feat(pygal): implement calibration-curve (#2354)

## Implementation: `calibration-curve` - pygal

Implements the **pygal** version of `calibration-curve`.

**File:** `plots/calibration-curve/implementations/pygal.py`

---
:robot: *[impl-generate
workflow](https://github.com/MarkusNeusinger/pyplots/actions/runs/20528205173)*

---------

Co-authored-by: github-actions[bot] <github-actions[bot]@users.noreply.github.com>
Co-authored-by: Claude Opus 4.5 <noreply@anthropic.com>
Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>

Author: github-actions[bot]

plots/calibration-curve/implementations/pygal.py

@@ -0,0 +1,149 @@
+""" pyplots.ai
+calibration-curve: Calibration Curve
+Library: pygal 3.1.0 | Python 3.13.11
+Quality: 91/100 | Created: 2025-12-26
+"""
+
+import numpy as np
+import pygal
+from pygal.style import Style
+
+
+# Data: Generate synthetic binary classification with realistic calibration
+np.random.seed(42)
+n_samples = 2000
+n_bins = 10
+
+# Generate true probabilities spread across 0-1 range
+true_prob = np.random.beta(2, 2, n_samples)
+y_true = (np.random.random(n_samples) < true_prob).astype(int)
+
+# Model 1: Well-calibrated model (Logistic Regression style)
+noise1 = np.random.randn(n_samples) * 0.08
+y_prob_model1 = np.clip(true_prob + noise1, 0.01, 0.99)
+
+# Model 2: Overconfident model (Random Forest / Neural Network style)
+# More extreme S-curve to make miscalibration pattern more visible
+y_prob_model2 = 1 / (1 + np.exp(-12 * (true_prob - 0.5)))
+y_prob_model2 = np.clip(y_prob_model2 + np.random.randn(n_samples) * 0.02, 0.02, 0.98)
+
+# Compute calibration data inline (KISS principle - no helper functions)
+bin_edges = np.linspace(0, 1, n_bins + 1)
+
+# Model 1 calibration
+bin_indices1 = np.digitize(y_prob_model1, bin_edges[1:-1])
+mean_pred1 = []
+frac_pos1 = []
+for i in range(n_bins):
+    mask = bin_indices1 == i
+    if mask.sum() > 0:
+        mean_pred1.append(np.mean(y_prob_model1[mask]))
+        frac_pos1.append(np.mean(y_true[mask]))
+
+# Model 2 calibration
+bin_indices2 = np.digitize(y_prob_model2, bin_edges[1:-1])
+mean_pred2 = []
+frac_pos2 = []
+for i in range(n_bins):
+    mask = bin_indices2 == i
+    if mask.sum() > 0:
+        mean_pred2.append(np.mean(y_prob_model2[mask]))
+        frac_pos2.append(np.mean(y_true[mask]))
+
+# Compute Brier scores inline
+brier1 = np.mean((y_prob_model1 - y_true) ** 2)
+brier2 = np.mean((y_prob_model2 - y_true) ** 2)
+
+# Custom style for 4800 x 2700 canvas with high-contrast colors
+custom_style = Style(
+    background="white",
+    plot_background="white",
+    foreground="#333333",
+    foreground_strong="#333333",
+    foreground_subtle="#CCCCCC",  # Lighter subtle color for more subtle guides
+    colors=("#888888", "#2E7D32", "#C62828"),  # Gray, dark green (high contrast), dark red
+    title_font_size=72,
+    label_font_size=48,
+    major_label_font_size=42,
+    legend_font_size=42,
+    value_font_size=36,
+    tooltip_font_size=36,
+    stroke_width=6,
+    opacity=0.9,
+    opacity_hover=1.0,
+    guide_stroke_color="#E0E0E0",  # Very light guide lines for subtlety
+    guide_stroke_dasharray="3,3",  # Subtle dashed pattern
+)
+
+# Create XY chart for calibration curve with legend at bottom
+chart = pygal.XY(
+    style=custom_style,
+    width=4800,
+    height=2700,
+    title="calibration-curve · pygal · pyplots.ai",
+    x_title="Mean Predicted Probability",
+    y_title="Fraction of Positives",
+    show_dots=True,
+    dots_size=14,
+    stroke_style={"width": 5},
+    show_x_guides=True,
+    show_y_guides=True,
+    x_value_formatter=lambda x: f"{x:.1f}",
+    range=(0, 1),
+    xrange=(0, 1),
+    legend_at_bottom=True,
+    legend_at_bottom_columns=3,  # Display legend items horizontally at bottom
+    legend_box_size=28,
+    truncate_legend=-1,
+    margin=50,
+    margin_top=80,
+    margin_bottom=200,  # Extra space for bottom legend
+)
+
+# Extend model curves to boundaries by computing expected values at 0 and 1
+# This creates visual balance with the perfect calibration line
+start_point1 = (0.0, 0.0)  # Well-calibrated model should start near origin
+end_point1 = (1.0, 1.0)  # Well-calibrated model should end near (1,1)
+start_point2 = (0.0, 0.0)  # Overconfident model starts at origin
+end_point2 = (1.0, 1.0)  # Overconfident model ends at (1,1)
+
+# Perfect calibration line (diagonal reference) - first in gray, dashed, no dots
+# Using dict format with 'value' key for pygal's tooltip system
+perfect_calibration = [
+    {"value": (0, 0), "label": "Perfect calibration reference"},
+    {"value": (0.25, 0.25), "label": "Predicted = Observed"},
+    {"value": (0.5, 0.5), "label": "Ideal: 50% predicted → 50% positive"},
+    {"value": (0.75, 0.75), "label": "Predicted = Observed"},
+    {"value": (1.0, 1.0), "label": "Perfect calibration reference"},
+]
+chart.add("Perfect Calibration", perfect_calibration, stroke_dasharray="15,8", dots_size=0, stroke_style={"width": 4})
+
+# Model 1 calibration curve - well-calibrated (dark green for high contrast)
+# Add boundary points for visual balance, then data points with interactive tooltips
+model1_points = [{"value": start_point1, "label": "Curve start (0,0)"}]
+model1_points.extend(
+    [
+        {"value": (pred, obs), "label": f"Bin: {pred:.2f} pred → {obs:.2f} actual ({int(obs * 100)}% positive)"}
+        for pred, obs in zip(mean_pred1, frac_pos1, strict=False)
+    ]
+)
+model1_points.append({"value": end_point1, "label": "Curve end (1,1)"})
+chart.add(f"Logistic Regression (Brier: {brier1:.3f})", model1_points)
+
+# Model 2 calibration curve - overconfident (dark red for contrast)
+# Shows characteristic sigmoid pattern of overconfident models with extended boundaries
+model2_points = [{"value": start_point2, "label": "Curve start (0,0)"}]
+model2_points.extend(
+    [
+        {"value": (pred, obs), "label": f"Bin: {pred:.2f} pred → {obs:.2f} actual (overconfident: Δ={pred - obs:+.2f})"}
+        for pred, obs in zip(mean_pred2, frac_pos2, strict=False)
+    ]
+)
+model2_points.append({"value": end_point2, "label": "Curve end (1,1)"})
+chart.add(f"Overconfident Model (Brier: {brier2:.3f})", model2_points)
+
+# Save as PNG for static preview
+chart.render_to_png("plot.png")
+
+# Save as HTML for interactive tooltips (pygal's distinctive feature)
+chart.render_to_file("plot.html")

