Full-sky volume visualizations for FastEMRIWaveforms (FEW) EMRI waveforms.

Fewview turns a FEW waveform into a spherical, retarded-time picture of the gravitational wave. The fully relativistic FastKerrEccentricEquatorialFlux modes are reconstructed into h_lm(t), combined with spin-weighted spherical harmonics over the whole sky, and evaluated at u = t_frame - r / c throughout a sphere, so a still or movie shows the wave propagating outward from the source.

A few stills produced with Fewview, across colour palettes and camera views.

magma, oblique	rainbow, oblique
cool, face-on	plasma, oblique
plasma, face-on	magma, face-on

Fewview reads FEW models directly, so install it into the same environment as your existing FastEMRIWaveforms installation:

pip install -e .

few is expected to already be importable; it is not installed as a dependency so that Fewview does not disturb a working FEW build.

• Tutorial — render a still and a movie, step by step (executable version: examples/tutorial.ipynb)
• Cluster rendering — distributed movies on Slurm

import numpy as np
from fewview import (
    generate_relativistic_mode_waveform,
    choose_max_delay,
    render_mode_frame,
)

wf = generate_relativistic_mode_waveform(
    M=1e6, mu=10.0, a=0.9, p0=12.0, e0=0.4, xI0=1.0,
    dt=10.0, T=0.01, lmax=10, nmax=55,
)
ref = wf.strain(theta=np.pi / 3.0, phi=0.0)
max_delay = choose_max_delay(wf.time, np.real(ref), -np.imag(ref))

render_mode_frame(
    wf,
    screenshot="emri.png",
    max_delay=max_delay,
    frame_time=wf.time[-1],
    waveform_start_time=wf.time[0],
    waveform_end_time=wf.time[-1],
    component="plus",
    opacity_profile="shells",
    color_scheme="rainbow",
)

For a movie, use render_mode_animation with the same arguments plus a start_time/end_time range and frames/fps.

Three knobs control the look.

Component — what scalar field is rendered:

Opacity profile — how values map to transparency: soft (broad fronts, the default), bands (symmetric levels), shells (nested translucent signed-strain sheets; needs plus/cross), flux (log-compressed energy flux; needs energy_flux).

Colour scheme — any Matplotlib colormap (magma, viridis, plasma, inferno, cividis, cool, blues) plus three tuned palettes: rainbow, aurora, cinematic. fewview.colormaps.available_color_schemes() lists them all.

The public API is re-exported from the top-level fewview package and grouped into topical submodules:

• fewview.waveform — generate_relativistic_mode_waveform, RelativisticModeWaveform, estimate_waveform_period, choose_max_delay, polarizations_from_complex
• fewview.volume — build_mode_retarded_time_volume, RetardedTimeVolume, to_pyvista, save_volume
• fewview.rendering — render_volume, render_mode_frame, render_mode_animation
• fewview.surface — StrainSurface, build_strain_surface, plot_strain_surface
• fewview.plotting — plot_volume_slice
• fewview.colormaps — palette and profile listings

Installing the package provides console scripts:

See docs/cluster_rendering.md for the cluster workflow.

• The polar axis carries an unavoidable seam: a spin-2 field cannot be combed continuously over the poles, so any scalar rendering of plus/cross is multivalued there. It is faint in oblique views and washes out under smoothing, but it is inherent to the representation, not a bug.
• The amplitude model is capped at nmax=55. Requesting a larger nmax emits a warning and uses the model's trained set; at high eccentricity the render will therefore be missing some high-overtone content.

Fewview's volume-rendering approach is inspired by two codes that pioneered this style of visualization for numerical relativity simulations:

• gwpv by Nils Vu, a ParaView-based renderer for gravitational waves from numerical relativity.
• ROSE, which renders the radiated field of numerical relativity simulations as an outgoing volume.

Fewview adapts that idea to the mode-resolved EMRI waveforms produced by FastEMRIWaveforms.

MIT. See LICENSE.