Simple mcplot and mcdisplay prototypes based on Scipp and Plopp

[willend]

Free-form text area

Please describe what your PR is adding in terms of features or bugfixes:

Plopp-based prototypes: mcplot-plopp and mcdisplay-plopp

Two new Python frontends built on ESS plopp/scipp, alongside the existing matplotlib and pyqtgraph tools. Both reuse mccodelib for loading and tracing. By default they open a one-cell JupyterLab notebook with the interactive UI (ipympl/pythreejs widgets); -o FILE writes a static png/pdf/svg instead.

mcplot-plopp / mxplot-plopp

• ASCII output is loaded via McCodeDataLoader and converted to scipp DataArrays:
  • 1D monitors get I_err² as variances and bin edges from xlimits
  • 2D monitors use dims ('y', 'x'), the same orientation as the matplotlib frontend
• NeXus output: a mccode.h5/.nxs file, or a directory with mccode.h5 and no mccode.sim (the same rule the matplotlib/pyqtgraph mcplot use before handing off to nexpy).
  • 2D monitors are stored (nx, ny) in NeXus, the transpose of the ASCII block, and are transposed to the same ('y', 'x') layout.
  • The NeXus errors block gives 2D monitors variances.
  • h5py is imported only when reading NeXus.
• Units are read from [...] in the axis labels (AA/Angs, meV, deg, us, cm, m, s). Unknown units stay in the label and the axis is dimensionless.
• Scans: each monitor is stacked over the scan parameter (1D→2D, 2D→3D).
  • ASCII: numbered step folders, scan values from mccode.dat.
  • NeXus: one NXentry per step, stacked over the first parameter that varies across entry*/simulation/Param.
• Event data: list-mode Monitor_nD pixel banks (neutron pixel min=… t, list all neutrons) in NeXus files are read via McStasToX.
  • McStasToX needs the source and sample components. The UI offers dropdowns (defaulting to components named like "source"/"sample"); the CLI takes --source/--sample.
  • Shows a pixel intensity map (3D scatter in the UI, z–x top view with -o) plus time-of-flight and wavelength histograms.
  • Wavelength per event uses the source → sample → pixel flight path, λ = 2π·AA2MS·t/L, with AA2MS from mcstas-r.h.
  • mcstastox is an optional dependency, imported only for event data.
• UI: a monitor dropdown showing one figure at a time; scan monitors get plopp's slicer with a ▶ player.

mcdisplay-plopp / mxdisplay-plopp

• Traces with --trace=2. Solids are drawn as 3D meshes with pp.mesh3d:
  • cylinder (solid or hollow), cone, disc, annulus, sphere
  • box (solid or hollow)
  • polyhedron (OFF and polygon geometry)
• Lines, multilines and circles are drawn as thin tubes, with small spheres at corners.
• Components are coloured per component; rays are thinner grey tubes.
• UI: a params field, a ray count, a Trace button, and a thickness slider that redraws without re-running mcrun.
• -o writes 2D z–x / z–y projections instead (solids as triangle outlines), since pythreejs can't render without a browser.
• -y/--yes is forwarded to mcrun (use defaults for unset parameters). Without it, mcrun prompts as usual. The UI always passes it, because a widget can't answer prompts.
• Shape conventions follow mcdisplay-webgl, the reference trace=2 renderer: shapes are centred on (x,y,z), a box's local y is turned onto n, and a cone's apex is at +h/2.

Dependencies

• scipp and plopp are required. Missing either is caught by guarded imports, which exit with a message plus conda and pip install hints. The UI also needs ipympl and jupyterlab.
• h5py (NeXus) and mcstastox (event data) are imported only when needed.

Build and packaging

• CMake subprojects tools/Python/{mcplot,mcdisplay}/plopp, set up in parallel to the matplotlib tools. The bash wrappers only check that the tool script is executable.
• Windows (MinGW):
  • .bat wrappers registered in InstallMCCODE.cmake
  • mkdist … nsis64 lines in build_windows_mcstas / build_windows_mcxtrace
  • both packages added to the MinGW metapackage .iss files
  • scipp and plopp are not added to meta-pkgs/windows/environment.yml; until users install plopp, the tools print the install message
• No Debian packages: scipp/plopp aren't packaged by Debian.

Testing

• Checked on macOS in a conda env with plopp/scipp 26.9.0.
• mcplot-plopp:
  • real single-run, scan and McStas 2.7 (Odin, 117 monitors) ASCII output
  • NeXus from mcrun --format=NeXus: a non-square 30×10 PSD plus a 1D monitor gives DataArrays identical to the same-seed ASCII run (dims, edges, values); a 3-step lambda scan stacks correctly; BNL_H8 single and scan load all 9 monitors
  • events: a scratch instrument with two 32×16 pixel banks (one rotated 60°) and an Incoherent sample gives wavelengths 1.50–2.51 Å for a 2 ± 0.5 Å source, time of flight 1.1–1.9 ms over 3 m, and banks placed correctly (mcstastox 0.0.12)
• mcdisplay-plopp: BNL_H8, mini, Polygon_ESS_butterfly_image, and a scratch instrument with Incoherent in every solid shape (bounding boxes match the parameters, including a hollow box rotated 30°).
• All UIs, including the 3D views, run without errors in executed notebooks (nbconvert --execute).
• Assert-based self-checks in test_mcplopp.py and test_mcdisplopp.py: units, orientation, a synthetic mccode.h5 (variances, scan stacking), tube and solid geometry.
• CMake configure/build/install checked standalone for both McStas and McXtrace.

Not yet tested: Windows/NSIS builds, Linux, the McXtrace runtime, and the 3D views in a live browser session.

Known gaps / follow-ups

• 2D monitors read from ASCII have no variances (mcplotloader skips the # Errors block); NeXus input has them.
• NeXus scans over several parameters use only the first one that varies as the scan axis.
• Event data is NeXus only. McStasToX doesn't read ASCII, and ASCII list files carry no rotations or pixel geometry. McStasToX also reads only entry1, the first step of a NeXus scan.
• The notebook cell must be run by hand. Voila could make it a self-starting app, but it would be a new dependency.
• The man-page .md.in docs aren't written yet.
• annulus()'s 5th argument is named inner_radius in the runtime but drawn as a ring width by both webgl and matplotlib. I followed the frontends; the naming could be clarified.

---

Declaration of use of AI-tools

• Please add a checkmark here if you used AI-tools during the work for this contribution
• Furter, please describe how / where and for what the tools were used:

^ 🤖 Generated with Claude Code

---

Development OS / boundary conditions

Please describe what OS you developed and tested your additions on, and if any special dependencies are required:

---

PR Checklist for contributing to McStas/McXtrace

For a coherent and useful contribution to McStas/McXtrace, please fill in relevant parts of the checklist:

• My contribution contains something else

  • Explanation is added in free form text above or below the checklist

---