gwnr.graph — visualization

Plotting utilities for posteriors, model-comparison studies and simulation data. Source: gwnr/graph/.

1. corner.py — corner plots
   1. class CornerPlot(MultiDDistribution)
   2. cornerplot_dataframe(df, cols=None)
2. misc.py — general plotting helpers
3. cbc.py
4. analysis_products.py — analysis-result figures
5. paraview.py — ParaView helpers
6. visualization.py — notebook embedding

---

corner.py — corner plots

class CornerPlot(MultiDDistribution)

Corner-plot generator built on the MultiDDistribution container: pass ≥2-dimensional data (columns = parameters) and call draw(...).

draw(params_plot, ...) is highly configurable; the main option groups are:

• Panel type: plot_type='scatter' or 'contour'; 2-D panels can show data points or percentile contours (contour_levels=[68.27, 90.0, 95.45], line styles, inline labels, optionally return enclosed areas via return_areas_in_contours)
• Truth values and priors: true_params_vals, params_oned_priors overlay injected values and 1-D priors
• Point coloring: param_color colors scatter points by a third parameter, with cmap, color_min/color_max, and a colorbar label
• 1-D marginals: histogram type/bins, median line (show_oned_median), percentile band (show_oned_percentiles=90.0), labeling and placement
• Layout: existing fig/axes_array reuse, axis limits (plim_low/plim_high), fonts, legends, grid, figure title

cornerplot_dataframe(df, cols=None)

One-liner corner plot from a pandas.DataFrame (defaults to all columns).

---

misc.py — general plotting helpers

• set_matplotlib_params() — house style for matplotlib.
• make_filled_contour_plot(x, y, z, ...) — filled contours from scattered (x, y, z) data, with a choice of interpolators (Rbf, griddata, SmoothBivariateSpline) and per-interpolator defaults.
• make_2Dplot_errorbars(Xs, Ys, Xerrs, Yerrs, ...) — multi-series 2-D plots with error bars.
• make_scatter_plot, make_scatter_plot3D, make_scatter_plot3D_mult, make_scatter_plot3D_multrow — color-mapped scatter plots in 2-D/3-D, singly or in multi-panel grids (used heavily for effectualness-vs-parameters figures).
• make_contour_plot_multrow, make_contourf_mult, make_parameters_plot — multi-panel contour figures.

cbc.py

ParamLatexLabels() — dictionary of LaTeX axis labels for standard compact-binary parameters (masses, spins, distance, angles, …), shared across plotting code.

---

analysis_products.py — analysis-result figures

Classes that read analysis outputs (HDF5 match/mismatch files) and produce publication figures:

• class plot_mismatches_sim — per-simulation NR error plots: mismatches between CCE extraction radii, resolutions (Levs), and extrapolation orders (plot_cce_mismatches_all, plot_cce_extrapolation_mismatches, plot_cce_max_mismatch, …).
• class plot_mismatches_sims — population-level versions across a catalog, to correlate NR errors with binary parameters (hist_cce_mismatch, …).
• class plot_effectualness_vs_totalmass — the full suite of effectualness / fitting-factor and parameter-bias figures versus total mass and intrinsic parameters, including multi-approximant panel grids, contour versions, and recovered-parameter plots (plot_effectualness_vs_totalmass, plot_parameterbiases_vs_parameters_multrow, plot_effectualness_contours_vs_parameters, …). Reads data written by the NR EffectualnessAndBias store.

---

paraview.py — ParaView helpers

class ParsePVD — edit ParaView .pvd collection files: RetrieveUniqueTimeSteps(), DownsampleTimeSteps(factor), RemoveTimeSteps(low, high), WriteFile(filename). Useful for trimming heavy 3-D visualization datasets from NR runs.

visualization.py — notebook embedding

play_movie(m) and embed_video(fname, mimetype) — display movies (e.g. SpECTRE field evolutions) inline in Jupyter notebooks.