File size: 6,516 Bytes
6fe482b | 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 | """
Interactive Plotly figures (returned to gr.Plot). Dark clinical styling to match
the EyeQC theme; hover, zoom, rotate and animation where it aids insight.
"""
from __future__ import annotations
import numpy as np
import plotly.graph_objects as go
from plotly.subplots import make_subplots
INK = "#0c1418"; PAPER = "rgba(0,0,0,0)"; TEAL = "#2aa5b8"; AMBER = "#e0b23c"
CORAL = "#e05252"; GREEN = "#28c07f"; GRID = "rgba(255,255,255,0.08)"
FONT = "Inter, system-ui, sans-serif"
def fig_to_iframe(fig, height=480):
"""Render a Plotly figure as a self-contained iframe so Play/Reset work."""
import html as _html
raw = fig.to_html(include_plotlyjs="cdn", full_html=True,
config={"displayModeBar": False, "responsive": True})
doc = _html.escape(raw, quote=True)
return (f'<iframe srcdoc="{doc}" loading="lazy" '
f'style="width:100%;height:{height}px;border:none;border-radius:12px;'
f'background:transparent;"></iframe>')
def _style(fig, h=430, legend=True):
fig.update_layout(
template="plotly_dark", paper_bgcolor=PAPER, plot_bgcolor=PAPER,
font=dict(family=FONT, color="#d7e2e6", size=12),
margin=dict(l=48, r=24, t=48, b=44), height=h,
showlegend=legend, legend=dict(bgcolor="rgba(0,0,0,0)"),
)
fig.update_xaxes(gridcolor=GRID, zeroline=False)
fig.update_yaxes(gridcolor=GRID, zeroline=False)
return fig
# ---------------------------------------------------------------- DSP curves
def dsp_figure(dsp):
"""Degradation-sensitivity curves: disease confidence & QC vs severity."""
kinds = list(dsp["curves"].keys())
fig = make_subplots(rows=1, cols=len(kinds), shared_yaxes=True,
subplot_titles=[k.capitalize() for k in kinds])
for c, k in enumerate(kinds, 1):
cur = dsp["curves"][k]
sev = cur["severity"]
fig.add_trace(go.Scatter(x=sev, y=cur["disease_prob"], name="disease p",
line=dict(color=CORAL, width=3), mode="lines+markers",
legendgroup="d", showlegend=(c == 1)), row=1, col=c)
fig.add_trace(go.Scatter(x=sev, y=[q/100 for q in cur["qc"]], name="QC quality",
line=dict(color=TEAL, width=3, dash="dot"), mode="lines+markers",
legendgroup="q", showlegend=(c == 1)), row=1, col=c)
fig.add_trace(go.Scatter(x=sev, y=cur["ungradable"], name="ungradable p",
line=dict(color=AMBER, width=2), mode="lines",
legendgroup="u", showlegend=(c == 1)), row=1, col=c)
fig.update_xaxes(title_text="severity", row=1, col=c)
fig.update_yaxes(title_text="probability / quality", range=[0, 1], row=1, col=1)
ttl = (f"Degradation Sensitivity - {dsp['top_disease']} | "
f"entanglement {dsp['entanglement_index']:.2f}")
fig.update_layout(title=dict(text=ttl, font=dict(size=14)))
return _style(fig, h=420)
def entanglement_dial(entanglement):
"""Radial gauge for the entanglement index."""
val = float(entanglement) * 100
color = CORAL if val > 40 else GREEN
fig = go.Figure(go.Indicator(
mode="gauge+number", value=val, number=dict(suffix="%", font=dict(size=34)),
title=dict(text="Entanglement index", font=dict(size=14)),
gauge=dict(axis=dict(range=[0, 100], tickcolor="#8aa"),
bar=dict(color=color, thickness=0.32),
steps=[dict(range=[0, 40], color="rgba(40,192,127,0.18)"),
dict(range=[40, 100], color="rgba(224,82,82,0.18)")],
bordercolor="rgba(0,0,0,0)")))
return _style(fig, h=300, legend=False)
# ---------------------------------------------------------- embedding explorer
def embedding_scatter(emb, labels, title="Embedding", dims=2):
labels = np.asarray(labels).astype(str)
fig = go.Figure()
palette = [TEAL, CORAL, AMBER, GREEN, "#9d7bd8", "#e08a3c", "#4db6ac"]
for i, g in enumerate(sorted(set(labels))):
sel = labels == g
col = palette[i % len(palette)]
if dims == 3 and emb.shape[1] >= 3:
fig.add_trace(go.Scatter3d(x=emb[sel, 0], y=emb[sel, 1], z=emb[sel, 2],
mode="markers", name=g,
marker=dict(size=5, color=col, opacity=0.9,
line=dict(width=0.5, color="#fff"))))
else:
fig.add_trace(go.Scatter(x=emb[sel, 0], y=emb[sel, 1], mode="markers",
name=g, marker=dict(size=11, color=col, opacity=0.9,
line=dict(width=1, color="#fff"))))
fig.update_layout(title=dict(text=title, font=dict(size=14)))
return _style(fig, h=460)
def animated_correction(emb_before, emb_after, batches, frames=24):
"""Animate points morphing from pre-correction to post-correction positions."""
b = np.asarray(batches).astype(str)
palette = [TEAL, CORAL, AMBER, GREEN, "#9d7bd8", "#e08a3c", "#4db6ac"]
uniq = sorted(set(b))
cmap = {g: palette[i % len(palette)] for i, g in enumerate(uniq)}
colors = [cmap[x] for x in b]
def frame_data(t):
p = (1 - t) * emb_before + t * emb_after
return go.Scatter(x=p[:, 0], y=p[:, 1], mode="markers",
marker=dict(size=11, color=colors, opacity=0.9,
line=dict(width=1, color="#fff")),
showlegend=False)
ts = np.linspace(0, 1, frames)
fig = go.Figure(
data=[frame_data(0)],
frames=[go.Frame(data=[frame_data(t)], name=f"{i}")
for i, t in enumerate(ts)])
# legend proxies
for g in uniq:
fig.add_trace(go.Scatter(x=[None], y=[None], mode="markers", name=g,
marker=dict(size=11, color=cmap[g])))
fig.update_layout(
title=dict(text="Batch harmonisation (press play)", font=dict(size=14)),
updatemenus=[dict(type="buttons", showactive=False, x=0.02, y=1.12,
buttons=[dict(label="▶ Play", method="animate",
args=[None, dict(frame=dict(duration=60, redraw=True),
fromcurrent=True, transition=dict(duration=0))]),
dict(label="⏮ Reset", method="animate",
args=[["0"], dict(frame=dict(duration=0, redraw=True),
mode="immediate")])])])
return _style(fig, h=460, legend=True)
|