Opened 10 years ago

Closed 10 years ago

#103 closed defect (fixed)

Atomspec specifying one model in hierarchy

Reported by: Tom Goddard Owned by: Conrad Huang
Priority: critical Milestone:
Component: Command Line Version:
Keywords: Cc: meng@…
Blocked By: Blocking:
Notify when closed: Platform: all
Project: chimera

Description

Atomspec parsing should be able to return the list of specified models not including parents. Currently with a command

color pink #1.2.3

The atomspec results models attribute returns models 1, 1.2 and 1.2.3. If this is a hierarchy of surfaces the color command would need to be able to get just model 1.2.3 without the parent models.

Change History (6)

comment:1 by Tom Goddard, 10 years ago

Cc: meng@… added

A related problem is that

open 1a0m
select #1.1

selects molecule #1 even when there is no submodel #1.1. And oddly

select #1.2

does not select model #1.

Elaine reported this:

From: Elaine Meng
Subject: possible "select" bug?
Date: October 7, 2015 at 1:27:02 PM PDT
To: Tom Goddard

Hi Tom,
I was checking to see whether “surface ligand” worked and happily it does, creating a separate surface model for each chain’s ligand:

open 4hhb
surf ligand
list

However, “select #1.1” then appears to select all of #1. Is that a bug? I thought #1 was the atoms and the submodels were the surfaces. To my surprise, however “select #1.2” only selects one surface.

Happy to submit a trac ticket if you want.
Thanks,
Elaine

comment:2 by Elaine Meng, 10 years ago

Priority: majorcritical

This problem makes it difficult to test whether some command works on a specific submodel (usually a molecular surface model) without its parent atomic model, because it makes it impossible to specify only that submodel.

comment:3 by Conrad Huang, 10 years ago

Resolution: fixed
Status: newclosed

Code was too smart for itself. If there are no submodels, the code was using 1 as the submodel number instead of returning "no match". Fixed in [develop 44fdb18].

comment:4 by Elaine Meng, 10 years ago

Resolution: fixed
Status: closedreopened

I'm afraid this is not fixed. For example,

open 4hhb
surf ligand

This gives atomic model #1 and surface models #1.1,#1.2,#1.3,#1.4

Although at first it seems to work to use "select #1.1" this approach does not select any of the other submodels (e.g. "select #1.2" does nothing). Then, when I try "select #1.1" again, it doesn't work either!

comment:5 by goddard@…, 10 years ago

This appears to be a different bug, the surface is selected but the green outline is not shown.  I’m looking at that.

comment:6 by Tom Goddard, 10 years ago

Resolution: fixed
Status: reopenedclosed

Conrad's fixed atomspec parsing worked correctly. It was another bug that caused selecting a molecular surface ("select #1.2") not to display. That is fixed too now.

Note: See TracTickets for help on using tickets.