StrainExtract seem not working

This question was expired because it remained in the 'Open' state without activity for the last 15 days.

Sorry to add another comment on this thread, but I would really appreciate some help

[Just adding another details and comment, in case that it might be helpful]
As the respond of the strainextract is the empty vtk file, I am not sure if this is the same case with this bug??
>>https://bugs.launchpad.net/esys-particle/+bug/1727586

So, I have tried what all I could do, reading and trying to understand through codes of main.cpp of either ExtractStrain and other tools (dump2vtk ,etc.). It seem that the main.cpp of strainextract tool just use the "infilename" variable, while others (like dump2vtk) just use "basefilename" instead.

Based on my understanding, my hypothesis is that the "infilename" only allows a single file for the input which mean that you are not allow to input "series" of checkpointer of a single timeframe. (In my case, I define 3 workers for simulation which make them produce 3 checkpointers per time frame.)

So, what I am thinking is that to make strainextract run properly, only a single worker must be define at the simulation (to produce only a single checkpointer file per timeframe, excluding the header). (This is why, I guessed, Tao said his strainextract still work well [https://answers.launchpad.net/esys-particle/+question/226129], because he might only have a single checkpointer.)

I am not sure if it is the right one, but to discuss and ask further, if this is the case, how can we edit the code of main.cpp to make strainextract also work for "series" of checkpointers (I am quite a beginner in programming, would appreciate very much if someone could help or guide me on this problem).

Hi Wasin,

Looking at the code for the DataExtractor.read(..) subroutine, it appears that StrainExtract needs you to supply the name of a checkpoint header file (ending in '_0.txt') as the input filename. Consequently, you should try a commandline like this:

 strainextract -i snapshot_t=600000_0.txt -o strain_60.vtu -grid 60 4 0 5.0

This should produce a single VTK file for timestep 600000, reading particle data from all of the worker's checkpoint files.

Assuming this works, you can issue the following command in the terminal to sequentially convert a whole series of checkpoint files:

for (( i = 0 ; i < 100 ; i++ )); do infilename='snapshot_'$i'0000_0.txt'; outfilename='strain_'$i'.vtu'; strainextract -i $infilename -o $outfilename -grid 60 4 0 5.0; done

This uses the shell 'for' loop to step through the sequence of checkpoint files.

Alternatively you can make a shell script to do this. Copy the following into a text file called "strain_convert.sh":

#start of strain_convert.sh
for (( i = 0 ; i < 100 ; i++ ))
do
   infilename='snapshot_'$i'0000_0.txt'
   outfilename='strain_'$i'.vtu'
   strainextract -i $infilename -o $outfilename -grid 60 4 0 5.0
done
#end of strain_convert.sh

Cheers,

Dion

Dear Dion,

Thank you for your kindly answer.
However, I have tried
>> strainextract -i snapshot_t=600000_0.txt -o strain_60.vtu -grid 60 4 0 5.0
and the results still be the empty vtk files.

Here is the respond from terminal:

version : 3
infilename : snapshot_t=600000_0.txt
nr. of filenames: 3
snapshot_t=600000_1.txt
snapshot_t=600000_2.txt
snapshot_t=600000_3.txt
inserted 0 particles

The command just work very quickly (less than a second) which I think it might not read those series of chekpointer files. (If i tried the in put as _1/_2/_3.txt file, it would take a while to process but still be the empty vtk result.)

Regards,
Wasin

Hi Wasin,

The output from the terminal suggests that strainextract is correctly reading the files when you supply the header file as the input filename.

Another possible reason for the failure to read particles is that your checkpoint files were generated in a simulation using NRotSphere as the particle type. It looks like strainextract expects the checkpoint files to contain rotational information for each particle (orientation quaternions and angular velocity). What particle type are you using in your simulations, NRotSphere or RotSphere?

Cheers,

Dion

Dear Prof. Dion,

I am using the RotSphere in my 2D simulation.

Regards.
Wasin

Thanks Dion Weatherley, that solved my question.