHERE type=23
AND Images LIKE 'https://image.tmdb.org%';

This removes good and bad TMDB urls.

Then do a Refresh-Search for Missing or Refresh all if you like to repopulate image urls.

Don't have a system with bad urls, but modifying a couple to bad urls seemed to work.

 

Hey thanks to you and @PenkethBoy for this! This method seems to work just fine only it took 8+ hours to finish in my case and that was only for movie libraries. Btw, do you know does Full Scan Library does the same as refresh missing metadata or is it important to do refresh missing first? 

Link to comment
Share on other sites

levander
9 minutes ago, PenkethBoy said:

8hrs - that included the lib scan etc - right?

Executing SQL with DB Browser took around 20 minutes on my library.db (~1Gb)

Refreshing missing metadata only on Movie library took more than 8 hours for sure but could be much more. (Not scan all libraries)

 Mac mini 2012

Edited by levander
Link to comment
Share on other sites

Happy2Play
1 hour ago, levander said:

Is there any way to perserve people images so if something similar happens that I can simply restore it without Internet?

That is the purpose of the People Backup plugin.  You allow Emby to import all your existing metadata and images offline, then you restore all the backed up people images.  You have to wait on the initial import to add the person to the database so the restore can apply info to it.  Same name actors will be a issue though,

But this has no effect if TMDB does house cleaning again and you do not have a local image.

Edited by Happy2Play
Link to comment
Share on other sites

levander
14 hours ago, Happy2Play said:

That is the purpose of the People Backup plugin.  You allow Emby to import all your existing metadata and images offline, then you restore all the backed up people images.  You have to wait on the initial import to add the person to the database so the restore can apply info to it.  Same name actors will be a issue though,

But this has no effect if TMDB does house cleaning again and you do not have a local image.

I’m not quite sure I understood everything you said. Let me try to summarize and you correct me if I’m wrong: 

People backup plugin can restore all people info + images that match those already in the database but could mix up people with same name?

 Questions:

- does people tags get backup up (since people backup plugin is older than people tags feature)?

- why does people with same name get mixed up if there is people id in people metadata?

- if for example plugin has backed up 10 000 people from one library setup and you use backup again in new library setup where there are less people imported does plugin updates/merge old backup with new data or overwrites it?

- what does the last sentence you wrote means? Can you elaborate little more?

Link to comment
Share on other sites

Happy2Play
2 minutes ago, levander said:

I’m not quite sure I understood everything you said. Let me try to summarize and you correct me if I’m wrong: 

People backup plugin can restore all people info + images that match those already in the database but could mix up people with same name?

 Questions:

- does people tags get backup up (since people backup plugin is older than people tags feature)?

- why does people with same name get mixed up if there is people id in people metadata?

- if for example plugin has backed up 10 000 people from one library setup and you use backup again in new library setup where there are less people imported does plugin updates/merge old backup with new data or overwrites it?

The issue with the plugin is it saves in the legacy format (by name folders) so it is impossible to do same name folder.  Same name people is one reason Emby has moved away from people by name.  

@pünktchen can  you comment more on this.

 

Link to comment
Share on other sites

pünktchen
55 minutes ago, levander said:

does people tags get backup up (since people backup plugin is older than people tags feature)?

Yes.

43 minutes ago, Happy2Play said:

why does people with same name get mixed up if there is people id in people metadata?

Happy2Play already answered this. It's because of historical reasons.

56 minutes ago, levander said:

if for example plugin has backed up 10 000 people from one library setup and you use backup again in new library setup where there are less people imported does plugin updates/merge old backup with new data or overwrites it?

A backup always overwrites existing data, but it won't delete existing data if the new backup has less people than the older existing backup.

  • Like 1
Link to comment
Share on other sites

  • 4 weeks later...
On 10/23/2020 at 9:20 PM, freddyfinn said:

 

I take no responsibility if this fails so make a backup!

Superfast way to fix this issue. 

Shutdown Emby Server

Download SQLite .  I used the portable version.

