From: Axel Kohlmeyer (
Date: Wed Mar 11 2015 - 05:54:56 CDT

On Wed, Mar 11, 2015 at 11:39 AM, Albert Solernou
<> wrote:
> Hi Axel,
> thanks for the tip. Indeed I imagine a variation of your Quicksurf/Polyhedra
> representation idea. I could load the nodes (positions for the vertices) and
> faces (the way that the triangles emerge from nodes) at
> plugin.open_file_read,
> but then make the nodes correspond to atoms, being able to label in a lot of

the plugin api allows you to read properties of vertices like
type/label/radius (in read_structure), the position of vertices (in
the read_next_timestep) and connections (in read_bonds), no faces or
anything else. also, the connection information is read only once, so
it cannot change over the course of a trajectory.

the polyhedra representation, however, will not use that connection
information, but build connections based on a distance cutoff (it was
conceived to highlight substructures in simple crystals) and the
quicksurf representation will use the position and radius information
to build an density map and then compute an isosurface from it.

> information, and creating a new drawing method in order to plot the
> triangles. That does not sound that difficult, does it? Would that work?

adding to a new drawing style to VMD is a significant effort.

like i mentioned before, you are trying to teach VMD to do something
that it is pretty much against its fundamental design decisions, so it
is to be expected that there are significant problems at every step of
the way. it would certainly desirable for some specialized
applications (e.g. hybrid DEM and FEM calculations) to have these kind
of features available, but you also need to keep in mind that this
would be a niche application of VMD and any changes that would
negatively impact the performance of VMD for common use cases are not
likely to be integrated. the other visualization tools that i
mentioned (and there are likely more of those) were meant to handle
such irregular meshes as you have as data and can also be taught to
deal with atomic data (just not as efficiently as VMD for large
datasets and long trajectories and don't have the specialized
scripting for analysis of position/molecular data).



> Cheers,
> Albert
> On 03/10/2015 08:11 PM, Axel Kohlmeyer wrote:
>> On Tue, Mar 10, 2015 at 1:12 PM, Albert Solernou <>
>> wrote:
>>> Hi John,
>>> what do you think about Pawel's approach? Is it feasible at all? To me
>>> that
>>> would be the preferred solution. In the meantime I am spending my time
>>> reading on VisIt.
>> you can try using the polyhedra representation and experiment with the
>> distance cutoff setting. it won't work for all kinds of geometries,
>> but might be sufficient for simple ones with fairly regular meshes.
>> another option to turn pure position data into surface objects could
>> be the Quicksurf representation.
>> axel.
>>> Best,
>>> Albert
>>> On 03/10/2015 12:44 PM, Axel Kohlmeyer wrote:
>>>> On Tue, Mar 10, 2015 at 11:53 AM, Albert Solernou
>>>> <> wrote:
>>>>> Hi John,
>>>>> thanks for the answer. Although it is not the ideal solution, and
>>>>> although I
>>>>> don't know tcl... I could give it a go. Could you point me to the right
>>>>> documentation?
>>>>> In addition I have a few extra questions:
>>>>> - Will I be able to define/label different bodies or parts so that I
>>>>> can
>>>>> click on them, change colors, and show/hide?
>>>> nope. VMD is not really a good tool for such things. you might be
>>>> better off to look into a tool like ParaView or VisIt, for example.
>>>>> - In that case how will that integrate with the next timestep given
>>>>> that
>>>>> it
>>>>> is a "different" molecule?
>>>> each timestep has to be a "molecule" (i.e. in "molecule" in VMD is the
>>>> term for dataset), the rest is handled by the plugin GUI, which
>>>> emulates the regular animation controls and instead shows/hides
>>>> individual molecules/datasets
>>>>> - I have currently a Tkinter interface that I plan to port to, and I
>>>>> read
>>>>> that you support it. Could you point me to the right documentation site
>>>>> to
>>>>> do so? Do you have any suggestion?
>>>> Tkinter is a significant challenge with VMD, since python runtimes are
>>>> so incredible non-portable and so large that they are no longer
>>>> bundled with VMD. so to have access to tkinter, you first will need to
>>>> compile a custom version of VMD that includes python.
>>>> axel.
>>>>> Cheers,
>>>>> Albert
>>>>> On 03/09/2015 06:00 PM, John Stone wrote:
>>>>>> Hi,
>>>>>> At present, VMD doesn't provide for time-varying mesh data.
>>>>>> In the short term, you could overcome this limitation by loading
>>>>>> different trajectory timesteps into separate molecules. Once loaded
>>>>>> you can animate the mesh trajectory using the multimolanim plugin.
>>>>>> Let us know if you need help to get this working.
>>>>>> Cheers,
>>>>>> John Stone
>>>>>> On Mon, Mar 09, 2015 at 03:11:26PM +0000, Albert Solernou wrote:
>>>>>>> Dear all,
>>>>>>> I am writing a new plugin that should enable VMD to work with Finite
>>>>>>> Element Trajectories. Currently, my plugin reads correctly the input
>>>>>>> file, through:
>>>>>>> plugin.open_file_read
>>>>>>> then loads and displays the surface triangles through:
>>>>>>> plugin.read_rawgraphics
>>>>>>> and finally releases my handler:
>>>>>>> plugin.close_file_read
>>>>>>> The next step should be to load the whole trajectory, and this is
>>>>>>> where I need some advice. I understand that this should be done
>>>>>>> through:
>>>>>>> plugin.read_next_timestep
>>>>>>> but, will it update the vertices of the triangles? How can I do so
>>>>>>> using molfile_timestep_t? Is there an easy workaround?
>>>>>>> Thanks in advance,
>>>>>>> Albert
>>>>>>> --
>>>>>>> ---------------------------------
>>>>>>> Dr Albert Solernou
>>>>>>> EPSRC Research Fellow,
>>>>>>> Department of Physics and Astronomy,
>>>>>>> University of Leeds
>>>>> --
>>>>> ---------------------------------
>>>>> Dr Albert Solernou
>>>>> EPSRC Research Fellow,
>>>>> Department of Physics and Astronomy,
>>>>> University of Leeds
>>> --
>>> ---------------------------------
>>> Dr Albert Solernou
>>> EPSRC Research Fellow,
>>> Department of Physics and Astronomy,
>>> University of Leeds
> --
> ---------------------------------
> Dr Albert Solernou
> EPSRC Research Fellow,
> Department of Physics and Astronomy,
> University of Leeds

Dr. Axel Kohlmeyer
College of Science & Technology, Temple University, Philadelphia PA, USA
International Centre for Theoretical Physics, Trieste. Italy.