- Chimera-dev - RBVI Mailing Lists

Re: [chimera-dev] [Chimera-users] Dialog w/ sliders?
by Eric Pettersen 21 Apr '08

21 Apr '08

On Apr 21, 2008, at 12:42 PM, Randy Heiland wrote: > I was wondering if there's a dialog in the gui code that has built-in > sliders? E.g., currently our extension module subclasses > ModelessDialog; however, our input panel has so many values that it > would be convenient if it was embedded in a widget such that sliders > appeared when it was resized smaller. Hi Randy, Not directly. One possibility is to use Pmw.ScrolledFrame (http:// pmw.sourceforge.net/doc/ScrolledFrame.html) as the parent widget for the contents of your dialog. Take the frame you get as an arg to the fillInUI() method and immediately create a Pmw.ScrolledFrame with that frame as the parent. Then fill your dialog in from the frame you get from Pmw.ScrolledFrame.interior(). You can do this almost painlessly like so: def fillInUI(self, parent): sf = Pmw.ScrolledFrame(parent) sf.grid(sticky="nsew") origParent, parent = parent, sf.interior() # remainder of fillInUI() remains the same... Another possibility is to use the Feature_Buttons_Panel widget from Chimera's CGLtk.Hybrid module. This is what Volume Viewer uses to hide/show sections of its interface. You'd have to look at the Volume Viewer code to see how it's called since I haven't used it myself. --Eric Eric Pettersen UCSF Computer Graphics Lab http://www.cgl.ucsf.edu

