Re: [Jmol-users] large animations

Bob Hanson Fri, 26 May 2006 11:00:02 -0700

[note: this message relates to the experimental prototype "Jmol 10.x" not
the public release Jmol 10.2. See
http://www.stolaf.edu/people/hansonr/jmol/test/proto/new.htm for details.]


eric capps wrote:

bob,

so if you have 1 model whose atoms you continually change with 1000
atoms you have i = 0 - 999? i may very well end up using this method
instead! let me know if there's any debugging/coding i can help with.
do you still consider these methods "highly experimental"?

oh, yeah. We are tapping into core classes directly via JavaScipt. Absolutely noguarantees. Be the first on your block....


i'm still trying to make the following work properly:

function loadNext(file){
    jmolSaveOrientation("save1");
    jmolScriptWait("load "+file);
    jmolRestoreOrientationDelayed("save1",0);
    makeNext();


var oldOrientation = jmolSaveOrientation("save1").replace(/1/,"0");
//changes this to "moveTo 0.0 ...." instead of "moveTo 1.0...."
jmolScriptWait("load "+file + ";" + oldOrientation);

}

you can ignore makeNext(), it just figures out which model to load
next. unfortunately there's still about a 1/10th of a second delay
between when the next file is loaded and when the orientation is
restored. and also, this approach unfortunately does not save any

information from label or render or anything else.


THAT's true.

i wonder, is it

possible to specify all of these parameters in a "load" statement?
failing that, is there a way to synchronize these commands tighter?
i'd be happy to work on this, but i'm still trying to figure out
exactly how jmol and jmol-new.js work together.


keep playing... and asking....


eric

On 5/25/06, Bob Hanson <[EMAIL PROTECTED]> wrote:

Eric,

yeah, that's it. Starting with 0 and NOT repeating for each model --
10 models of benzene would have i=0,1,2,....,i=119. Even if the data
are coming from different files, that number still just starts at 0
and runs.

We don't recommend people fiddling with that, but in this case, it's
an experiment. That's the only reliable tag you have.

It's also the number that's displayed if you use

select *;label %D

It's called the atom "index" number.

Bob

eric capps wrote:

> wow, yeah. so does move, save, move, restore (w/out the second save).
>
> regarding the jmolSetAtomCoord function, what does 'i' refer to? i'm
> guessing it's the atom number in the order that jmol read them?
>
> On 5/25/06, Bob Hanson <[EMAIL PROTECTED]> wrote:
>
>> Eric,
>>
>> I think there's something wrong with the data for the "home" position.
>> If you  first move the model around and THEN save, move some more,
>> save, they all work, right?
>>
>> Bob
>>
>>
>> eric capps wrote:
>> > bob,
>> >

>> > may be my environment but this isn't quite working for me. therestore>> > delayed function doesn't seem to do anything at all! does thiswork on

>> > your end?
>> >
>> > eric
>> >
>> > On 5/25/06, Bob Hanson <[EMAIL PROTECTED]> wrote:
>> >
>> >> what delay? Oh, THAT delay. Sure. see
>> >>
>> >> http://www.stolaf.edu/people/hansonr/jmol/test/proto/Jmol-new.js
>> >> http://www.stolaf.edu/people/hansonr/jmol/test/proto/new.htm
>> >>
>> >> jmolRestoreOrientationDelayed(id,delay,targetSuffix)
>> >>
>> >> -- id should correspond to same for jmolSaveOrientation()
>> >> -- delay defaults to 1 second
>> >> -- can be set to anything
>> >> -- targetSuffix optional
>> >>
>> >> Bob
>> >>
>> >> eric capps wrote:
>> >>
>> >> > bob,
>> >> >

>> >> > can you think of a way to do a restore orientation w/out thedelay?

