Job submission gui for Windows and Linux (python/tk based)

[PaulAltair]

Important news, as of 26th September 2024, the gui submission tool has reached a level of maturity that it is now migrated from a confluence download to be included in the official Tools repository:

https://github.com/OpenRadioss/Tools/tree/main/openradioss_gui

Since the original post in march 2023, many improvements have been made, see below for latest gui appearance: and download and contribute in the tools repository

Thanks again for the contributions of @hotaosa to this tool

---

To help make job submission easier (especially for the precompiled windows version) I have written a small, basic python/tk based job submission gui, Windows and Linux versions are available

Download and instructions for use are available at confluence pages:
python/tk guis in confluence

There are also python based launchers to run the Anim-vtk and T01-csv conversion tools

---

[PaulAltair]

Linux version now updated to support mpi submissions

[PaulAltair]

version now available supporting mpi under windows if installed

[ZdendaH]

Hello, Thank you for the work you do. I found it but I do not know how to use it. I double-click on anim-vtk.bat and the window shows just milliseconds and disappeared. The same happened with a file for submitting the job. I tried to run the Python script file in Visual Studio Code but I did not have success. I do not know what I am doing wrong.

Does anyone have an idea of what could help to solve it?
Thank you for your reply.

[PaulAltair]

For windows execution, there are now exes available that don't require python install as pre-requisite, the instructions for use are here GUI Job Submission Guide in Confluence Hope this helps

[PaulAltair]

FYI I today have uploaded v2 of the job submission guis, now supporting:

Multiple Engine Files (e.g. 0001.rad, 0002.rad, 0003.rad etc) select all in Engine Selection at runtime
Running 'only' Engine File(s) (for restarts, e.g. launch a 0002.rad and restart an existing run) Select only engine file(s)
Running 'only' Starter (if you want to check the Starter only), still creates .rst file(s) Select only starter file

The ability to run .k or .key (Dyna Format models) remains in place, and there is a version allowing selection of a vars.bat to assist mpi running under windows.

Let me know if you have any issues in your application, or any other ideas for improvement

Available from confluence pages:
Gui submission tools at confluence

[bthomas84]

Hi I encounter an issue. Trying to launch a radioss run that already ran on the commercial radioss, i have an error :
*** ERROR : Unknown command line argument: days
Since i don't write the command line myself thanks to the GUI I don't really understand where does it come from.

Kind regards

[PaulAltair]

Is there a space in your install path or filename?

The robustness of the gui when it/OpenRadioss are installed in a directory with spaces is not guaranteed!

Possibly it is trying to read part of the submission filename/path as an argument to the script? Try renaming file without space?

Paul

[bthomas84]

You're completely right... space in the directory name, thank you Paul!

[PaulAltair]

Is this under windows or linux btw... I should try and code something better for this when I get some time, but it isn't that simple due to the pushd/popd the script is doing to get the files in/out of the submission folder...

[bthomas84]

It's on Windows, don't worry a simple change of directory plays the trick. Thank you vey much for your so quick answer

[hotaosa]

Hello. I have written a gui(for Windows) supporting for STOP/KILL control, multiple engine files and queue of jobs. I referred to your script. I hope it helps you to imporove your gui.

openradioss_run_script_for_windows_mpi.zip

Please modify line 18 of "openradioss_run_script.bat" to set MPI_PATH
Multiple jobs can be queued and will be executed in order
If the next engine file(ex.0002.rad) exist in the same directory, it will executed automatically
When you submit starter file(0000.rad), the result files are deleted if they exist in the same directory

[PaulAltair]

Really nice, thank you! I like the sub windows and queuing a lot, its a lot neater than mine! :) I will take a deeper look and try to combine our efforts and port it to linux version. In the meantime, may I share your gui 'as is' also on the confluence page? Job submission guis page at OpenRadioss confluence

[PaulAltair]

hmm, ok, thanks for the catch, I hadn't considered jobs being run in the same folder... I'll give it some thought tomorrow!

