Assigning orientations from file

[xshang93]

Hi, I'm trying to assign orientation for my 3-scale tessellation structure. Now I'm testing the -ori 'file(filename)' option on a 2-scales tessellation.

First of all, is there a format that the file needs to follow or do I just do something like:
1 2 3
1 2 3
...
I used this format, output the .ori file, no error was generated. But the orientations are just not what i specified, they are still the same as the randomly generated orientation.

Second, in the neper documentation on page 19, it says "if Rodrigues vector (rodrigues), the descriptor can be omitted." but on page 79, I found "When loading orientations from an external file, ‘file(file_name)’ assumes orientations described by Euler angles in Bunge convention (and active convention).". So which one of these is used by default?

Third, for a 2 or 3 scale tessellation, is there a way to find the grain IDs so I know which one to assign? Or do I just use 3 different files separated by :: ?

Many thanks in advance.
XS

[smallrain1996]

I've also noticed that for the same orifile, the following two commands get different .ori files
neper -T -n 2::4::3
-ori "file(orifile,des=rodrigues)"
-format "tess,ori"
-o test1

neper -T -n 24
-ori "file(orifile,des=rodrigues)"
-format "tess,ori"
-o test2

[xshang93]

I think that's expected. my guess is, when you give -ori "file(orifile,des=rodrigues)" in your three scale tessellation, it only reads it for the first level, and uses the default (which is random) for the second and third levels. At the end of the day, all the output orientations should only be defined by your third level -ori file, and should have nothing to do with the first and second level, yet you still need to specify them.

I still need to do more test on this.

XS

[rquey]

@xshang93 is correct. See #325 (comment) for a proper solution.

[xshang93]

I did some more tests and I think I can answer some of my questions:

1. the file format is correct.
2. It's using rodrigues by default. I tested that out by doing a simple one scale tessellation.
3. I think the correct way to use -ori is, using a 2-scale tessellation as example, -ori "file1::file2". But when I did this, neper stucked, and is using 99.9% of my CPU but not moving forward. The output from neper is attached below (I'm assigning random orientations for the second tessellation level):

========================    N   e   p   e   r    =======================
Info   : A software package for polycrystal generation and meshing.
Info   : Version 4.2.1-7
Info   : Built with: gsl|muparser|opengjk|openmp|nlopt|libscotch (full)
Info   : Running on 8 threads.
Info   : <https://neper.info>
Info   : Copyright (C) 2003-2021, and GNU GPL'd, by Romain Quey.
Info   : No initialization file found (/Users/Xiao/.neperrc').
Info   : ---------------------------------------------------------------
Info   : MODULE  -T loaded with arguments:
Info   : [ini file] (none)
Info   : [com line] -n 2::from_morpho -id 1::1 -morpho
         gg::lamellar(w=file(lamwidth),v=(1,0,0)) -ori
         file(ori,des=euler-bunge)::file(ori2,des=euler-bunge) -format
         tess,ori -o Polycrystal_laminated
Info   : ---------------------------------------------------------------
Info   : Reading input data...
Info   : Creating domain...
Info   : Creating tessellations...
Info   :   [level] Running in multiscale mode (2 scales)...
Info   :   [    1]  1/1 Info   :   - Setting seeds... 100%              
Info   :   - Generating crystal orientations...
Info   :     [i] Parsing file 'ori'...
Info   :     [i] Parsed file 'ori'.
Info   :   - Running tessellation...
Info   :     > Initial solution: f   =13.325428096
Info   :     > Iteration    218: fmin=1.800334242 f=1.800334242  
Info   :     > Final solution  : f   =1.800334242 (218 iterations)
Info   :     > Reached 'eps' criterion.
Info   :   [    2]  1/2 Info   :   - Running tessellation...            
Info   :   - Generating crystal orientations...

Can someone help with this?

[smallrain1996]

It may take a developer of the software to fundamentally reply to this technique, but I can offer two compromises.

If you are in need of subsequent calculations using Fepx, you can actually use a separate orientation file to provide orientation information regardless of how Neper generates the orientation.

If you just need to present the correct orientation information in the .tess file, you can write a script that modifies the content of the specified area in the .tess, because .tess is essentially a text file, and this is not difficult.

