File size: 10,345 Bytes
d2486ac | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 | """
Shared matplotlib styling primitives: palette, axis styling, and the
time-series plot helpers used across the panel builders.
Split out of the former monolithic dashboard_core.py (Phase 1 of the
dashboard refactor). _series_plot/_series_enhanced/_energy_enhanced's
former redundancy was unified into a single _series_plot (Phase 2) --
same visual output, one parameterized function instead of three.
"""
import numpy as np
import pandas as pd
import matplotlib.pyplot as plt
import matplotlib.patches as mpatches
from matplotlib.colors import LinearSegmentedColormap
C = dict(
bg = '#03050e',
panel = '#070b18',
grid = '#0f1628',
border = '#18243c',
title = '#e2eaf8',
label = '#4a6080',
tick = '#2e4a6a',
prob = '#38bdf8',
energy = '#00e0ff',
entropy = '#f43f7a',
purity = '#4ade80',
grad = '#fbbf24',
noise = '#fb923c',
theta = '#a78bfa',
dom = '#ffffff',
accent = '#00ff9d',
warn = '#ff6b35',
)
MONO = {'fontfamily': 'monospace'}
FILL_A = 0.08
_cmap_prob = LinearSegmentedColormap.from_list(
'tureq_prob', ['#0a1f3d', '#38bdf8', '#f43f7a'], N=256
)
BG_ART = '#030305'
PANEL_ART = '#07070a'
CYAN_N = '#00f3ff'
PINK_N = '#ff0055'
PURP_N = '#7a00ff'
GOLD_N = '#ffaa00'
# Repeated across build_panel_fisica/vqe_results/md_results/performance's
# plt.figure(facecolor=...) -- distinct from C['bg'] (a different literal
# in the original), centralized here (Phase 2) as one named constant
# instead of 5 duplicated string literals in panels.py.
FIG_BG_DARK = '#010409'
def _ax_style(ax, title='', xlabel='', ylabel='', spine_alpha=0.6):
ax.set_facecolor(C['panel'])
for sp in ax.spines.values():
sp.set_edgecolor(C['border'])
sp.set_linewidth(0.7)
sp.set_alpha(spine_alpha)
ax.tick_params(colors=C['tick'], which='both', length=3, width=0.5, labelsize=7.5)
ax.xaxis.label.set_color(C['label']); ax.xaxis.label.set_fontsize(8)
ax.yaxis.label.set_color(C['label']); ax.yaxis.label.set_fontsize(8)
ax.grid(True, ls='--', lw=0.30, alpha=0.25, color=C['grid'])
if title:
ax.set_title(title, color=C['title'], fontsize=9.5, fontweight='bold',
pad=6, loc='left', **MONO)
if xlabel: ax.set_xlabel(xlabel, **MONO)
if ylabel: ax.set_ylabel(ylabel, **MONO)
def _fill(ax, x, y, col):
ax.fill_between(x, y, alpha=FILL_A, color=col)
def _rolling(y, win):
return pd.Series(y).rolling(win, center=True, min_periods=1).mean().values
def _badge(ax, text, color):
"""Top-right value badge."""
ax.text(0.98, 0.97, text,
transform=ax.transAxes, ha='right', va='top',
fontsize=7.5, color=color, **MONO,
bbox=dict(boxstyle='round,pad=0.25',
facecolor=C['bg'], edgecolor=C['border'], alpha=0.72))
def _series_plot(ax, df, col, color, title, xlabel, ylabel, *,
show_raw=True, badge=False, badge_fmt='{:.4g}',
ref_line=None, detect_plateau=False,
mark_convergence=False):
"""Unified time-series: raw line + fill + rolling mean + last-value
annotation, with optional badge / reference-line / barren-plateau-span /
convergence-marker overlays.
