5. Diagnostics ¶

5.1. Output files of GKV ¶

When finishing a run of GKV, all simulation output will be dumped in the output directory DIR set in the run/shoot script (for example in Listing 4.3 , DIR=/data/maeyama/gkv_training/test01/ ), as classified into the following directories,

DIR/
- log/ (Log files on simulation runs)
- cnt/ (Binary data for restart)
- fxv/ (Binary data of distribution functions $tilde{f}_{mathrm{s}bm{k}}(k_x,k_y,v_parallel,mu)$ at several positions of $z$)
- phi/ (Binary data of potentials, fluid moments, and variables in the entropy balance equation)
- hst/ (Ascii data of the GKV standard output)
- ... (Others are back up of the source code and environmental settings)

List of GKV output is summarized in Appendix List of GKV namelist .

Users may diagnose these output data by themselves. The present version of GKV provides two post-processing tools, gkvfig and diag_python on GKV GitHub page .

5.2. gkvfig - Generating PDF of GKV standard ASCII output ¶

gkvfig is a Python package that generates a summary figure PDF of GKV standard ASCII output. (Previous fig_stdout tool using shell/awk/gnuplot/latex is converted to Python.)

5.2.1. Installation ¶

To install from PyPI:

            $ pip install gkvfig

           

Or install the latest development version from GitHub:

            $ pip install git+https://github.com/GKV-developers/gkvfig.git

           

5.2.2. Usage ¶

5.2.2.1. (i) Basic usage: As a command line tool ¶

             $ python -m gkvfig -d DIR

            

The argument DIR is the path of GKV output directory. The namelist file DIR/gkvp.namelist.001 , log file DIR/log/gkvp.000000.0.log.001 , and hst directory DIR/hst/ should exist. You get a summary PDF file CWD/figpdf_yyyymmdd_hhmmss/fig_stdout.pdf .

5.2.2.2. (ii) As a Python function ¶

             from gkvfig import gkvfig

gkvfig(gkv_stdout_dir="YOUR GKV OUTPUT DIR")

You get a summary PDF file CWD/figpdf_yyyymmdd_hhmmss/fig_stdout.pdf , always in the current working directory CWD .

5.2.3. Dependencies ¶

gkvfig requires the following Python packages: - numpy , matplotlib , pandas , reportlab , pypdf

5.3. diag_python - Post-processing tool for BINARY output ¶

diag_python is a set of Python scripts, which read Zarr format files of GKV binary output. (Python version of the previous Fortran post-processing tool diag .)

5.3.1. How to use diag_python ¶

Copy whole diag_python/ into the output directory of GKV. For example,

YOUR_GKV_EXECUTED_DIR/
- diag_python/
- cnt/ (Zarr format files *.zarr will be read by diag_python)
- fxv/ (Zarr format files *.zarr will be read by diag_python)
- phi/ (Zarr format files *.zarr will be read by diag_python)
- hst/ (gkvp.mtr.001 will be read by diag_python)
- src/ (gkvp_header.f90 will be read by diag_python)
- log/
- gkvp_namelist.001 (gkvp_namelist.001 will be read by diag_python)

Initial settings in main.py :

            import sys
sys.path.append("./src/")
from diag_rb import rb_open, rb_get_tri_filelist
from diag_geom import geom_set
### Read Zarr store gkvp.phi.*.zarr/ by xarray ###
xr_phi = rb_open('../phi/gkvp.phi.*.zarr/')
xr_Al  = rb_open('../phi/gkvp.Al.*.zarr/')
xr_mom = rb_open('../phi/gkvp.mom.*.zarr/')
xr_fxv = rb_open('../fxv/gkvp.fxv.*.zarr/')
xr_cnt = rb_open('../cnt/gkvp.cnt.*.zarr/')
xr_trn = rb_open('../phi/gkvp.trn.*.zarr/')
tri_filelist = rb_get_tri_filelist('../phi/gkvp.tri.*.zarr/')
xr_tri_list=[]
for file in tri_filelist:
    xr_tri=rb_open(file + '.*.zarr/')
    xr_tri_list.append(xr_tri)
### Set geometric constants ###
geom_set(headpath='../src/gkvp_header.f90', nmlpath="../gkvp_namelist.001", mtrpath='../hst/gkvp.mtr.001')

           

Call functions, e.g.:

            from out_mominxy import phiinxy
# Plot phi[y,x] at t[it], zz[iz]
it = 3
iz = 8
phiinxy(it, iz, xr_phi, flag="display")

           

See help(phiinxy) for details.

For more details, Appendix Diagnostics modules in the post-processing program diag shows some examples of diagnostics modules.