[xshang93]

These are very good points. Thanks. I think this is what I'll do for now.

XS

[rquey]

I did some more tests and I think I can answer some of my questions:

1. the file format is correct.

Yes.

2. It's using rodrigues by default. I tested that out by doing a simple one scale tessellation.

Yes.

3. I think the correct way to use `-ori` is, using a 2-scale tessellation as example, `-ori "file1::file2"`. But when I did this, neper stucked, and is using 99.9% of my CPU but not moving forward. The output from neper is attached below (I'm assigning random orientations for the second tessellation level):

[...]

Thanks for the report. Neper now generates an error when the input file is not properly formatted (instead of looping infinitely). Otherwise, see #325 (comment) for a proper solution.

[xshang93]

Thank you. I had the feeling that not moving forward while eating up CPU signifies an error. Thank for confirming this. I guess the most recent issue I had, i.e. #326 , is due to such an issue too. However, I was not able to figure out what cost the error and it would be great if you could give some insight.

[rquey]

Hi, I'm trying to assign orientation for my 3-scale tessellation structure. Now I'm testing the -ori 'file(filename)' option on a 2-scales tessellation.

First of all, is there a format that the file needs to follow or do I just do something like: 1 2 3 1 2 3 ... I used this format, output the .ori file, no error was generated. But the orientations are just not what i specified, they are still the same as the randomly generated orientation.

The file should be formatted as a .ori file, so, just a list of orientations.

Second, in the neper documentation on page 19, it says "if Rodrigues vector (rodrigues), the descriptor can be omitted." but on page 79, I found "When loading orientations from an external file, ‘file(file_name)’ assumes orientations described by Euler angles in Bunge convention (and active convention).". So which one of these is used by default?

Thanks for pointing this out. It is Rodrigues vectors. Fixed in version 4.2.1-14.

Third, for a 2 or 3 scale tessellation, is there a way to find the grain IDs so I know which one to assign? Or do I just use 3 different files separated by :: ?

Many thanks in advance. XS

This may seem a bit tricky at first, but the way to do it is as follows. For a 2-scale tessellation for which you would want to specify the orientations of the cells at scale 2:

neper -T -n 3::4 -ori "random::file(scale2_file)"

where scale2_file is a multiscale file that should contain, for each scale-1 cell, the orientation distribution. It could be random (as at scale 1), but it can also be of the type file(...) to load custom orientations, or any other input described in option -ori. In your case, the content of scale2_file could be:

1 file(cell1)
2 file(cell2)
3 file(cell3)

where cell1 would be an orientation file, which would contain the orientations of the 4 cells subdividing the 1st scale-1 cell, e.g.,

   0.704910491805    1.813296457703    0.720973368800
  -1.755963753575    1.732321781503    4.425698589080
   2.557193098119   -1.540193224736   -0.227983822696
  -0.349045017160    0.088659451979   -0.322358107095

Same for cell2 and cell3.

This can be generalized to 3 scales and more.

[xshang93]

Thanks for the reply. I'll test it out tonight!

[janzenchoi]

Hello @rquey,

I just tried the above commands with the newest version of Neper. However, I got a segmentation fault from running them.

Are those commands / formats no longer supported in the newest version of Neper? Is there a new set of commands I could use instead?

Thanks.

[rquey]

Please provide the input necessary to reproduce the bug (index.txt) and copy text as text (using Markdown formatting), not images... This is cleaner and saves us from re-typing commands...

[janzenchoi]

Hello @rquey,

Sure thing. The commands and files should be quite similar to the ones you suggested above.

The index.txt file contains:

1 file(cell1.txt)
2 file(cell2.txt)
3 file(cell3.txt)

The cell1.txt, cell2.txt, and cell3.txt files all contain:

   0.704910491805    1.813296457703    0.720973368800
  -1.755963753575    1.732321781503    4.425698589080
   2.557193098119   -1.540193224736   -0.227983822696
  -0.349045017160    0.088659451979   -0.322358107095

Finally, after creating those files, I ran the following command.

neper -T -n 3::4 -ori "random::file(index.txt)" -o img

I seem to get a segmentation fault from it.