plots/calibration-curve/metadata/pygal.yaml

@@ -0,0 +1,24 @@
+library: pygal
+specification_id: calibration-curve
+created: '2025-12-26T19:36:53Z'
+updated: '2025-12-26T20:03:31Z'
+generated_by: claude-opus-4-5-20251101
+workflow_run: 20528205173
+issue: 0
+python_version: 3.13.11
+library_version: 3.1.0
+preview_url: https://storage.googleapis.com/pyplots-images/plots/calibration-curve/pygal/plot.png
+preview_thumb: https://storage.googleapis.com/pyplots-images/plots/calibration-curve/pygal/plot_thumb.png
+preview_html: https://storage.googleapis.com/pyplots-images/plots/calibration-curve/pygal/plot.html
+quality_score: 91
+review:
+  strengths:
+  - Excellent use of pygal interactive tooltip system with descriptive labels for
+    each data point
+  - High-contrast color scheme (dark green, dark red, gray) ensures accessibility
+  - Brier scores displayed in legend provide quantitative calibration metrics
+  - Correct title format following pyplots.ai convention
+  - Both PNG and HTML outputs leverage pygal dual static/interactive capability
+  - Clean KISS-style code structure with inline calibration computation
+  weaknesses:
+  - None significant - previous issues have been addressed