Segmentations tool lists misleading ID numbers

[Elaine Meng]

When I create segmentations in the Segmentations tool it lists them in the table with names ending in (#1) (#2) etc. even though these bear no relation to the actual model numbers. The actual model numbers should be given instead, or no numbers at all. The numbers that are shown are misleading because they look like model numbers, and useless because you can't use them for anything (they don't work in commands etc.). Tested in UCSF ChimeraX version: 1.9.dev202405090425 (2024-05-09)

[Elaine Meng]

Session attached, created in same daily build. I guess these "(#N)" things are part of the name, so my original description actually applies to the model names that are generated by Segmentations. Further, it loses track so that multiple models can have the same N in (#N) in the name. Opening the session shows that the segmentation model names include (#1) for two segmentation models and (#2) for the third. Their actual model numbers are #2, #6, and #7.

[Zach Pearson]

I need to fix the logic clearly but the underlying intent was to have the segmentations be distinguishable from one another in the menu. What about "segmentation x of foo", where x has no hash that would make it appear to be a model number?

[Elaine Meng]

That's fine with me!  Or, if you decide a shorter name might be better, "seg 1 of foo" or "seg1 of foo" or something along those lines.

[Zach Pearson]

OK, I've pushed a ​fix. Currently if you, say, create three segmentations and delete all of them and create three new ones, you'll get new segmentations 1-3.

Also, if you create a segmentation, open one from a file, and create a new one, you'll see "segmentation 1", "(whatever the name of the DICOM was)", "segmentation 3" -- I'm just tracking the number of currently open segmentations and adding 1 for the next one.

Let me know if you prefer tracking the total number of segmentations ever created for a volume and incrementing from there even if previous numbers are available.

[Elaine Meng]

That sounds good to me, just starting over at 1 with the numbers if you deleted the previous segmentations for that image.