Database: Refactor MP1 database for multi seat usage (2 Viewers)

[Anthony Vaughan]

Never noticed this, used quite often, the last time was the day before yesterday on the test 1.29.

I did a complete reinstall of standard 1.29 and used 12 Years a Slave as my test movie. Scanned it in then did a re-scan with refresh and artist info selected.

I thought it was so weird that I'd better ask.

 

[Anthony Vaughan]

These are the settings I was using.

 

[Anthony Vaughan]

I've tried it again and it din't happen. Go figure .

I'll keep an eye on that. It must have been a glitch moving between versions.

 

[Anthony Vaughan]

In the IMDB processing, the processing goes to fetch actor details irrespective of whether the actor is already in the database.

I think it would be more efficient to check before fetching data from the internet to see whether we have it already.

Is there any reason I don't know about for not doing that?

This check really speeds things up (because it stops going for things we already have).

 

[catavolt]

The actor details are specific for that movie, e.g. his or her role...

 

[Anthony Vaughan]

The actor details are specific for that movie, e.g. his or her role...

Yes. I get that. It's a starting point - thinking about how to tackle this.

The thing is that I thought I would refresh the movies I had on my system - 119 of them. After a while I gave up because it was going to take hours to complete - with me sitting there clicking twice for each movie. I can't even let it get the job done behind the scenes. The first phase does go through without user intervention (collecting the movies to process). It's the second phase that is the killer - and it's the getting actors bit that takes so long.

I am thinking that maybe we should make the fetch process more intelligent. For example, if we have the actor details don't ask for it again. But we need to get the role the actor plays for each movie So, get the role for the given movie. Making that change reduces the task a bit. Then there's the graphics like covering art. If you've already got it, you don't need to get it again.

Maybe I'm trying to do the impossible or there isn't enough benefit in doing it. When you add movies over time, the problem seems less. But when you try to do a lot of movies in one go, it's a killer. I was just hoping to find a way to make the process more efficient.

Any, and all, ideas would be most welcome.

 

[Anthony Vaughan]

Oh. One other question. The actorinfo table doesn't seem to get populated though the data is collected in the fetch phase. It never has been - to my knowledge.

Is that a glitch? Is it something I should look into?

 

[framug]

Then there's the graphics like covering art. If you've already got it, you don't need to get it again.

Depending on movies, actors can have short hair, long hair, mustached or not, blend, brown or red hair, even bald.
Look at X-men, sometimes, blue skin for mystic, for example where, normal skin for another movie...

 

[Anthony Vaughan]

Depending on movies, actors can have short hair, long hair, mustached or not, blend, brown or red hair, even bald.
Look at X-men, sometimes, blue skin for mystic, for example where, normal skin for another movie...

True. But won't the images be per movie. I'm talking about when you already have a picture of the actor for that movie.

I get that there are two layers here: the actor and the movies that the actors were in.

Please keep the comments coming. This is really helping me understand the true nature of the issues .

 

[Anthony Vaughan]

Sorry to bother you again. I'm looking at the picture scan.

Suppose I have 100 pictures to scan in, the counter shows the picture count going from 1 to 100. At the end of the scan the total is shown as 200.

There's a bit of code that doubles the count (to account for different sized thumbnails.

if (!noLargeThumbnails)
{
// Display double count (for generate small and large thumbnail)
totalFiles = availableFiles.Count*2;
}
else
{
totalFiles = availableFiles.Count;
}

That's fine internally but as a user I am expecting to see that all 100 have been processed and don't really care that the system is holding 200 thumbnails.

I find the 200 final figure a bit misleading. Shouldn't the final figure be the actual number of images scanned in?

I also keep getting the following type of error:

[2021-11-04 08:55:11,534] [Config ] [PoolThread33] [ERROR] - Picture:ReCreateThumbnail Could not find file '\\safehelp08\MediaPortal TV Server\Thumbs\Test\Pictures\44\2133543444L.jpg'.

I think this is coming from Core. Despite the error, all of my pictures are in the database - so should it really be an error? I get stacks of these errors every time I do a scan of pictures.