[PaulAltair]

think this fixes it, windows only, will sort out linux tomorrow:
win_scripts_mk3.zip

[PaulAltair]

ok, actually, fixed linux too, and recreated win exe, confluence page zips are updated, or here:
UpdatedGuis240723b.zip

[PaulAltair]

I will tidy up some more tomorrow... the feature isn't needed for starter phase, and I think I can code something that removes the need for the 'running_job' file completely

[PaulAltair]

Ok, I couldn't get rid of the status file(s), but, I gave the stop, kill, anim, h3d buttons different behaviour during starter since none of them would work there (Stop/Kill will cause job to end at end of starter, Anim, h3d give message), and since I'd implemented that and I could reuse the status check, I added 'starter only' checkbox to the main submission window, this allows user to run starter only even if engines are present. Dropdown now includes link to confluence page too.
UpdatedGuis240723c.zip

[PaulAltair]

I have now updated the confluence page reflecting the latest updates: Confluence guis page

[PaulAltair]

@hotaosa I will publicise this more widely via linkedin etc next week, I will credit you by your github username, if you also wish to be credited in person (linked in profile etc) let me know!

[hotaosa]

Thank you for offering to credit me. I don’t need any more specific credit. I'm looking forward to it!

[PaulAltair]

I posted to LinkedIn today, https://www.linkedin.com/posts/paul-sharp-08a4a436_opensourcecommunity-opensource-openradioss-activity-7089596719952211968-2Ry_?utm_source=share&utm_medium=member_android

[MzHf0]

Hello,
I am trying to use the gui for job submission but always get the message:

ERROR :
Reader configuration files are not available
Check and set RAD_CFG_PATH variable.
Standard configuration file installation is
%ALTAIR_HOME%\hwsolvers\radioss\cfg

Can anyone help me with that?
Tank you.

[hotaosa]

Hello. It seems the path to the solver directory is not set up correctly. The script folder should be extracted the installation folder of OpenRadioss. (or you need to modify "openradioss_run_script.bat".) Please refer to here: https://openradioss.atlassian.net/wiki/spaces/OPENRADIOSS/pages/58785793/Latest+gui+scripts

[MzHf0]

Thank you for your quick answer! I just figured out the problem. Somehow some files where not unziped correctly. A new download and unzipping of Open Radioss solved the problem.

[JasGarnier]

I got this error too. It seemed to be caused by having the script files folder in the openradioss folder. I moved it elsewhere and explicitly set the path via "set OPENRADIOSS_PATH=C:\OpenRadioss" and the error disappeared.

[ser2177]

Hello,
I am attempting to utilize the latest gui script for submitting job applications, but I consistently receive the following message:

'findstr' is not recognized as an internal or external command,

operable program or batch file.

[ser2177]

There are no error messages, but radioss doesn't work.

[PaulAltair]

Strange, going back to basics, can you confirm that you have placed the gui submissions folder inside your openradioss installation (same level as exec) that OpenRadioss is installed in a path with no spaces in it, and that you are trying to run a filename with no spaces from a directory/path with no spaces?

Does a commandline submission work ok?

[ser2177]

I agree with you. I will uninstall OpenRadioss and perform a clean installation. I have changed too many parameters. There is a possibility that it is not working due to one of them.

I can run it from the command line.

[PaulAltair]

if it works from commandline, it ought not be necessary to uninstall/reinstall, but it is worth a go...

when you state that there are no errors but radioss doesn't work, what does happen? do you get the console pop up and no content? or the console never launches?

[JasGarnier]

I got that too. Solved it by adding PATH=C:\Windows\System32 in the Windows environment variables. Normally I'd have thought that would be there already but some software installs might screw it up.

[ser2177]

The console opens as shown in the image below.

[PaulAltair]

Ok, not sure what the issue is behind this still then. Looks like the python/tk bits are working ok, but there is still an issue with the bat script for your setup.

I'm out of office today, but tomorrow I will look into it further and see if we can narrow down the cause by running the bat from commandline outside of the gui.

Thanks for your patience

[ser2177]

Thank you for your interest. Today, I performed a clean installation and added some commands to the batch file as system paths. I also believe that there are no issues with Python/tk. However, after the clean installation, when I run it from the command line, I encounter a new error.

[ser2177]

I deleted the line “set OPENRADIOSS_PATH=%ABS_PATH%” and defined it on the system. It It works in both SMP and MPP modes. Thank you for your assistance.

[PaulAltair]

Ah, ok, thanks for the update, that line would normally pick up the OpenRadioss install path based on the location of the run script folder (inside OpenRadioss install) not sure why that wouldn't work for you, but glad you have a workaround!