>> >> > since i'm trying to use it in place of the built in animation
>> >> > capability, i want it to be able to snap back in position
>> immediately.
>> >> >
>> >> > On 5/23/06, Bob Hanson <[EMAIL PROTECTED]> wrote:
>> >> >
>> >> >>
>> >> >>
>> >> >> Eric Capps wrote:
>> >> >>
>> >> >> > bob,
>> >> >> >

>> >> >> > i believe i have the most recent version of jmol, but myjmol.js

>> >> >> > doesn't have these functions anywhere in it. the jmol-new.js
>> on the

>> >> >> > page you linked me to seems to be older, but it does havethese

>> >> >> > functions. what's the deal with that? does my version not have
>> that
>> >> >> > functionality?
>> >> >>
>> >> >> oops, sorry, it isn't older -- it's just that I've ignored the
>> header.
>> >> >> the two key files are JmolAppletProto.jar and Jmol-new.js in
>> >> >>
>> >> >> http://www.stolaf.edu/people/hansonr/jmol/test/proto
>> >> >>
>> >> >> if you use this, comment out your jmolInitialize() call and have
>> >> >> JmolAppletProto.jar in the same directory as your HTML page.
>> >> >>
>> >> >> >
>> >> >> > what you are describing sounds very interesting but is
>> >> unfortunately a
>> >> >> > bit over my head, as i'm new to both jmol and javascript.
>> still, i'd
>> >> >> > like to give it a shot.
>> >> >> >
>> >> >> > so am i right in thinking that you are suggesting writing a
>> method
>> >> >> > which would perform animation by moving an atom or a group of
>> atoms

>> >> >> > rather than reading an entire list of coordinates? thiswould be

>> >> >> > particularly helpful as the animations i am trying to make
>> typically
>> >> >> > only have one or two atoms moving at once (out of thousands).
>> would
>> >> >> > this be written in the jmol source code or the jmol.js file?
>> >> >>
>> >> >> Ah, there you go. Then you are all set! Can you figure out which
>> atoms
>> >> >> are
>> >> >> changing? That's interesting. OK, just for you, Eric:
>> >> >>
>> >> >> http://www.stolaf.edu/people/hansonr/jmol/test/proto/new.htm
>> >> >>
>> >> >> Fun, eh? This should give you some ideas.
>> >> >>

>> >> >> It will probably not work on Macs. What are your parametersin that

>> >> >> regard?
>> >> >>
>> >> >> If you want to change a WHOLE LOT of atom positions, though, we
>> should
>> >> >> pass the
>> >> >> applet a string and let it parse the string for atoms.
>> >> >>
>> >> >>
>> >> >> Bob
>> >> >>
>> >> >>
>> >> >> >
>> >> >> > -eric
>> >> >> >
>> >> >> > On 5/23/06, Bob Hanson <[EMAIL PROTECTED]> wrote:
>> >> >> >
>> >> >> >> what a challenge!
>> >> >> >>
>> >> >> >> Keeping the orientation after a file load is no problem. Jmol
>> 10.x
>> >> >> >> allows the
>> >> >> >> following:
>> >> >> >>
>> >> >> >> jmolSaveOrientation()
>> >> >> >> jmolScriptWait("load....")
>> >> >> >> jmolRestoreOrientation()
>> >> >> >>
>> >> >> >> So then the orientation resets to the original. You can see
>> this in
>> >> >> >> action at

>> >> >> >> http://fusion.stolaf.edu/chemistry/jmol/xtalx . There arelinks

>> >> in the
>> >> >> >> lower
>> >> >> >> left panel under "orientation" for "save" and "restore".
>> These work
>> >> >> >> across file
>> >> >> >> loads.
>> >> >> >>

>> >> >> >> 10.x also allows proper loading of multiple files. Theproblem

>> >> you are
>> >> >> >> going to

>> >> >> >> run into is that every atom has to be saved individually,so you

>> >> are
>> >> >> >> looking at
>> >> >> >> 100000 x (nAtoms) of atoms in memory. Not feasible.
>> >> >> >>

>> >> >> >> The better solution, it seems to me, would be to have alimited

>> >> number
>> >> >> >> of active

>> >> >> >> frames (say, 10) and then actively CHANGE the atompositions as

>> >> >> >> needed. Think of
>> >> >> >> the set of atoms as a "buffer" that we are going to fill as
>> needed.
>> >> >> >> When the
>> >> >> >> user is watching frame 3, we are loading frame 2; they go to
>> >> frame 4,
>> >> >> >> we load
>> >> >> >> frame 3. Sort of the way a word processor processes a large
>> >> document.
>> >> >> >> It doesn't
>> >> >> >> have to load the whole thing -- just what you are looking at.
>> One
>> >> >> >> would just
>> >> >> >> fill them as needed and display them in some sequence.
>> >> >> >>
>> >> >> >> So, it seems to me, the solution is to have the capability of
>> >> loading
>> >> >> >> atom
>> >> >> >> positions independently of the atoms themselves -- sort of a
>> >> >> "replace" or
>> >> >> >> "loadCoordinates" capability rather than a "load" capability,
>> >> and at
>> >> >> >> the frame

