RCM viewer component

[neuroelf]

%% note for the XBS team: I am aware that @EkaterinaPoddubskaya is currently enjoying some time-off; and I simply wanted to get this request into the system! :)

For images of type "RCM: mosaic," the "enlarge" button currently shows an error icon, most likely due to the fact that browsers do not natively support Cloud-optimized GEO-Tiff as an image format.

We would like XBS to implement, as a minimum viewer, code (similar or identical) available here, whereby instead of displaying the (JPG) image (for other image types), a new DIV element is displayed that is initialized with the correct ISIC ID to show the RCM mosaic.

[neuroelf]

Hello @EkaterinaPoddubskaya -- and welcome back! :) Would you mind letting us know if you have any questions on how to implement this minimal viewer for images with that specific type?

[EkaterinaPoddubskaya]

Hello @neuroelf , glad to be back! I've already started working on this issue and don't have any questions so far

[IharXB]

Hi @neuroelf , let me clarify one moment regarding the feature implementation. As user clicks "Enlarge" on any image, including tiff file, applies zoom in/out several times and than clicks "Metadata" switcher, the image is returned to its' initial zoom state. So zoom state is not saved. Having many levels of zoom this situation may not be comfortable for users. We will implement zoom state saving to improve UX.

[neuroelf]

Hello Ihar,

We will implement zoom state saving to improve UX.

Yes, please make sure that as a user brings up other aspects of this particular UI widget (the zoomable RCM mosaic viewer) that the viewport (zoom, position) remains the same.

[EkaterinaPoddubskaya]

Hi @danlamanna and @brianhelba,

I've noticed that the melanoma thickness filter does not work with the value "5.0-*". I tested on https://api.isic-archive.com, and while all other values work correctly, queries using "5.0-*" fail due to the unspecified upper bound.

Could you please clarify the correct syntax for specifying an open-ended range (from 5.0 to infinity) in this filter?

Thanks in advance!

[danlamanna]

This was a bug on our end, it should work now, thanks for catching that.

…

On Thu, Aug 7, 2025 at 5:35 AM EkaterinaPoddubskaya < ***@***.***> wrote: *EkaterinaPoddubskaya* left a comment (ImageMarkup/isic-gallery#123) <#123 (comment)> Hi @danlamanna <https://github.com/danlamanna> and @brianhelba <https://github.com/brianhelba>, I've noticed that the melanoma thickness filter does not work with the value "5.0-*". I tested on https://api.isic-archive.com <https://api.isic-archive.com>, and while all other values work correctly, queries using "5.0-*" fail due to the unspecified upper bound. Could you please clarify the correct syntax for specifying an open-ended range (from 5.0 to infinity) in this filter? Thanks in advance! image.png (view on web) <https://github.com/user-attachments/assets/94694e52-2451-4e95-9a8e-00e5a79d8175> image.png (view on web) <https://github.com/user-attachments/assets/2a45967a-499c-4a12-9f9a-a1fab07c46f4> — Reply to this email directly, view it on GitHub <#123 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AAEUPZOW6SNDM3WYNUJVWID3MMMXJAVCNFSM6AAAAAB75QRR2GVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZTCNRTGMYTIOBWHE> . You are receiving this because you were mentioned.Message ID: ***@***.***>

-- Dan LaManna Kitware, Inc.

[EkaterinaPoddubskaya]

This was a bug on our end, it should work now, thanks for catching that.
…

It now works on api.isic-archive.com, but it still doesn't work in the gallery at https://gallery.isic-archive.com/. Could you please take a look at this as well?

[danlamanna]

It looks like the gallery is sending the asterisk as undefined: https://api.isic-archive.com/api/v2/images/search/?limit=15&query=mel_thick_mm%3A%5B5%20TO%20undefined%7D&collections=

[EkaterinaPoddubskaya]

Yes, thats because this asterisk value doesnt come from the server (see the missing "to" value in the attached pic). Could you please add it to the response?

[danlamanna]

Got it. It should be set now.

[EkaterinaPoddubskaya]

Hello @neuroelf,

We've noticed a behavior related to zooming within the enlargement window that we wanted to bring to your attention.

In the enlargement window, there are zoom in (+) and zoom out (-) buttons located below the image. When a user zooms in very close to the edges of the picture and then presses the zoom out button, the image slightly shifts. You can see this behavior live here: https://add-tiff-support.isic-gallery.pages.dev/.

This appears to be the default zoom behavior in the library (you can also observe it in this snippet with the default library buttons: https://codepen.io/brianhelba/pen/XJrBdqp). Both our custom buttons and the default library buttons exhibit the same behavior.

Would you prefer us to look into implementing some custom behavior for the buttons to make the zoom smoother, or is this behavior something you consider expected?

Thank you!

[neuroelf]

Hello @EkaterinaPoddubskaya,

Thanks for pointing this out! After looking at the draft (add-tiff-support.) deploy sub-domain together with Nick, and playing with this, I find the behavior acceptable. It's admittedly mildly (visually) displeasing, but since this is an external library, we find this to be an acceptable trade-off.

So, please "ignore" this sort of visual "bug" in the library :)

[EkaterinaPoddubskaya]

Hi @neuroelf ,

Just wanted to let you know that before I got your reply, I looked into implementing a custom zoom behavior for the buttons and came up with an idea for a solution to this mildly displeasing behavour :) It didn’t take long to implement it, so I hope that’s okay!

I’ll push the changes to my branch soon, and you can check them out whenever you like.

[brianhelba]

@EkaterinaPoddubskaya @neuroelf It's definitely possible to configure OpenLayers to not shift the camera on zoom-out. It's currently moving the camera in an attempt to keep as much of the content on the viewport as possible (otherwise you could pan partially off the content extent).

I don't have a strong opinion which behavior the viewer uses:

• keep the content maximally on the viewport (the current default)
• keep the camera pan position maximally stable