[GNSamuelGao]

When I using this gui in Centos7, there is error message, as below. Could you help to figure out how to solve this error? I already install python3-tk, or not installed right?
Traceback (most recent call last):
File "openradioss_run_script_for_linux_mpi_ps.py", line 18, in
icon_image = tk.PhotoImage(file='./icon/ross.png')
File "/usr/lib64/python3.6/tkinter/init.py", line 3545, in init
Image.init(self, 'photo', name, cnf, master, **kw)
File "/usr/lib64/python3.6/tkinter/init.py", line 3501, in init
self.tk.call(('image', 'create', imgtype, name,) + options)
_tkinter.TclError: couldn't recognize data in image file "./icon/ross.png"

[martijnt1]

Seems to be centOS7 lacking PNG support, have you installed libraries [libpng-dev] and [zlib], or tried converting the PNG into a JPG + updateling the icon extension in the script? Mar 20, 2024, 02:58 by ***@***.***:

…

When I using this gui in Centos7, there is error message, as below. Could you help to figure out how to solve this error? I already install python3-tk, or not installed right? Traceback (most recent call last): File "openradioss_run_script_for_linux_mpi_ps.py", line 18, in icon_image = tk.PhotoImage(file='./icon/ross.png') File "/usr/lib64/python3.6/tkinter/> init> .py", line 3545, in > init Image.> init> (self, 'photo', name, cnf, master, **kw) File "/usr/lib64/python3.6/tkinter/> init> .py", line 3501, in > init self.tk.call(('image', 'create', imgtype, name,) + options) _tkinter.TclError: couldn't recognize data in image file "./icon/ross.png" — Reply to this email directly, > view it on GitHub <#1231 (comment)>> , or > unsubscribe <https://github.com/notifications/unsubscribe-auth/A4LGM5OPNDAYMQJLBIWEKBLYZDULZAVCNFSM6AAAAAAV3XBBNKVHI2DSMVQWIX3LMV43SRDJONRXK43TNFXW4Q3PNVWWK3TUHM4DQNBXGU4TK>> . You are receiving this because you are subscribed to this thread.> Message ID: > <OpenRadioss/OpenRadioss/repo-discussions/1231/comments/8847595> @> github> .> com>

[GNSamuelGao]

Thanks martijnt1. I have installed many libraries related png, jpg... But it's still error. I have remove all the image code, and it can run now. Without any good appearance, but it works at least. :)

[GNSamuelGao]

Another issue is how to use MPI in Linux. I cound not find the OneAPI in OpenRadioss folder. As below code quote:

export PATH=$OPENRADIOSS_PATH/extlib/intelOneAPI_runtime/linux64:$PATH

[MarianBulla]

Hi @GNSamuelGao,
on Linux, you can also use OpenMPI. I'm showing it here: https://www.youtube.com/watch?v=ddvH2CNfaYg&t=466s
But you should also be able to use IntelMPI.
I did not test it on Linux, but on windows here: https://www.youtube.com/watch?v=Uo5miT_a-YY&t=393s

Hope it helps you ?

Thanks and best regards,
Marian

[martijnt1]

Glad its running, w your workaround.Not knowing if you tried the 2nd possible solution mentioned, to adjust the PNG image+code for a JPG. Confirming Marian's feedback, MPI is running at least on Debian & Ubuntu, unable to comment on CentOS (RHEL) yet. Mar 21, 2024, 07:33 by ***@***.***:

…

Thanks martijnt1. I have installed many libraries related png, jpg... But it's still error. I have remove all the image code, and it can run now. Without any good appearance, but it works at least. :) — Reply to this email directly, > view it on GitHub <#1231 (reply in thread)>> , or > unsubscribe <https://github.com/notifications/unsubscribe-auth/A4LGM5OELXPZK3JJXIRDYBLYZJ5KNAVCNFSM6AAAAAAV3XBBNKVHI2DSMVQWIX3LMV43SRDJONRXK43TNFXW4Q3PNVWWK3TUHM4DQNRRGQ2DQ>> . You are receiving this because you are subscribed to this thread.> Message ID: > <OpenRadioss/OpenRadioss/repo-discussions/1231/comments/8861448> @> github> .> com>

[GNSamuelGao]

Hi @martijnt1 & @MarianBulla ,

Thanks both. I am a new hand for Github. I have tried different way, but I still have below questions

1. For icon issue, I have tried both PNG and JPG and installed many lib, but it still can not recognize the image file. But it can work at least~
   Error: _tkinter.TclError: couldn't recognize data in image file "./icon/ross.jpg" .
   

2. For MPI, maybe I need to build the "engine_linux64_gf_ompi" by myself? Because from the latest GUI code, it needs intelOneApi lib in the exec folder, see pic, but I can't find the lib in exec folder.
   