Go into C:\Users\(YourAccountName)\AppData\Roaming\Emby-Server\programdata\data

And make backup of library.db

Open SQLite Select 'Open Database' and navigate to C:\Users\(YourAccountName)\AppData\Roaming\Emby-Server\programdata\data

and open 'library.db'.

Download the attached 'Emby.sql'

In SQLite go to the 'Execute SQL' tab and load Emby.sql

Press the play button to Execute SQL

Close the database and SQLite.

Start Emby Server go into server settings and open 'Library'

Do a Metadata refresh by clicking on the three buttons on each folder.

Select 'Search For Missing Metadata' and you can leave 'Replace Existing Images unchecked' 

Done!

 

 

Emby.sql 148 B · 6 downloads

Thank you for this @freddyfinn . As I have a lot of actors' images missing, I wanted to try this.

I used DB browser for SQLite on openSUSE and got this result:

no such column: UserDataKey: UPDATE MediaItems
SET ProviderIds = NULL,
    PresentationUniqueKey = NULL,
    Images = NULL,
    UserDataKeyID = NULL,
    UserDataKey = NULL

Am I doing anything wrong?

Would appreciate any help. Thanks.

Edited by Simorgh
typo
Link to comment
Share on other sites

Please don't run that SQL as it will wipe out all images for everything and not just actors.

Use what Happy2Play posted which is

UPDATE MediaItems
SET Images = NULL
WHERE type=23
AND Images LIKE 'https://image.tmdb.org%';

That will only wipe actor images from tmdb.org so that only those images need to be reloaded.

The database to edit is library.db

  • Thanks 1
Link to comment
Share on other sites

  • 2 years later...
parhamsan

Hello,

I know I am late to the party but recently I have realized that emby is having issues with detecting the correct cast and crew.

For example for the guardians of the galaxy 3 it has recognized Caitriona Balfe as drax the destroyer; also Curtiss Cook as Nebula, aslo Brian Huskey as Groot, also Rocket Voice is wrong!!!

image.png

This also happens to Tv shows:

Hijack Tv Show, Jesse Eisenberg as Sam Nelson, which supposed to be Idris Elba.

image.thumb.png.289ab1feafff3a63265d43b6c8ad33a0.png

I have tried identifying using IMDB Id and TheTVDB Id and refreshing the Metadata after but same result happens.

Any help would be appreciated.

Thanks.

 

Edited by parhamsan
Link to comment
Share on other sites

1 hour ago, parhamsan said:

Hello,

I know I am late to the party but recently I have realized that emby is having issues with detecting the correct cast and crew.

For example for the guardians of the galaxy 3 it has recognized Caitriona Balfe as drax the destroyer; also Curtiss Cook as Nebula, aslo Brian Huskey as Groot, also Rocket Voice is wrong!!!

image.png

This also happens to Tv shows:

Hijack Tv Show, Jesse Eisenberg as Sam Nelson, which supposed to be Idris Elba.

image.thumb.png.289ab1feafff3a63265d43b6c8ad33a0.png

I have tried identifying using IMDB Id and TheTVDB Id and refreshing the Metadata after but same result happens.

Any help would be appreciated.

Thanks.

 

Hi, please try refreshing the metadata on that movie and then attach the emby server log. thanks.

Link to comment
Share on other sites

parhamsan
39 minutes ago, Luke said:

Hi, please try refreshing the metadata on that movie and then attach the emby server log. thanks.

Hi Luke,

I have attached the part of the log file right after I refreshed the metadata on the movie.

 

The odd thing is after I changed the "Series Metadata Downloader" to TheTVDB (on my TV shows library), now the Hijack TV show is getting the correct cast and crew:

image.thumb.png.d78e0474fc375f32ad60c24b1d865878.png

image.thumb.png.655ea2bbf64e11a9681a94261fb616b1.png 

Let me know if you want the whole log file to be attached.

 

Thanks.

embyserver.txt

Link to comment
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now
×
×
  • Create New...