Updated image extraction to pull cover image rather than one of the s…

[TheStashMaster]

…ample images.

Generated by an automatic template. Can be removed if not applicable.

Scraper type(s)

• performerByName
• performerByFragment
• performerByURL
• sceneByName
• sceneByQueryFragment
• sceneByFragment
• sceneByURL
• groupByURL
• galleryByFragment
• galleryByURL
• imageByFragment
• imageByURL

Examples to test

https://vrconk.com/video/god-of-war-freya-a-porn-parody/
https://arporn.com/video/special-after-party/
https://vrbangers.com/video/panty-hunter/

Images were pulling sample images which were inconsistent. I've updated it to retrieve the cover art.

[feederbox826]

cc @nrg101 your change targeted the image selector #2259

[nrg101]

ok, I'll check and compare the resulting scraped images, hopefully later today

[nrg101]

I have to use CDP configured with a VPN in order to access the sites... this doesn't resolve the site access for some sites for me

Here is what I get with my current version of the scraper:

• https://vrconk.com/video/god-of-war-freya-a-porn-parody/ --> scraping fails with error:
  parse error: expected string near offset 1144 of 'partitionKey'
• https://arporn.com/video/special-after-party/ --> the cover image
• https://vrbangers.com/video/panty-hunter/ --> parse error: expected string near offset 898 of 'partitionKey'

afaik, this partitionKey error is something to do with the CDP-seen response rather than a scraped page

[nrg101]

For the arporn example, the proposed xpath for the image can be simplified to:

//section[@id="video-page-banner"]/picture/div[contains(@class, "app-image")]/img/@src

This resolves to https://content.arporn.com/uploads/2024/12/676338ee0c168851352077.jpg which is 2000x714 (very wide aspect) which will result in the sides being cropped in a Stash thumbnail. I would consider this image more of a banner image for the scene.

The existing xpath resolves the first expression to https://content.arporn.com/uploads/2024/12/676338ee0c944958049212.jpg which is 1300x731 (16:9 aspect ratio) which is much more suitable in a Stash thumbnail... this image is the one that is shown in the video player.

The proposed xpath fetches a very wide aspect ratio version of the cover image which isn't really suitable for Stash thumbnails

vrconk:

proposed: https://content.vrconk.com/uploads/2023/04/643f0910087a8445965650.jpg --> 2000x714 (very wide aspect ratio)
existing: https://content.vrconk.com/uploads/2023/04/644049a847cf8627744787.jpg --> 850x489 (near to 16:9 aspect ratio)

vrbangers:

proposed: https://content.vrbangers.com/uploads/2024/11/6733d704cba33825192694.jpg --> 2000x714 (very wide aspect ratio)
existing: https://content.vrbangers.com/uploads/2024/11/6733d704cbe4d773348800.jpg --> not a cover image, looks like just one of the photos from the photo set

so, from this limited set of example URLs, arporn and vrconk look like they can use the image of the video player as that is a 16:9 version of the cover image, whereas vrbangers seems to at least sometimes just use an image from the photo set

what is slightly annoying is that the 16:9 cover image version for "Panty Hunter" does exist, it's just that you often have to find it at vrporn instead, e.g.

https://vrporn.com/panty-hunter/ gives https://mcdn.vrporn.com/files/20250213003950/Panty-Hunter-Melanie-Marie-VR-Bangers-vr-porn-video.jpg which is 1300x731 (16:9 aspect ratio)

interestingly, https://vrporn.com/special-after-party/ is also there, as is https://vrporn.com/god-of-war-freya-a-porn-parody/ which gives a 1300x731 cover image https://mcdn.vrporn.com/files/20230811083807/God-Of-War-Freya-A-Porn-Parody-Armani-Black-VRConk-vr-porn-video-1.jpg, higher resolution than at vrconk.com

as the slugs are identical, maybe this scraper could add a secondary URL for https://vrporn.com/<the-slug-from-the-studio-site> as that should be able to consistently scrape a 1300x731 (16:9) version of the cover image (albeit each scene release at vrporn.com tends to lag behind the studio site, which is acceptable imo)

[nrg101]

TL;DR:

• I disagree with scraping the 2000x714 versions of the image as a cover, as these end up heavily cropped at the sides in Stash thumbnails.
• The existing xpath expression seems to work most of the time
  • https://stashdb.org/edits/36ef2940-da1b-4b18-9947-7825e130c4e5
• Adding a secondary URL at vrporn.com could provide a convenient way to get a 1300x731 (16:9) version of the cover image (I have done this manually many times for VRB site scenes, so this would be very useful to me)
  • https://stashdb.org/edits/eccbeb09-9bd0-44bb-af46-b2aa9b5d128d
  • https://stashdb.org/edits/81e9723c-c987-4ed7-ab9c-ff6ef8010778

[nrg101]

https://stashdb.org/edits/ded5caf5-7cac-4e56-ac6e-dccac9d90342

[feederbox826]

Agreed, closing PR. Will see if it's possible to add alt selector just for one site