vmecdash 0.1.0

Dash and VS Code tooling for inspecting VMEC wout NetCDF equilibria

VMECdash

An interactive Dash application for exploring VMEC wout_*.nc stellarator equilibria.
It reconstructs magnetic surfaces with JAX, renders 1‑D/2‑D/3‑D plots and so on, motivated by the original MATLAB tool VMECplot.m.

Motivation: Quickly inspect the VMEC equilibrium quantities of interest without relying on difficult-for-starter-to-compile programs (such as libstell) or memory-heavy commercial software (such as MATLAB).

---

Features

• File upload for standard VMEC wout NetCDF output.
• Summary dashboard with key scalar diagnostics and highlighted metadata (free/fixed boundary, mgrid file, etc.).
• 1‑D profiles for rotational transform, safety factor, pressure, volume derivative, beta metrics, and more.
• 2‑D visualisations:
  • R‑Z cross sections with geometry or interpolated scalar fields.
  • θ‑ζ flux-surface contours across a field period.
• 3‑D flux surfaces with optional coordinate‑free background shading.
• Field lines viewer for $\alpha-\zeta$ coordinates on selected flux
  • A nice choice for visualising magnetic ripple
  • 2D plots of $|B|(\alpha, \zeta)$
  • 1D plots of single period field line traces with different initial $\alpha$ values
  • Single field line transistions.
• JAX acceleration for geometry reconstruction.

---

Requirements

Python 3.10+ is recommended. Install dependencies with conda:(optional)

conda activate your_env_name
pip install -r requirements.txt

Note for JAX: If you don't have a GPU/TPU or your gpu is not supported for FP64(like metal), then simply running pip install jax[cpu] to install the CPU-only version is good enough. GPU/TPU wheels require platform-specific instructions from the JAX documentation.

---

Running the App

python VMECdash.py

Dash defaults to http://127.0.0.1:8050/. The layout is responsive, so you can resize the browser to focus on plots or the control sidebar.

The packaged entry point is also available after installation:

pip install -e ".[dash]"
vmecdash serve

---

VS Code Native Preview

VMECdash now includes the first native VS Code integration surface:

• a Dash-free Python backend at python -m vmecdash.vscode_backend --stdio;
• a workspace VS Code extension under extension/;
• a Custom Readonly Editor for wout*.nc files;
• a Webview UI that renders backend Plotly figures with bundled plotly.min.js.

For local development, install the Python package in the interpreter VS Code should use:

pip install -e .

Then open extension/ as a VS Code extension development project or package it as a VSIX. In Remote-SSH, install vmecdash in the remote Python environment and set vmecdash.pythonPath if VS Code does not pick the desired interpreter.

---

Usage Tips

1. Upload a VMEC NetCDF file (wout_*.nc) via the drag‑and‑drop area in the sidebar.
2. Use Visualization Mode to switch between:
   • Summary Dashboard (shows statistics + multi-panel plots),
   • 1D Profiles, 2D Cross Sections, 2D Flux Surfaces, and 3D Geometry.
3. Adjust sliders for toroidal angle (phi), flux surface index (s), and choose different physical quantities from the dropdowns.
4. Toggle Coordinate-free background in 3‑D mode for clean screenshots.
5. Click Download Plot to export the currently visible figure as a PNG.

The app caches equilibrium metadata, so switching modes or variables is fast.
For heavy 2‑D physics overlays, a pre-computation step runs on the server while keeping the UI responsive.

---

Repository Layout

Path	Description
VMECdash.py	Main Dash entry point (layout + callbacks).
vmec_jax.py	JAX-powered data processor for VMEC equilibria.
views/	Modular UI components (overview, profiles, 2D/3D, fieldlines).
ui/	Shared UI helper components.
requirements.txt	Python dependencies.
example/wout_PO.nc	Example VMEC equilibrium (use your own files for new cases).

---

TroubleShooting

Feel free to open issues or pull requests to add new physical quantities, UI tweaks, or performance optimisations. Enjoy exploring your VMEC equilibria!

Next step

• Add jax-based boozer coordinate transformation.
• Fast evaulation of EffetiveRipple, GammaC, maybe slow without gpu.