Opened 3 years ago

Closed 3 years ago

#7962 closed defect (fixed)

split submodels not hiding

Reported by: Elaine Meng Owned by: Tom Goddard
Priority: moderate Milestone:
Component: Structure Editing Version:
Keywords: Cc: pett
Blocked By: Blocking:
Notify when closed: Platform: all
Project: ChimeraX

Description

I am mystified as to why after using this splitting example

open 2gbp
split #1 atoms ligand atoms solvent atoms :1-10 atoms :11-100 atoms :101-end

...unchecking the "eye" icon in the Model Panel for the new ligand submodel does NOT hide the ligand, nor does that work for the new solvent submodel either. The show/hide commands generated by checking/unchecking are purely by model number, also not working:

hide #1.1 models
hide #1.2 models

These also do not work

hide #1.1 atoms
hide #1.2 atoms

However, this works

select #1.1

Tested in UCSF ChimeraX version: 1.6.dev202211030048 (2022-11-03)

Change History (8)

comment:1 by pett, 3 years ago

As Elaine has surmised, the ligand is both in the first 'atoms' (ligand) and the last one (:101-end). The question is whether that is a bug or not.

in reply to:  2 ; comment:2 by Elaine Meng, 3 years ago

Hmmm, the name "split" suggests mutual exclusivity but perhaps it is useful that it is not mutually exclusive?  Documentation is silent on that issue but I can amend it if we want this behavior.  (Also I guess my brain was thinking "end"=end of peptide chain but of course that is not true.)

comment:3 by Tom Goddard, 3 years ago

The split code does not insist that multiple "atoms" option be non-overlapping. It could be made to give an error if they overlap.

in reply to:  4 ; comment:4 by Elaine Meng, 3 years ago

The current behavior is OK with me.  Unless you decide it should  behave differently, I just need to add something about it to the documentation.

comment:5 by pett, 3 years ago

Maybe it should _warn_ if they're non-overlapping?

comment:6 by pett, 3 years ago

Oops, I mean overlapping, not non-overlapping.

in reply to:  7 ; comment:7 by Tom Goddard, 3 years ago

I want to think if I can imagine any case where overlapping atoms specifications would be useful.  If not, it clearly can lead to confusion and maybe it should be disallowed.  Warning that there is overlap is also an option.  

comment:8 by Tom Goddard, 3 years ago

Resolution: fixed
Status: assignedclosed

I couldn't think of a case where overlapping atom specifications seems useful and it can be very confusing when they overlap so I made it give an error when atoms overlap. In the example in the ticket it will say:

split #1 atoms ligand atoms solvent atoms :1-10 atoms :11-100 atoms :101-end

The split command requires non-overlapping atoms, "ligand" and ":101-end" have 12 atoms in common, "solvent" and ":101-end" have 214 atoms in common

Note: See TracTickets for help on using tickets.