1 0

Chimera in windows: seeing samba mounted disk
by Jean-Didier Maréchal 18 Apr '08

18 Apr '08

Dear all, I usual use chimera on my linux box, but for some practicities, I need to use it in windows. I realized though that I can not access the disk F: of my windows box which is a mounted disk from the linux one. Would it be difficult that the file browser of chimera recognize samba mounted disk? Our is ther a trick I miss? All the best, JD On Wed, 2008-04-09 at 12:00 -0700, chimera-dev-request(a)cgl.ucsf.edu wrote: > Send Chimera-dev mailing list submissions to > chimera-dev(a)cgl.ucsf.edu > > To subscribe or unsubscribe via the World Wide Web, visit > http://www.cgl.ucsf.edu/mailman/listinfo/chimera-dev > or, via email, send a message with subject or body 'help' to > chimera-dev-request(a)cgl.ucsf.edu > > You can reach the person managing the list at > chimera-dev-owner(a)cgl.ucsf.edu > > When replying, please edit your Subject line so it is more specific > than "Re: Contents of Chimera-dev digest..." > > > Today's Topics: > > 1. Re: Multi-touch and Chimera (Tom Goddard) > 2. Re: Multi-touch and Chimera (Tom Goddard) > > > ---------------------------------------------------------------------- > > Message: 1 > Date: Tue, 08 Apr 2008 12:39:58 -0700 > From: Tom Goddard <goddard(a)cgl.ucsf.edu> > Subject: Re: [chimera-dev] Multi-touch and Chimera > To: Gergely Nagy <gergelynagy76(a)gmail.com> > Cc: chimera-dev(a)cgl.ucsf.edu > Message-ID: <47FBCA0E.8090702(a)cgl.ucsf.edu> > Content-Type: text/plain; charset=ISO-8859-1; format=flowed > > Hi Greg, > > Chimera is written in Python and C++. The C++ is used for OpenGL > rendering and for optimizing calculations. The event loop and user > interface is all in Python. The touch events and the actions they > trigger should definitely be in Python. It looks very reasonable to use > Boost Python to wrap the C++ touchlib library or parts of it to make it > accessible to Python. > > http://wiki.python.org/moin/boost.python/GettingStarted > > http://www.boost.org/doc/libs/1_35_0/libs/python/doc/index.html > > We use an in-house wrapping program (WrapPy). Wrapping C/C++ for Python > is often done with package called SWIG. > > We use the Tcl/Tk event loop using the Python/Tk interface module > called Tkinter. You can put custom Tk event types such as touch events > in the event queue but I haven't figured out a way to do it. The > trouble is that you typically will get an event from a callback in C/C++ > code. I recently tried passing such an event (for a Space Navigator > 6-dof input device) to a Python routine. It resulted in crashes because > Tkinter is designed so that only it can call Python code. It has some > complicated thread locking code that releases a Python lock and my event > handler called into Python without having that Python lock, tried to do > some Tk call and crashed. Probably the Python lock can be acquired in > the C/C++ callback, but then if that Python code does Tk calls to add an > event there may be additional problems with Tkinter's Tk lock which is > inaccessible. So some work is needed to figure out how to get the C/C++ > callback events into the Tk event loop. Once that is solved everything > can be done in Python with event handlers processing those Tk events to > cause any desired Chimera actions. > > Tom > > > Gergely Nagy wrote: > > Hi Tom, > > > > Thanks for your reply. > > Our multi-touch interface won't work with touch-pads, since it is > > based on another technology. > > The user will need to touch the actual object on the screen since > > there won't be any pointers. The interaction is direct between the > > user finger and the programs objects. So basically when the user > > touches a menu item for example, there is a sort of mouse clicked > > event at the given coordinates on the screen. > > Talking about events, there exists a library written in C++ called > > Touchlib (http://nuigroup.com/touchlib/) that pass those touch events > > to any program written in C++ and this is what we would like to use. > > Of course in this case we will need to wrap that code to be able to > > use it in python. I thought using Boost > > (http://www.boost.org/users/index.html) for that. > > I was asking you about the sources because at the end we should have > > something compiled and working. It doesn't matter if its not the > > latest version, what would be important is that we will be able to > > demonstrate the interface working. Would it be possible? > > > > Concerning the atom-atom distance, the user would zoom in on the part > > he is interested in, after he would choose the appropriate display > > form and then he would change from moving-zooming mode to selection > > mode and he would touch the atoms he would like to know the distance > > between. So you are right, visual feedback would be necessary, as it > > was the case with a mouse click. > > So to sum up the interaction events, there would be something like > > finger_down, finger_up, dragging. And also these events are dispatched > > by touchlib, so we would need to implement them in Chimera. > > > > What do you think about it? > > > > Thanks