g(r) Pair Distribution Viewer

Single-panel viewer for the 2-body g(r) of one supercell or a regime-ladder comparison of several. The same g2_viewer.html template is used for both; behaviour is driven by whether the embedded JSON payload contains a single profiles array (single-cell mode) or a series list (compare mode).

Single-cell mode¶

# write a standalone HTML file
cell.export_g2_html("mro_g2.html", r_max=8.0, title="Si MRO g(r)")

# or view inline in Jupyter (returns IPython.display.HTML wrapping an iframe)
cell.plot_g2(r_max=8.0, title="Si MRO g(r)")

Controls:

• pair dropdown: picks which species pair is plotted. Binary materials (SiO₂) show Si-Si, Si-O, O-O; ternaries (SrTiO₃) show all six unique pairs; single-element cases show one.

• overlay all pairs checkbox: draws every pair on the same axis with an inline legend (useful for quickly finding which pair peaks share radii).

A faint dashed vertical marker at the reference pair_peak from shell_target (if available) helps locate the expected first-shell radius on the current curve.

Compare mode¶

cells = {
    "liquid":          cells_liquid,
    "amorphous":       cells_amorphous,
    "SRO":             cells_sro,
    "MRO":             cells_mro,
    "ext. MRO":        cells_ext_mro,
    "nanocrystalline": cells_nano,
}

# inline Jupyter display
am.plot_g2_compare(cells, r_max=8.0, title="Silicon regime comparison")

# standalone HTML for docs / sharing
am.export_g2_compare_html(cells, "si_g2_compare.html", r_max=8.0)

Behaviour:

• Stacked y-offset. Each series sits in its own horizontal slot with a small gap (offset = 35 % of the per-curve peak height). The input order is preserved (passing a dict iterates insertion order), so the regen scripts’ liquid → nanocrystalline order puts the most disordered curve at the bottom and the most ordered at the top. Numeric y-axis ticks are hidden in this mode because a single number no longer maps to one g(r) value.

• Inline labels. Each series label is drawn in its curve’s colour at the near-origin anchor (r ≈ 0.3 Å), just above the curve. No legend box; the labels double as the legend.

• Per-series reference line. A faint dashed g = 1 line is drawn at every series’s baseline so you can read off the over- / under-density of each regime at a glance.

Input shapes¶

atomode.export_g2_compare_html accepts any of:

# dict (keys become labels; insertion order = stack order)
am.plot_g2_compare({"liquid": c1, "MRO": c2, "NC": c3})

# list of (cell, label) tuples
am.plot_g2_compare([(c1, "liquid"), (c2, "MRO")])

# list of cells (uses each cell.label)
am.plot_g2_compare([c1, c2, c3])

All supplied cells must share the same species set. Measurements reuse the g3 machinery under the hood (phi_num_bins is set low for speed because the compare viewer doesn’t plot angular data).

Example: silicon regime ladder¶

The SiO₂ example in this repo opens with exactly this compare viewer embedded under its Overview heading; the silicon analogue is visible on the silicon example page. Pair dropdown switches which species pair is shown; the six stacked curves identify each regime by colour and inline label.