Features not Working properly

[learner-codec]

Hi I am trying to make use of slicer_cli_webs xml configuration to design a a front end for models parameters, but during my initial testing some of the features are not working as expected, such as whel selecting multi points, there are some numbers like -1 and -2 are being inserted randomly, is this behavior expected? here is an example: in small docker Example 1, First I have chosen multi points and selected two points, here is how the selection looks inside the text box. -1,-1,-2,-1,-2,-2,-1,-2,-1,-1,10539,5363,-1,-1,12711,6596,

Also while selecting points and putting the cursor on the image, the point location is being recorded but no marker is being shown on the screen. Could you please provide some insights, it will be really helpful.
Thank you

[learner-codec]

Hi, Just an addition, The second part of not showing the marker has been solved by cloning the latest git repo.

[manthey]

In the actual slicer-cli specification, the only supported region type is a three dimensional box. We have extended it by using a different arrangement of the vector array element (since the schema allows for this). See the region section here https://github.com/girder/slicer_cli_web#docker-clis for how this is done. For use in the HistomicsUI library, you must be on a new enough version for the interface to fully support this.

[learner-codec]

Thanks for your reply, I have seen the region section. I think I can get around that by having multiple panels for different shapes. I also have another question regarding it, Is it possible to select a shape and then use it indefinitely until I hit cancel or select a different shape, currently every time I want to use a shape, lets say "point" , after selecting once I have to go back to the shape selection panel and re-select it every time.

[manthey]

If you have the parameters set so the Submit option appears, when that is selected then selecting the shape will remain active (and each time you select a shape it will submit). Otherwise, you have to click on the region control before drawing a shape each time -- we certainly could add a feature to do that. Can you give a sample of your workflow so I can better appreciate why this is useful? It might help inform how we implement a toggle for this behavior.

[learner-codec]

Sure sir, For example I might want to select a specific region in a WSI and do some detection in it, a good example would be to use segment anything model, Instead of having to read the WSI over and over again for We could preload the region we have selected or even the whole WSI using large image or some other library and do detection on it. Also It would be really useful to have some functionality that clears all the selected point and regions after submission or a button that clears all the region information. Also if you could provide any useful resource that we can use to design our own features that can be added to SCW, It would be very helpful. Thank you.