a lot for your help! > > > > Regards, > > Greg > > > > > > > > 2008/4/8, Tom Goddard <goddard(a)cgl.ucsf.edu > > <mailto:goddard@cgl.ucsf.edu>>: > > > > Hi Greg, > > > > Sounds neat. Will a multi-touch interface work with a standard > > laptop touch-pad? > > > > The source code for Chimera is on the web though it isn't up to date. > > > > http://www.cgl.ucsf.edu/chimera/sourcecode.html > > > > All of the functions you are interested in are implemented in > > Python code which is included in the distributed Chimera. The > > Python code is in files (with ".py" suffix) in directories under > > > > chimera/share > > > > or on Mac under directory Chimera.app/Contents/Resources/share. > > But it is quite difficult to find the Chimera Python routines you > > need. Our programming reference guide (automatically generated > > from code comments) is not too useful but we provide programming > > examples that might help. > > > > > > http://www.cgl.ucsf.edu/chimera/docs/ProgrammersGuide/Examples/index.html > > > > The easiest approach is to ask us by email for the routines you > > need, e.g. list all chain identifiers, select chain given id, > > color equivalent to actions menu, move models, rotate models, show > > atom-atom distance, find atom under mouse, draw label.... Eric > > Pettersen is the most familiar with that code and what simple > > routines already exist to do those things. He is on vacation now > > but expected back around Thursday. > > > > You are interested in atom-atom distances. Will the user choose > > an atom through the touch interface? It seems that would require > > visual feedback in the Chimera window -- equivalent of the mouse > > pointer. I suppose you could take over the mouse pointer, or you > > could create a new "cursor" implemented perhaps as a Chimera 2d label. > > > > Tom > > > > > > > > > > > > > ------------------------------ > > Message: 2 > Date: Tue, 08 Apr 2008 12:49:01 -0700 > From: Tom Goddard <goddard(a)cgl.ucsf.edu> > Subject: Re: [chimera-dev] Multi-touch and Chimera > To: Gergely Nagy <gergelynagy76(a)gmail.com> > Cc: chimera-dev(a)cgl.ucsf.edu > Message-ID: <47FBCC2D.2010402(a)cgl.ucsf.edu> > Content-Type: text/plain; charset=ISO-8859-1; format=flowed > > Greg suggest a direct call from C/C++ to queue an event. > > Tom > > Greg Couch wrote: > > Tom, > > > > You need to link with tcl (actually tcl stubs) and use the > > Tcl_ThreadQueueEvent function. > > > > - Greg > > > > On Tue, 8 Apr 2008, Tom Goddard wrote: > > > >> From: Tom Goddard <goddard(a)cgl.ucsf.edu> > >> Sender: chimera-dev-bounces(a)cgl.ucsf.edu > >> To: Gergely Nagy <gergelynagy76(a)gmail.com> > >> Cc: chimera-dev(a)cgl.ucsf.edu > >> Date: Tue, 08 Apr 2008 12:39:58 -0700 > >> Subject: Re: [chimera-dev] Multi-touch and Chimera > >> Received-SPF: pass (cgl.ucsf.edu: 169.230.27.3 is authenticated by a > >> trusted > >> mechanism) > >> Received-SPF: pass (cgl.ucsf.edu: 169.230.27.3 is authenticated by a > >> trusted > >> mechanism) > >> > >> Hi Greg, > >> > >> Chimera is written in Python and C++. The C++ is used for OpenGL > >> rendering and for optimizing calculations. The event loop and user > >> interface is all in Python. The touch events and the actions they > >> trigger should definitely be in Python. It looks very reasonable to use > >> Boost Python to wrap the C++ touchlib library or parts of it to make it > >> accessible to Python. > >> > >> http://wiki.python.org/moin/boost.python/GettingStarted > >> > >> http://www.boost.org/doc/libs/1_35_0/libs/python/doc/index.html > >> > >> We use an in-house wrapping program (WrapPy). Wrapping C/C++ for Python > >> is often done with package called SWIG. > >> > >> We use the Tcl/Tk event loop using the Python/Tk interface module > >> called Tkinter. You can put custom Tk event types such as touch events > >> in the event queue but I haven't figured out a way to do it. The > >> trouble is that you typically will get an event from a callback in C/C++ > >> code. I recently tried passing such an event (for a Space Navigator > >> 6-dof input device) to a Python routine. It resulted in crashes because > >> Tkinter is designed so that only it can call Python code. It has some > >> complicated thread locking code that releases a Python lock and my event > >> handler called into Python without having that Python lock, tried to do > >> some Tk call and crashed. Probably the Python lock can be acquired in > >> the C/C++ callback, but then if that Python code does Tk calls to add an > >> event there may be additional problems with Tkinter's Tk lock which is > >> inaccessible. So some work is needed to figure out how to get the C/C++ > >> callback events into the Tk event loop. Once that is solved everything > >> can be done in Python with event handlers processing those Tk events to > >> cause any desired Chimera actions. > >> > >> Tom > > > > ------------------------------ > > _______________________________________________ > Chimera-dev mailing list > Chimera-dev(a)cgl.ucsf.edu > http://www.cgl.ucsf.edu/mailman/listinfo/chimera-dev > > > End of Chimera-dev Digest, Vol 51, Issue 3 > ******************************************

