Central FanArt Folder Improvments

[Tiki]

I had another thought that may apply to the scraping tools. I know we said that we want to leave the file names intact so that the scraper can compare file names and know what it already downloaded. However, how do we handle the case where some of the scraped fan art is poor quality and we want to delete it? If we simply delete, will the scraper automatically re-download?

[evilpenguin]

Originally Posted by Tiki

I had another thought that may apply to the scraping tools. I know we said that we want to leave the file names intact so that the scraper can compare file names and know what it already downloaded. However, how do we handle the case where some of the scraped fan art is poor quality and we want to delete it? If we simply delete, will the scraper automatically re-download?

I'm making it so my scrapper automatically downloads the first X number of images (default 3) for of each type. So yes, if you were to delete one of those, I would automatically re-download it every time you scraped.

A way around it would be to only download images if the folder doesn't already exist. That way the first time it would populate everything, but every subsequent time it wouldn't try to download anything, even if new fanart is added to the scraping website.

[MeInMaui]

What about appending a new extension (like .del or .rm) to the files to be ignored instead of deleting them? The scraper could check against those files as well as the originals and the fan art class could be updated to exclude files with that extension.

Aloha,
Mike

[jaminben]

I don't see this as an issue if we have the abilty to choose between three different images from within the SageMC UI. Its been my expirence so far that at least one of those three would be a good image and you'd be quite unlucky to get three bad ones. Even if you did get 3 bad images the idea of limiting the maximum downloaded images is to allow the user to do some work on their favorite images themselves rather than to spam TVDB for 30+ backgrounds in once instance.

Just my thoughts

edit:

Another thought might be to only download the three highest user rated images which should be the best three available.

[stuckless]

How about this.....

in each folder that contains metadata, have an optional "Images" files. This file contains the name of each image that has been added to the folder. When a scraper downloads an image, it's checks to see if the image name exists in the Images file. If it does, then the image is skipped, if not the image is downloaded and then it's name is added to the Images file.

This way, a user can manually remove images, and not have to worry about it being redownloaded. Also, users can rename images, or the scrapers can rename images for that matter, and not download the image more than once.

If a user wants to redownload fanart, then they'll have to manually remove the name from the Images file.

A sample images file may look like

poster_1920.jpg
poster_lowres.jpg

The names in the Images file must be the original filename as retrieved from the site.

This way the fanart apis don't need to be udpated, and it's up to scrapers to "honor" the Images file or not.

[Ikarius]

Stuckless,
Considering that various sites which contain fanart commonly have odd names for the images which do not conform to our naming schema, how should we tie an image a scraper has downloaded to a local filename? Can I assume the "images" file should contain both the remote source and the local filename?

[evilpenguin]

Originally Posted by Ikarius

Stuckless,
Considering that various sites which contain fanart commonly have odd names for the images which do not conform to our naming schema, how should we tie an image a scraper has downloaded to a local filename? Can I assume the "images" file should contain both the remote source and the local filename?

In Stuckless's plan the scraper wouldn't care what the local name was, it would just see that the images has previously been downloaded and skip it.

I personally like Mike's idea because checking for the existence of a file is easier than reading in a text file and doing string compares.

[Tiki]

Maybe another possibility would be to only automatically look for new fan art if no fan art exists. Once a folder is created, don't keep automatically checking for new art, but allow a manual method to force it to check for new art if desired.

I probably missed it in an earlier post, but what triggers the scraper to go out and look for new images?

[jaminben]

How would this also work for when you start getting new seasons? Would it see that its a new season and download new backgrounds, posters etc that are related to that season or would it just use what it already had?

Thought I'd throw that one into the pot as well

[stuckless]

While my ideas of using an "Images" file is a little more work for the scrapers, it actually a lot more intuitive for the end users... ie, they just delete fanart images that they don't want, and the folders are a little neater in the end. As it stands, I have to update the scraper to do "something", but with this solution, I don't have to update the phoenix api as well

jaminben, season artwork is downloaded into its own folder... so it would only look into its own folder for existing artwork. ie, the scrapers would not check all seasons for the existence of a piece of artwork (at least bmt won't )

[MeInMaui]

Just for the record, I like stuckless' plan best for handling deleted fan art. I really like that it makes it so the scraper no longer cares what the filenames are.

Aloha,
Mike

[VikingCrown]

Not sure if this one is relevant to the new format that is coming, but thought i throw it out there. Zap2it currently populates generic show information for some shows. Specifically The Starter Wife on UniversalHD. I am trying to record this for my wife, but each episode only has the same generic information for the series. I checked the info for the episodes and they all have the same show id as well as the same original air date. How would this new format treat an instance like this? I am currently sorting through the episodes to see what goes where, once i figure out which is which how should I go about renaming them to get the scrapers to pull the correct information?

If this doesn't belong here, please move it to somewhere more relevant. Can't wait to see what you guys have been working on, its gonna be awesome!

Regards,
Brian

[stuckless]

Instead of emailing "snapshots" of the phoenix jar to developers, I've decided to add a release to the downloads area. I'll post snapshots there as new functionality is added.

[Julianus]

Is the source code available?

[stuckless]

Phoenix is open source (apache 2.0), but currently the project is being developed by a limited # of people. I'll try to post src snapshots when I post binary builds.

[Fuzzy]

Regarding Actor images (I know, it's been pseudo hashed out and forgotten). Could it have a fallback to a generic Actor folder? So it would check:

TV\SeriesTitle\Season #\Actors\
TV\SeriesTitle\Actors\
\Actors\

and

Movies\MovieTitle\Actors\
\Actors

This would allow actor images to be displayed, even if there are no Show specific shots of the actor?