Skip to content
Projects

Gallery helpers

The example scripts of the struphy-hub gallery write their results with plasma_plots.gallery, so a copied example runs on its own after pip install "plasma-plots[gallery]". Every helper writes to the current directory and names its files after the example’s stem; the website’s build reads exactly these files:

FileWritten by
<stem>.png, <stem>.plotly.json, <stem>.htmlsave_figure, save_extra_figure
<stem>.metadata.jsonmerge_metadata, save_extra_figure
<stem>-profile.h5 and its plot dataexport_profiling

The scripts also run under MPI. The simulation, the post-processing and the analysis run on every rank, and only rank 0 writes files, so a script needs no rank checks of its own. Importing the module sets Struphy’s logging level to INFO, which prints one block per time step, as one wants to see in a CI log.

from plasma_plots import gallery
stem = "landau_damping"
phi = out.evaluate("em_fields/phi")
gallery.save_figure(
gallery.space_time_figure(
phi.isel(eta2=0, eta3=0),
space="eta1",
title="Potential",
colorbar_title="φ",
),
stem,
)
extra = gallery.save_extra_figure(
gallery.heatmap_figure(
f.isel(t=-1),
x="eta1",
y="v1",
title="Distribution function",
xaxis_title="x",
yaxis_title="v",
),
stem,
"distribution",
alt="The distribution function in phase space at the last time",
caption="Phase-space holes form as the wave damps.",
)

save_figure writes a static PNG, the Plotly JSON and a standalone HTML page. heatmap_figure draws a 2-D array, space_time_figure a space-time map with colors symmetric about zero, and heatmap_movie a 3-D array as an animated heatmap over sweep. save_extra_figure saves a further figure and returns its entry for the page’s figures list.

rate = energy.plasma.analysis.damping_rate(window=(5.0, 20.0))
gallery.merge_metadata(stem, damping_rate=float(rate.rate), figures=[extra])
# a run made with profiling_activated=True
gallery.export_profiling(sim, stem)

merge_metadata adds fields to <stem>.metadata.json and keeps the ones already there. export_profiling writes the raw scope-profiler HDF5 file and the plot data the example page draws (see Profiling). is_root() and barrier() are the rank checks the helpers use, for code of your own.