2 1

Re: [chimera-dev] Multi-touch and Chimera
by Tom Goddard 08 Apr '08

08 Apr '08

Hi Greg, Chimera is written in Python and C++. The C++ is used for OpenGL rendering and for optimizing calculations. The event loop and user interface is all in Python. The touch events and the actions they trigger should definitely be in Python. It looks very reasonable to use Boost Python to wrap the C++ touchlib library or parts of it to make it accessible to Python. http://wiki.python.org/moin/boost.python/GettingStarted http://www.boost.org/doc/libs/1_35_0/libs/python/doc/index.html We use an in-house wrapping program (WrapPy). Wrapping C/C++ for Python is often done with package called SWIG. We use the Tcl/Tk event loop using the Python/Tk interface module called Tkinter. You can put custom Tk event types such as touch events in the event queue but I haven't figured out a way to do it. The trouble is that you typically will get an event from a callback in C/C++ code. I recently tried passing such an event (for a Space Navigator 6-dof input device) to a Python routine. It resulted in crashes because Tkinter is designed so that only it can call Python code. It has some complicated thread locking code that releases a Python lock and my event handler called into Python without having that Python lock, tried to do some Tk call and crashed. Probably the Python lock can be acquired in the C/C++ callback, but then if that Python code does Tk calls to add an event there may be additional problems with Tkinter's Tk lock which is inaccessible. So some work is needed to figure out how to get the C/C++ callback events into the Tk event loop. Once that is solved everything can be done in Python with event handlers processing those Tk events to cause any desired Chimera actions. Tom Gergely Nagy wrote: > Hi Tom, > > Thanks for your reply. > Our multi-touch interface won't work with touch-pads, since it is > based on another technology. > The user will need to touch the actual object on the screen since > there won't be any pointers. The interaction is direct between the > user finger and the programs objects. So basically when the user > touches a menu item for example, there is a sort of mouse clicked > event at the given coordinates on the screen. > Talking about events, there exists a library written in C++ called > Touchlib (http://nuigroup.com/touchlib/) that pass those touch events > to any program written in C++ and this is what we would like to use. > Of course in this case we will need to wrap that code to be able to > use it in python. I thought using Boost > (http://www.boost.org/users/index.html) for that. > I was asking you about the sources because at the end we should have > something compiled and working. It doesn't matter if its not the > latest version, what would be important is that we will be able to > demonstrate the interface working. Would it be possible? > > Concerning the atom-atom distance, the user would zoom in on the part > he is interested in, after he would choose the appropriate display > form and then he would change from moving-zooming mode to selection > mode and he would touch the atoms he would like to know the distance > between. So you are right, visual feedback would be necessary, as it > was the case with a mouse click. > So to sum up the interaction events, there would be something like > finger_down, finger_up, dragging. And also these events are dispatched > by touchlib, so we would need to implement them in Chimera. > > What do you think about it? > > Thanks a lot for your help! > > Regards, > Greg > > > > 2008/4/8, Tom Goddard <goddard(a)cgl.ucsf.edu > <mailto:goddard@cgl.ucsf.edu>>: > > Hi Greg, > > Sounds neat. Will a multi-touch interface work with a standard > laptop touch-pad? > > The source code for Chimera is on the web though it isn't up to date. > > http://www.cgl.ucsf.edu/chimera/sourcecode.html > > All of the functions you are interested in are implemented in > Python code which is included in the distributed Chimera. The > Python code is in files (with ".py" suffix) in directories under > > chimera/share > > or on Mac under directory Chimera.app/Contents/Resources/share. > But it is quite difficult to find the Chimera Python routines you > need. Our programming reference guide (automatically generated > from code comments) is not too useful but we provide programming > examples that might help. > > > http://www.cgl.ucsf.edu/chimera/docs/ProgrammersGuide/Examples/index.html > > The easiest approach is to ask us by email for the routines you > need, e.g. list all chain identifiers, select chain given id, > color equivalent to actions menu, move models, rotate models, show > atom-atom distance, find atom under mouse, draw label.... Eric > Pettersen is the most familiar with that code and what simple > routines already exist to do those things. He is on vacation now > but expected back around Thursday. > > You are interested in atom-atom distances. Will the user choose > an atom through the touch interface? It seems that would require > visual feedback in the Chimera window -- equivalent of the mouse > pointer. I suppose you could take over the mouse pointer, or you > could create a new "cursor" implemented perhaps as a Chimera 2d label. > > Tom > > > >