I was trying to apply crystal orientations to a multiscale and got a similar error, which is how I found this discussion post.

[rquey]

Here is what I get:

========================    N   e   p   e   r    =======================
Info   : A software package for polycrystal generation and meshing.
Info   : Version 4.5.1-2
Info   : Built with: gsl|muparser|opengjk|openmp|nlopt|libscotch (full)
Info   : Running on 8 threads.
Info   : <https://neper.info>
Info   : Copyright (C) 2003-2022, and GNU GPL'd, by Romain Quey.
Info   : Loading initialization file `/home/rquey/.neperrc'...
Info   : ---------------------------------------------------------------
Info   : MODULE  -T loaded with arguments:
Info   : [ini file] -tesrformat ascii
Info   : [com line] -n 3::4 -ori random::file(index.txt) -o img
Info   : ---------------------------------------------------------------
Info   : Reading input data...
Info   : Creating domain...
Info   : Creating tessellations...
Info   :   [level] Running in multiscale mode (2 scales)...
Info   :   [    1]  1/1 Info   :   - Setting seeds...                   
Info   :   - Generating crystal orientations...
Info   :   - Running tessellation...
Info   :   [    2]  1/3 Info   :   - Setting seeds...                   
Info   :   - Generating crystal orientations...
Info   :     [i] Parsing file `cell1.txt'...
Info   :     [i] Parsed file `cell1.txt'.
Info   :   - Running tessellation...
Info   :   [    2]  2/3 Info   :   - Setting seeds...                   
Info   :   - Generating crystal orientations...
Info   :     [i] Parsing file `cell2.txt'...
Info   :     [i] Parsed file `cell2.txt'.
Info   :   - Running tessellation...
Info   :   [    2]  3/3 Info   :   - Setting seeds...                   
Info   :   - Generating crystal orientations...
Info   :     [i] Parsing file `cell3.txt'...
Info   :     [i] Parsed file `cell3.txt'.
Info   :   - Running tessellation...
Info   : Flattening 2-scale tessellation...
Info   :   [level] Flattening 2-scale tessellation...
Info   :   [    1]  1/1: id=1                                           
Info   : Writing results...
Info   :     [o] Writing file `img.tess'...
Info   :     [o] Wrote file `img.tess'.
Info   : Elapsed time: 0.051 secs.
========================================================================

[janzenchoi]

In text form, mine is:

========================    N   e   p   e   r    =======================
Info   : A software package for polycrystal generation and meshing.
Info   : Version 4.5.1-2
Info   : Built with: gsl|muparser|opengjk|openmp|nlopt|libscotch (full)
Info   : Running on 16 threads.
Info   : <https://neper.info>
Info   : Copyright (C) 2003-2022, and GNU GPL'd, by Romain Quey.
Info   : No initialization file found (`/home/janzen/.neperrc').
Info   : ---------------------------------------------------------------
Info   : MODULE  -T loaded with arguments:
Info   : [ini file] (none)
Info   : [com line] -n 3::4 -ori random::file(index.txt) -o img
Info   : ---------------------------------------------------------------
Info   : Reading input data...
Info   : Creating domain...
Info   : Creating tessellations...
Info   :   [level] Running in multiscale mode (2 scales)...
Info   :   [    1]  1/1 Info   :   - Setting seeds...
Info   :   - Generating crystal orientations...
Info   :   - Running tessellation...
Info   :   [    2]  1/3 Info   :   - Setting seeds...
Info   :   - Generating crystal orientations...
Info   :     [i] Parsing file `cell1.txt'...
Info   :     [i] Parsed file `cell1.txt'.
Info   :   - Running tessellation...
Info   :   [    2]  2/3 Info   :   - Setting seeds...
Info   :   - Generating crystal orientations...
Info   :     [i] Parsing file `cell2.txt'...
Info   :     [i] Parsed file `cell2.txt'.
Info   :   - Running tessellation...
Info   :   [    2]  3/3 Info   :   - Setting seeds...
Info   :   - Generating crystal orientations...
Segmentation fault

That really is weird. I think the only difference in our outputs before the segmentation fault is the ini file.

[rquey]

Please attach your txt files.

[janzenchoi]

Sure thing.
files.zip