Unifies what used to be three near-duplicate functions (_series_plot,
which this absorbs, plus _series_enhanced and _energy_enhanced, Phase 2
of the dashboard refactor) into one parameterized function -- same
visual output as each of the three original call shapes:
- old _series_plot(...) == _series_plot(...) (defaults)
- old _series_enhanced(..., ref_line=X) == _series_plot(..., badge=True, ref_line=X)
- old _energy_enhanced(ax, df) == _series_plot(ax, df, 'VQE_Energy',
C['energy'], 'VQE Energy', 'Epoch',
'E (Ha)', badge=True,
badge_fmt='Eₙ={:.4g} Ha',
mark_convergence=True)
"""
_ax_style(ax, title, xlabel, ylabel)
if df.empty or col not in df.columns:
ax.axis('off')
ax.text(0.5, 0.5, f'[ {col} — no data ]',
ha='center', va='center', color=C['label'],
fontsize=8.5, transform=ax.transAxes, **MONO)
return
x = df.index.values
y = df[col].values
if show_raw:
ax.plot(x, y, color=color, lw=1.4, alpha=0.7)
_fill(ax, x, y, color)
win = max(3, len(y) // 15)
rm = _rolling(y, win)
ax.plot(x, rm, color=C['title'], lw=1.1, alpha=0.6, ls='--',
label='rolling mean')
ax.annotate(f'{y[-1]:.4g}',
xy=(x[-1], y[-1]), xytext=(-4, 6),
textcoords='offset points',
color=color, fontsize=7.5, fontweight='bold',
ha='right', **MONO)
x_f = df.index.to_numpy(dtype=float)
if badge:
_badge(ax, badge_fmt.format(y[-1]), color)
if ref_line is not None:
ax.axhline(ref_line, color=color, lw=0.6, ls='--', alpha=0.28)
if detect_plateau:
_thr = 0.01 * np.abs(y).max() if np.abs(y).max() > 0 else 1e-9
_bp = np.abs(y) < _thr
if _bp.sum() > 3:
_edges = np.where(np.diff(_bp.astype(int)))[0]
if len(_edges) >= 2:
ax.axvspan(x_f[_edges[0]], x_f[_edges[1]],
alpha=0.11, color=C['warn'])
ax.text(
(x_f[_edges[0]] + x_f[_edges[1]]) / 2,
y.max() * 0.88,
'plateau', ha='center', fontsize=6.5,
color=C['warn'], alpha=0.80, **MONO,
)
if mark_convergence and len(y) > 2:
_conv = int(np.argmin(np.gradient(y)))
ax.axvline(x_f[_conv], color=C['warn'], lw=0.8, ls=':', alpha=0.55)
ax.annotate(
f'∇min@{_conv}',
xy=(x_f[_conv], y[_conv]),
xytext=(6, -14), textcoords='offset points',
color=C['warn'], fontsize=7.0,
arrowprops=dict(arrowstyle='-', color=C['warn'], lw=0.5),
**MONO,
)
def _interp_colour(c1_hex, c2_hex, t):
t = float(np.clip(t, 0, 1))
def h(s): return [int(s.lstrip('#')[i:i+2], 16) / 255 for i in (0, 2, 4)]
r1, g1, b1 = h(c1_hex)
r2, g2, b2 = h(c2_hex)
to_hex = lambda v: f'{int(v*255):02x}'
return f'#{to_hex(r1+(r2-r1)*t)}{to_hex(g1+(g2-g1)*t)}{to_hex(b1+(b2-b1)*t)}'
def _noise_profile_plot(ax, noise_model, noise_p, n_qubits,
prob_ideal, prob_noisy=None):
_ax_style(ax, 'Noise Analysis', '', '')
ax.set_xlim(0, 1); ax.set_ylim(-0.08, 1.08)
ax.axis('off')
model_col = C['warn'] if noise_model != 'ideal' else C['accent']
ax.text(0.5, 0.97, noise_model.upper().replace('_', ' '),