>> >> >> >> level rather than the "model set" level. Or, for thatmatter, at

>> >> the
>> >> >> >> atom level,
>> >> >> >> not the model level.
>> >> >> >>
>> >> >> >> It would only work for situations where the bonding and atoms
>> were
>> >> >> >> identical in

>> >> >> >> each "frame", but I think this is what you are thinking.Since

>> >> all you
>> >> >> >> need are
>> >> >> >> the xyz coordinates, the syntax is trivial -- I could imagine
>> >> >> >> accepting xyz

>> >> >> >> files, since they are so simple, but one could alsoimagine just

>> >> >> >> straight x y z
>> >> >> >> x y z x y z down the line. A very simple reader.
>> >> >> >>
>> >> >> >> Maybe the syntax would be something like:
>> >> >> >>
>> >> >> >> loadCoordinates FILENAME [optional model number]
>> >> >> >>
>> >> >> >> Personally, I would drive the whole thing with JavaScript
>> based on
>> >> >> >> that sequence

>> >> >> >> file, and as the frames change and send back theirframe-change

>> >> >> callback

>> >> >> >> message, execute the loads. In fact, if you set up theanimation

>> >> as a
>> >> >> >> loop 1 2 3
>> >> >> >> 4 5 1 2 3 4 5 1 2 3 4 5, then you get a message back that you
>> >> are on
>> >> >> >> frame X any

>> >> >> >> time. The JavaScript would then call to load frame X+3with file

>> >> Y+3,
>> >> >> >> and the
>> >> >> >> animation could just run and run, and the user could twiddle
>> >> with the
>> >> >> >> mouse all
>> >> >> >> he/she wants, and it would just be a gas. They could start,
>> stop,
>> >> >> >> rewind, etc.

>> >> >> >> You would just have to carefully sync the frame loadingwith the

>> >> >> >> user's trajectory.
>> >> >> >>
>> >> >> >> Certainly doable. Sound interesting?
>> >> >> >>
>> >> >> >> Bob Hanson
>> >> >> >>
>> >> >> >>
>> >> >> >>
>> >> >> >> Eric Capps wrote:
>> >> >> >> > i'm trying to create a very large animation. it maps the
>> >> movement of

>> >> >> >> > around 1000 atoms, over potentially hundreds ofthousands of

>> >> steps,
>> >> >> >> > with as many as 10000 unique states (there are situations
>> where a
>> >> >> >> > series of steps repeat themselves many times). i am new to
>> >> jmol, and

>> >> >> >> > from the demonstrations it seems that animation onlyworks on>> >> >> >> > multi-frame files. for hundreds of thousands of steps,this is

>> >> >> >> > unfeasible.
>> >> >> >> >
>> >> >> >> > i have each unique state stored in its own .xyz file, and
>> another
>> >> >> file

>> >> >> >> > which lists the order in which they occur. i thought tomake a>> >> >> >> > multi-frame .xyz of only the unique states and specifythem by

>> >> frame
>> >> >> >> > number when they repeat, but this still creates a very
>> large file
>> >> >> >> > (around 100 megs), far larger than i feel it has to be.
>> >> >> >> >
>> >> >> >> > now, i am trying to create a script which uses load
>> statements as
>> >> >> >> > necessary to load the next, potentially repeated .xyz file,
>> but i
>> >> >> want

>> >> >> >> > the user to be able to make changes to the view and havethem

>> >> not be
>> >> >> >> > reset every time a new .xyz is loaded. is this possible? if
>> not,
>> >> >> does
>> >> >> >> > anyone have any ideas on how to implement very large
>> animations?
>> >> >> >> >
>> >> >> >> > thanks,
>> >> >> >> > eric
>> >> >> >> >
>> >> >> >> >
>> >> >> >> > -------------------------------------------------------
>> >> >> >> > All the advantages of Linux Managed Hosting--Without the
>> Cost and
>> >> >> Risk!
>> >> >> >> > Fully trained technicians. The highest number of Red Hat
>> >> >> >> certifications in

>> >> >> >> > the hosting industry. Fanatical Support. Click to learnmore

>> >> >> >> >
>> >> >> >>
>> >> >>
>> >>