January 2023 - ChimeraX-users - RBVI Mailing Lists

Re: [chimerax-users] Remote control / Threading in ChimeraX
by Albert Smith-Penzel 05 Jan '23

05 Jan '23

Hi Tom, Thanks for the explanation. I’ll have to think and investigate a bit to see if my thread use is causing some problems elsewhere. Most of our code has been pretty stable, but there are still a few occasional glitches that I haven’t completely understood. Anyway, session.ui.thread_safe is perfect for resolving the issues we have with adding the surface to the empty session and running commands remotely. The curl solution also works; I had something like this implemented previously for some other purposes, but my own preference is to send commands/data directly rather than writing out a file to be read on the other end. Thanks again! Cheers, Albert ––––––––––––––––––––––––––––––––––– Albert Smith-Penzel, Ph.D. DFG Project Leader Institute for Medical Physics and Biophysics Leipzig University 04107 Leipzig, Germany E-mail: albert.smith-penzel(a)uni-leipzig.de <mailto:albert.smith-penzel@medizin.uni-leipzig.de> > On 4. Jan 2023, at 21:41, Tom Goddard <goddard(a)sonic.net> wrote: > > WARNUNG: Diese E-Mail kam von außerhalb der Organisation. Klicken Sie nicht auf Links oder öffnen Sie keine Anhänge, es sei denn, Sie kennen den Absender und wissen, dass der Inhalt sicher ist. > > WARNUNG: Diese E-Mail kam von außerhalb der Organisation. Klicken Sie nicht auf Links oder öffnen Sie keine Anhänge, es sei denn, Sie kennen den Absender und wissen, dass der Inhalt sicher ist. > > Hi Albert, > > ChimeraX is not designed to run arbitrary Python code multithreaded. There are many reasons. First the Qt window toolkit is not thread safe -- all Qt GUI calls have to be done in the main thread. Violating that often results in a crash. But even without the GUI, imagine the main thread is looping over the list of open models while a second thread decides to add a model to that list -- this could break the looping thread many ways. Python has a global interpreter lock which means it only can run one thread at a time. Still if you pause one thread and then change the state of the ChimeraX session in another thread, the first thread could throw errors in countless ways when it resumes at an arbitrary point in the code with the data structures changed. In general very little software is thread safe because it is very difficult to implement. Usually that is only done by making threads not modify any data structures used by other threads. > > Here are two solutions. 1) Have your listener thread run the Python code it receives in the main thread rather than the listener thread. There is a ChimeraX function that takes a function and queues it to be run in the main thread. > > session.ui.thread_safe(func, args...) > > Here is the code if you want to see how it works > > https://github.com/RBVI/ChimeraX/blob/2e33222fabb48edec76df6e82a10efedb54a8… > > A second solution is to use the ChimeraX remotecontrol communication for everything. That mechanism just runs ChimeraX commands, but you can run Python code using it by just sending the command "open script27.py" where you put the Python you want to run in a file script27.py. > > Tom > > >> On Jan 4, 2023, at 4:40 AM, Albert Smith-Penzel via ChimeraX-users <chimerax-users(a)cgl.ucsf.edu> wrote: >> >> Hello, >> >> I am working on a Python project that uses ChimeraX (currently running 1.5 to interactively view dynamics data by mapping it onto a molecule. All the figures/movies in this paper, https://www.nature.com/articles/s41467-021-27417-y <https://www.nature.com/articles/s41467-021-27417-y>, were made possible by Chimera! Also thanks to Tom Goddard for explaining to me how some of those figures could be made awhile ago (https://rbvi.github.io/chimerax-recipes/spherical_harmonics/spherical_harmo… <https://rbvi.github.io/chimerax-recipes/spherical_harmonics/spherical_harmo…> ). >> >> So we’re controlling ChimeraX remotely; we launch a session and then can send commands from Python over to Chimera and edit images from within our python code (while still allowing us to interact via the ChimeraX interface). The basic way this works is as follows >> >> launch chimeraX from our code, including a short python script >> python script imports our libraries, launches “cmxr” (chimeraX remote) >> cmxr uses multiprocessing.connection.Listener to listen for commands/data coming from Client in our python code >> !!!The Listener runs in a thread (threading.Thread) so that we continue to be able to interact normally with ChimeraX (launching the Listener without using a thread freezes ChimeraX while waiting for commands, rendering it unusable) >> Then, we just tell it to run some function(s) that we’ve imported into ChimeraX, and provide some data, atom ids, etc. over the Client/Listener connection >> >> There are a few things that I can’t figure how to make work via this process. We’ve had a few long-standing workarounds, but I’ve always been a little disturbed that we can’t get this all to work from python alone. At the moment, when we launch ChimeraX, we also enable remote control via curl, and some commands are then sent that way (remotecontrol rest start port XXXX) >> >> In particular, if I want to send commands that would normally be executed at ChimeraX’s command line, in principle I should be able to use chimerax.core.commands.run. If I execute “run” from a simple python script, everything works, but if I run it from within a thread (as required by our Listener setup), then ChimeraX freezes. So, we send these commands via curl on the remotecontrol instead. >> >> Second, is there a right way to open models/surfaces from the ChimeraX python shell? At the moment, we also open pdbs via curl commands. For pdbs, maybe I just don’t know the right python command. However, I am re-implementing spherical harmonics referenced above. Once I have prepared the surface (“s”), if I add it to session.models, it crashes chimeraX if it is being added within the listener thread, and if there are no models already open. If another model is already open, then it works just fine, and if I load the same modules/execute the same commands from within the python shell provided in ChimeraX, it also works (regardless of other models being open). It seems like maybe ChimeraX isn’t able to fully initialize the session from within a thread. >> >> I’ve grouped these problems together because I guess mainly they have to do with being unable to do some actions from within threads, and if there’s an explanation/workaround for dealing with that. >> >> Thanks for your help! >> >> Cheers, >> Albert >> >> >> >> ––––––––––––––––––––––––––––––––––– >> Albert Smith-Penzel, Ph.D. >> DFG Project Leader >> Institute for Medical Physics and Biophysics >> Leipzig University >> 04107 Leipzig, Germany >> E-mail: albert.smith-penzel(a)uni-leipzig.de <mailto:albert.smith-penzel@medizin.uni-leipzig.de> >> >> >> >> >> >> _______________________________________________ >> ChimeraX-users mailing list >> ChimeraX-users(a)cgl.ucsf.edu >> Manage subscription: >> https://www.rbvi.ucsf.edu/mailman/listinfo/chimerax-users >