   I also try to install oneapi, but there is also error: cannot find libmpi.so.40. I check my resource, only libmpi.so.12
   

Br
Thanks

[PaulAltair]

Sorry, I had missed this reply, apologies for the lengthy time to respond. I think in my ubuntu install (where I wrote the gui), I installed ompi from apt-get rather than compiling it, I should update it to use a compiled ompi per the install documentation. I'll look into it.

[PaulAltair]

The installation for OPENMPI under linux is decribed in the HOWTO.md

in the openradioss_run_script_ps.sh can you then try adding these 2 lines? it is working for me then with self compiled open mpi per the HOWTO

export LD_LIBRARY_PATH=/opt/openmpi/lib:$LD_LIBRARY_PATH
export PATH=/opt/openmpi/bin:$PATH

[GNSamuelGao]

Hi Paul,

Thanks for your reply. I have tried again by following HowTo and you advice. It works now. Thanks.
But here come another issue is the GUI is not fit to Win11. Which parameter I can change to change the GUI to show all text. I can change the GUI width but not height.

[PaulAltair]

ok, I use win11 myself, and it is ok here, so perhaps it is down to graphics scaling settings (do you have text scaling set higher than 100%?)

I just tried setting mine at e.g. 150% and can replicate the issue you have

the size of the dialog is set in the file 'openradioss_run_script_for_windows_mpi_ps.py' lines 18-20

you can change the values there or set 'root.resizable' as True, True to allow you to change the height

you can also change the size of the buttons if you need to in the same file on lines 183, 185, 187, 189

I'm not sure how to do a more general fix for robustness on all installs, I found some info about importing ctypes for DPI awareness in tk, but they don't seem to help here, I'll keep looking into it!

[GNSamuelGao]

Thanks for the detail info. It's useful.

[PaulAltair]

With the great initiative of the Vortex-Radioss anim-d3plot converter: Discussion 2361 I have now have updated the job submission gui: available here: Job Submission guis

So that there is now an option to run the Vortex-Radioss d3plot conversion at the end of a run if it is installed

Let me know if you have any difficulties

[Vortex-CAE]

Thank you Paul! This is a great development to help improve workflow!

[PaulAltair]

Based on feedback, I made some small updates to the packages today, both linux and windows version include a helper d3convert.py script

[hdrz]

Hi,

I started using the run script on linux, and it runs fine. Just a small suggestion, maybe add a shebang line at the top of the script, so double clicking it (or directly run it from the commandline) wil work:

#!/usr/bin/env python3

This should work on all linux distributions.

Thanks!

[PaulAltair]

Thanks, I have added this in the latest (mk5) version

[PaulAltair]

Latest Mk5 versions available, the gui now detects if Vortex-Radioss python module is present or not, and presents the gui with/without d3plot conversion options as appropriate.

Also, if present, a d3plot button is added to run window, allowing conversion of any anim files present in the run directory while the job is running

[PaulAltair]

As of September 26th 2024, The gui is now migrated to the official OpenRadioss tools repository, downloads and contributions should be done there from now on https://github.com/OpenRadioss/Tools/tree/main/openradioss_gui

[hoanpqh]

Hello, I can not see Kill, Stop, Anim button. Please show me how to solve this problem!

[PaulAltair]

Are you using the current build in tools repository? In Linux or Windows? Does resizing the window make them visible? (they disappear if you make the window smaller)

[PaulAltair]

I just submitted a request to stop this happening (buttons should remain visible on resizing)

[hoanpqh]

Are you using the current build in tools repository? In Linux or Windows? Does resizing the window make them visible? (they disappear if you make the window smaller)

Hello, I use the current build in tools repository in Linux.
Resizing the window does not make them visible!

[PaulAltair]

I just tested current build in my Lubuntu Linux and it is working ok for me there: What linux version are you using, maybe it is due to graphics/desktop implementation there, does the gui otherwise work ok? e.g. queue window, submission window buttons?

[hoanpqh]

I increased display solution and now I can see the buttons. By the way, I can not open the GUI by running OpenRadioss_gui.bash. I have to run python3 OpenRadioss_gui.py instead.

[PaulAltair]

what happens when you try to run the bash?

[PaulAltair]

I think if you take the latest build from tools the modification I made the other day should work even at the lower resolution btw

[hoanpqh]

what happens when you try to run the bash?

It works now. Thank you very much.

[PaulAltair]

Great, your screenshot still shows the older version though!

The newer release should be in next stable build release too, so if you have this one working now then you can wait for that (It changes check boxes to drop-down menu as well as change to window buttons)