sphWarpPlotting 0.4.0

GPU-accelerated SPH neighbor search and operators built with NVIDIA Warp and PyTorch

sphWarpPlotting

sphWarpPlotting is a visualization layer for SPH particle data built on top of sphWarpCore. It is designed to make SPH field inspection easy while keeping plotting configurable and update-friendly.

The library centers around two functions:

• visualizeParticlesNew(...): build a plot from particle state, quantity, and plotting options.
• updatePlot(...): update an existing visualization state with new quantities and/or styling.

This repository also includes a full notebook demo in demo.ipynb and generated outputs in output/.

What This Library Does

• Visualizes particle scalar fields as scatter plots.
• Visualizes scalar fields on a regular grid (interpolated from particles).
• Applies SPH operations (for example gradients) during plotting.
• Maps vector/tensor quantities to scalars for coloring (x, y, L2, etc.).
• Overlays streamlines for vector fields.
• Supports stateful updates of an existing figure for interactive workflows.

Installation

From this repository:

pip install -e .

Notebook extras:

pip install -e .[notebooks]

Core API

from warpPlot import (
    visualizeParticlesNew,
    updatePlot,
    PlottingOptions,
    GridVisualization,
    Mapping,
    PlotScaling,
    ColorMap,
    StreamLineLocation,
)

Main Inputs

• particleState: sphWarpCore.ParticleState
• domain: sphWarpCore.DomainDescription
• quantity: tensor to visualize (scalar or vector; can be mapped)
• options: PlottingOptions(...) for rendering and operation behavior

Main Output

visualizeParticlesNew returns a VisualizationState object containing figure/axis handles, filtered particle states, and artist handles used by updatePlot.

Minimal Example

import matplotlib.pyplot as plt
from warpPlot import visualizeParticlesNew, PlottingOptions, ColorMap

fig, ax = plt.subplots(1, 1, figsize=(6, 5))

state = visualizeParticlesNew(
    fig,
    ax,
    particleState=particleState,
    domain=domain,
    quantity=f_smoothed,
    options=PlottingOptions(
        colorMap=ColorMap.rocket,
        markerSize=4,
        plotTitle="Smoothed Quantity",
    ),
)

fig.tight_layout()

Stateful Update Example

from warpPlot import updatePlot, ColorMap

updatePlot(
    state,
    particles=particleState,
    domain=domain,
    quantity=-f_smoothed,
    colorMap=ColorMap.viridis,
)

Demo Gallery (from demo.ipynb)

1) Linear Gradient Validation

Computes a known linear field gradient with WarpSPH and compares against analytical reference.

2) WarpSPH vs DiffSPH Gradient Comparison

Visual side-by-side comparison of gradient components and their differences.

3) Scatter vs Grid Visualization

Shows the same smoothed scalar field as direct particle scatter and as a grid-mapped field.

4) Updating an Existing Plot State

Demonstrates stateful update of the same visualization object with new quantity/style settings.

5) On-the-fly Plotting Operation + Mapping

Computes gradient during plotting and maps to component views (Mapping.x, Mapping.y).

6) Grid + Streamlines

Shows mapped magnitude coloring with streamline overlay, sampled before mapping.

7) Streamline Figure Update

Stateful update of the streamline visualization.

Typical Workflow

1. Build or load particleState and domain from your SPH pipeline.
2. Choose a quantity (scalar or vector) to inspect.
3. Configure PlottingOptions:
   • colormap/scaling
   • optional plottingOperation
   • optional mapping
   • optional GridVisualization and streamlines
4. Call visualizeParticlesNew(...).
5. Reuse returned state with updatePlot(...) for iterative inspection.

Notes

• mapping is essential when visualizing vector quantities as a color field.
• Grid visualization is often easier to read for dense particle sets.
• updatePlot is useful for simulation loops where geometry stays fixed but values change.

Running the Demo

Open and run all cells in demo.ipynb. The notebook saves figures into output/.