1 1

Multi-touch and Chimera
by Gergely Nagy 08 Apr '08

08 Apr '08

Hello all, I am a graduate student at the University of Applied Sciences in Switzerland and in my diploma work, I would like to develop a multi-touch user interface for Chimera. By multi-touch I mean the capability of interpreting simultaneous user interactions. You can find more info on the technology here: http://cs.nyu.edu/~jhan/ftirtouch/ <http://cs.nyu.edu/%7Ejhan/ftirtouch/> I already have a working hardware, now I have to find the software to work on. Of course the time and resource limitations of this diploma work wouldn't allow to interface all the functions of Chimera for multi-touch, so I'd be aiming to choose a couple of them, those that are used mainly by the users of Chimera that I know. So basically my aim would be to take for example the Select-chain menu, the Actions-color menu and the Tools-Structure Analysis-Distances tool and make it available to touch interaction. And of course the user would be able to move around the molecule with one finger along the X and Y axis, he would be able to turn it around using 3 fingers simultaneously (2 defining the axis and the third making the rotation) and zoom in and out using 2 fingers simultaneously. So my question is, would it be possible to have the source code that would let me do this development? Thanks a lot for taking the time to read my inquiry. Regards, Greg (Gergely Nagy)

2 1

Re: [chimera-dev] [Chimera-users] list of atoms attributes
by Eric Pettersen 01 Apr '08

01 Apr '08

On Feb 22, 2008, at 7:28 AM, Jean-Didier Maréchal wrote: > Hi again, > > When using the python shell, I am often struggling to find the > attribute for the atoms I work on. > for at my list with the selected atoms, I caught that at[i].name gives > the atom name of the entry at[i], at[i].residue provides with the name > residue.chain, etc. > > Where can I find in the programmer's guide the full list of available > attributes? What would be the way to ask for the serialNumber from the > IDLE? > > Thanks lot > JD Hi JD, dir(a) will list the attributes of 'a'. help(a) will produce documentation for the attributes of 'a' that come from the C++ layer (but not ones that were added in the Python layer). 'a' could be a class rather than an instance. For example, help(chimera.Atom) lists the methods/attributes of Atom and provides some info about them. We do what we can to keep the programmer's guide up to date, but for attributes (since they get added/changed pretty frequently) it is better to query with help()/dir() than to fall back on the relatively static programmer's guide documentation. --Eric Eric Pettersen UCSF Computer Graphics Lab http://www.cgl.ucsf.edu

3 2

Wrapping Chimera in a Java program
by Fabrice Jossinet 31 Mar '08

31 Mar '08

Dear developers, I would like to launch Chimera from my Java program. I would like to send messages to Chimera to highlight in its own display selections I do in my Java program's display. I already did that with PyMOL using a pipe communication. So, several questions: - is it possible to start Chimera from a classical python distribution (not the embedded python of Chimera)? One of my ideas is to start Chimera from Jython, a Java implementation of Python - is it possible to communicate with Chimera using a pipe ? - and the most general question: is it possible to communicate with Chimera from the outside? Thank you for your answers Fabrice -- Dr. Fabrice Jossinet Laboratoire de Bioinformatique, modelisation et simulation des acides nucleiques Universite Louis Pasteur Institut de biologie moleculaire et cellulaire du CNRS UPR9002, Architecture et Reactivite de l'ARN 15 rue Rene Descartes F-67084 Strasbourg Cedex France Tel + 33 (0) 3 88 417053 FAX + 33 (0) 3 88 60 22 18 f.jossinet(a)ibmc.u-strasbg.fr fjossinet(a)gmail.com http://www-ibmc.u-strasbg.fr/arn/Westhof/index.html http://fjossinet.u-strasbg.fr/

