Re: [Jmol-users] color from external dataset FOLLOWUP

[Bob H. <ha...@st...>]

Good work and good summary, Tim. Sorry about "property_" not being clear.
The dual-column file format sounds excellent to me. What we will need is a
function name. But load() isn't it. load() just puts the file data into 
x. So the filter isn't there.
So it has to be a property of the data command. That's where the actual 
creation of the
parameters comes from. How about this?

set dataFormat n  # take data from nth column

set dataFormat -n # take data from nth column, reading atom values in 
column 1.

That might be pretty powerful. Very simple to implement.
 
I think you would still have to apply this as a filter to the selected 
set, because in a
multifile setting you don't want to have to worry about which file was 
loaded first.

I guess I don't see why careful selection shouldn't do the trick for the 
assignment of
values. What was it that was difficult about that?

Bob

Timothy Driscoll wrote:

>hi,
>
>thought I would followup this thread with some feedback, having tried  
>to implement various flavors of this recently.
>
>
>################
>1. coloring based on a built-in property works well and is very cool:
>
>y = {*}.occupancy.all
>data "property_charge @y"
>propertyColorScheme = "bwr"
>color atoms property_charge
>
>
>one caveat: the name of your new property must begin with  
>"property_".  this was not clear to me initially, and I spent a while  
>wondering why "my_charges" kept failing. :-)
>
>note: propertyColorScheme is also very useful and easy to implement.
>
>
>
>################
>2. coloring based on values in an external data file was cumbersome  
>when using any subset of atoms in a structure:
>
>y = load("charge.dat")
>data "property_charge @y"
>propertyColorScheme = "bwr"
>color atoms property_charge
>
>
>I could not easily align the values in my data file with the atom  
>count in the selected set in Jmol; the atom count was always  
>different, which threw off the coloring scheme. in the end, I put the  
>values in the occupancy column of the pdb file and used method 1,  
>which worked great.
>
>a suggestion: use a more explicit atomno:value syntax in the data  
>file, and have Jmol ignore values for atoms that are not in the  
>selected set.
>
>
>
>kudos for great work,
>
>tim
>  
>