ha='center', va='top', color=model_col,
fontsize=16, fontweight='bold', transform=ax.transAxes, **MONO)
bar_y = 0.80
ax.add_patch(mpatches.FancyBboxPatch(
(0.05, bar_y - 0.025), 0.90, 0.05,
boxstyle='round,pad=0.005',
facecolor=C['grid'], edgecolor=C['border'], lw=0.8,
transform=ax.transAxes, zorder=2))
fill_w = 0.90 * min(noise_p / 0.10, 1.0)
if fill_w > 0:
fill_col = _interp_colour('#00ff9d', '#ff6b35', noise_p / 0.10)
ax.add_patch(mpatches.FancyBboxPatch(
(0.05, bar_y - 0.025), fill_w, 0.05,
boxstyle='round,pad=0.005',
facecolor=fill_col, edgecolor='none',
transform=ax.transAxes, zorder=3))
ax.text(0.5, bar_y + 0.06, f'p = {noise_p:.4f}',
ha='center', va='bottom', color=C['title'],
fontsize=11, fontweight='bold', transform=ax.transAxes, **MONO)
ax.text(0.05, bar_y - 0.07, '0', ha='left', va='top',
color=C['label'], fontsize=7.5, transform=ax.transAxes, **MONO)
ax.text(0.95, bar_y - 0.07, '0.10', ha='right', va='top',
color=C['label'], fontsize=7.5, transform=ax.transAxes, **MONO)
if prob_noisy is not None and noise_model != 'ideal':
fid_bc = float(np.sum(np.sqrt(np.maximum(prob_ideal, 0) *
np.maximum(prob_noisy, 0))))
tvd = float(0.5 * np.sum(np.abs(prob_ideal - prob_noisy)))
rows = [
('Bhattacharyya Fidelity', f'{fid_bc:.6f}', C['purity']),
('Total Variation Distance', f'{tvd:.6f}', C['entropy']),
('Noise Channel', noise_model.replace('_', ' '), C['noise']),
]
else:
rows = [
('Channel', 'ideal — no noise applied', C['accent']),
('Fidelity', '1.000000', C['purity']),
('TVD', '0.000000', C['label']),
]
for j, (k, v, col) in enumerate(rows):
yy = 0.58 - j * 0.14
ax.text(0.03, yy, k, ha='left', va='center', color=C['label'],
fontsize=8, transform=ax.transAxes, **MONO)
ax.text(0.97, yy, v, ha='right', va='center', color=col,
fontsize=8.5, fontweight='bold',
transform=ax.transAxes, **MONO)
ax.axhline(yy - 0.05, xmin=0.01, xmax=0.99, color=C['grid'], lw=0.35)
if noise_model in ('ideal', 'depolarizing'):
ins = ax.inset_axes([0.04, 0.04, 0.92, 0.26])
ins.set_facecolor(C['panel'])
for sp in ins.spines.values():
sp.set_edgecolor(C['border']); sp.set_linewidth(0.5)
ps = np.linspace(0, 0.1, 200)
d = 2 ** n_qubits
fid_curve = ((1.0 - ps * (d - 1) / d) ** n_qubits).clip(0, 1)
ins.plot(ps, fid_curve, color=C['purity'], lw=1.2)
ins.axvline(noise_p, color=C['warn'], lw=0.9, ls='--', alpha=0.8)
ins.fill_between(ps, fid_curve, alpha=0.07, color=C['purity'])
ins.set_xlim(0, 0.10); ins.set_ylim(0, 1.05)
ins.tick_params(colors=C['tick'], labelsize=6.5, length=2)
ins.set_xlabel('p', color=C['label'], fontsize=7, **MONO)
ins.set_ylabel('F(p)', color=C['label'], fontsize=7, **MONO)
ins.set_title('Theoretical Fidelity Curve',
color=C['label'], fontsize=7, pad=2, **MONO)
ins.grid(True, ls=':', lw=0.3, alpha=0.25, color=C['grid'])
|