2 1

new keyword arg for added commands
by Eric Pettersen 26 Mar '08

26 Mar '08

If you have code that adds commands to Chimera using the addCommand () function, there is a new keyword to that function so that animation scripts can work better. The keyword is 'changesDisplay' and is a boolean that defaults to True. If your command typically changes the display in the main graphics window somehow the keyword should be True, otherwise False. It's needed so that as a script is executed, Chimera knows after each line whether a new frame is necessary or not. For instance, for the 'alias' command the value is False since after executing that command in a script you would not want to record a new frame in an animation. --Eric Eric Pettersen UCSF Computer Graphics Lab http://www.cgl.ucsf.edu

1 0

fit map in map -- also fit a/pix at the same time ?
by stefan.backstrom＠medchem.umu.se 26 Mar '08

26 Mar '08

Hi. Would it be possible to optimize the a/pix, or scale value when one is fitting either a map in map, or model in map ? Often it seems as if the scale is not 100 % correct between a crystal structure and a cryo-EM reconstructinon, and it would be nice to be able to choose to fit the scale of the EM-map at the same time as the fit rotation and translation. Best wishes, Stefan Bäckström Medical Biochemistry and Biophysics Umeå University Sweden

2 1

iPhone SDK
by Tom Ferrin 11 Mar '08

11 Mar '08

The iPhone SDK was announced last week. Turns out the iPhone includes a lot of the same OS layers as does the Mac. I didn't see anything about Python, but there is a hardware accelerated OpenGL ES in there. Also the standard set of development tools (interface builder, xcode, etc) and a package called "Instruments" that let's you capture on a Mac lots of execution trace information on the iPhone via an interface cable. One novel user interface component: the iPhone includes a 3D accelerometer with software interface. The video I watched showed a shoot-em-up space wars game where you steered by moving the phone around. Someone is bound to come up with an interactive molecular viewer; I suspect Warren DeLano is working on Pymol right now. --tom

2 1

Re: [chimera-dev] [Chimera-users] Driving matchAtoms from the IDLE
by Eric Pettersen 22 Feb '08

22 Feb '08

On Feb 22, 2008, at 3:45 AM, Jean-Didier Maréchal wrote: > Hi Guys, > > I'd like to superpose two inorganic complexes using two differently > "ordered" atom sets. My compound has quite similar substructures but I > am sure at all what is the best "ordered" selection of atom pairs to > align. Therefore, I'd like to make a "extensive" search (with an > heuristic approach to define...) passing by the idle. I am > unfortunately > missing something before getting to this point! > > Lets say that I just want to align the nitrogen atoms of the my two > structures. I have: > > import chimera > import numpy > from chimera import match > from chimera import runCommand > runCommand("select #1@N=") > an1=chimera.selection.currentAtoms() > runCommand("select #0@N=) > an2=chimera.selection.currentAtoms() > chimera.match.matchAtoms(an1,an2) > > And then it's when I am already stuck! > > the resulting prompt does not provide mistake an gives: > chimera.match.matchAtoms(an1,an2) > (<_chimera.Xform object at 0x9463ef8>, 0.034256915079603358) > > But how this results can be send to the graphical window? I guess I > have > to use the Xform object, but I can not clearly understand how to do > this. > > Thanks for any help! > > JD Hi JD, I'm not 100% certain what your goal is, but you could just run the match command directly: from chimera import runCommand runCommand("match #1@N= #0@N=") If you need access to the RMSD of the match, you could call the underlying match() function: from Midas import match mobileAtoms, refAtoms, rmsd = match("#0@N=", "#1@N=") which will also superimpose the structures. --Eric Eric Pettersen UCSF Computer Graphics Lab http://www.cgl.ucsf.edu

1 0