1 0

Remote control / Threading in ChimeraX
by Albert Smith-Penzel 04 Jan '23

04 Jan '23

Hello, I am working on a Python project that uses ChimeraX (currently running 1.5 to interactively view dynamics data by mapping it onto a molecule. All the figures/movies in this paper, https://www.nature.com/articles/s41467-021-27417-y <https://www.nature.com/articles/s41467-021-27417-y>, were made possible by Chimera! Also thanks to Tom Goddard for explaining to me how some of those figures could be made awhile ago (https://rbvi.github.io/chimerax-recipes/spherical_harmonics/spherical_harmo… <https://rbvi.github.io/chimerax-recipes/spherical_harmonics/spherical_harmo…> ). So we’re controlling ChimeraX remotely; we launch a session and then can send commands from Python over to Chimera and edit images from within our python code (while still allowing us to interact via the ChimeraX interface). The basic way this works is as follows launch chimeraX from our code, including a short python script python script imports our libraries, launches “cmxr” (chimeraX remote) cmxr uses multiprocessing.connection.Listener to listen for commands/data coming from Client in our python code !!!The Listener runs in a thread (threading.Thread) so that we continue to be able to interact normally with ChimeraX (launching the Listener without using a thread freezes ChimeraX while waiting for commands, rendering it unusable) Then, we just tell it to run some function(s) that we’ve imported into ChimeraX, and provide some data, atom ids, etc. over the Client/Listener connection There are a few things that I can’t figure how to make work via this process. We’ve had a few long-standing workarounds, but I’ve always been a little disturbed that we can’t get this all to work from python alone. At the moment, when we launch ChimeraX, we also enable remote control via curl, and some commands are then sent that way (remotecontrol rest start port XXXX) In particular, if I want to send commands that would normally be executed at ChimeraX’s command line, in principle I should be able to use chimerax.core.commands.run. If I execute “run” from a simple python script, everything works, but if I run it from within a thread (as required by our Listener setup), then ChimeraX freezes. So, we send these commands via curl on the remotecontrol instead. Second, is there a right way to open models/surfaces from the ChimeraX python shell? At the moment, we also open pdbs via curl commands. For pdbs, maybe I just don’t know the right python command. However, I am re-implementing spherical harmonics referenced above. Once I have prepared the surface (“s”), if I add it to session.models, it crashes chimeraX if it is being added within the listener thread, and if there are no models already open. If another model is already open, then it works just fine, and if I load the same modules/execute the same commands from within the python shell provided in ChimeraX, it also works (regardless of other models being open). It seems like maybe ChimeraX isn’t able to fully initialize the session from within a thread. I’ve grouped these problems together because I guess mainly they have to do with being unable to do some actions from within threads, and if there’s an explanation/workaround for dealing with that. Thanks for your help! Cheers, Albert ––––––––––––––––––––––––––––––––––– Albert Smith-Penzel, Ph.D. DFG Project Leader Institute for Medical Physics and Biophysics Leipzig University 04107 Leipzig, Germany E-mail: albert.smith-penzel(a)uni-leipzig.de <mailto:albert.smith-penzel@medizin.uni-leipzig.de>

2 1

How to get the average pTM scores from an Alphafold multimer prediction
by Chen, Henry 04 Jan '23

04 Jan '23

Dear Chimarax devs/users, I'm using ChimeraX Alphafold tool to do a multimer structure prediction. I was wondering how can I determine the average pTM score for each structure prediction? I get PAE.png and plddt.png file, but not a text file or anything that tells me the average pTM. (I was hoping the ranking file would give the average score but it does not) The models come with the pdb file as well as the scores.json files, but again I don't see anything that tells me the average pTM for the structure. Clearly the pTM per residue scores are saved in my pdb file, is there a command to average them all up? Thanks! Henry

3 3

Capping termini ends
by Pranav Shah 04 Jan '23

04 Jan '23

Hi Team, Happy New Year! I look forward to all new developments you guys have in store for ChimeraX in 2023. Speaking of, is it possible to somehow display the ends of a protein termini with a spheres or something? I have certain view of my model in which the termini get partially obscured so it would be nice to prominently have them displayed in some manner... Best, Pranav -- Pranav Shah Postdoctoral Research Fellow. Division of Structural Biology, Wellcome Trust Centre for Human Genetics, University of Oxford, Roosevelt Drive, Oxford OX3 7BN, UK

2 2