feat(highcharts): implement shap-summary (#2991)

## Implementation: `shap-summary` - highcharts

Implements the **highcharts** version of `shap-summary`.

**File:** `plots/shap-summary/implementations/highcharts.py`

**Parent Issue:** #2923

---
:robot: *[impl-generate
workflow](https://github.com/MarkusNeusinger/pyplots/actions/runs/20612964334)*

---------

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

Author: github-actions[bot]

plots/shap-summary/implementations/highcharts.py

@@ -0,0 +1,225 @@
+""" pyplots.ai
+shap-summary: SHAP Summary Plot
+Library: highcharts unknown | Python 3.13.11
+Quality: 91/100 | Created: 2025-12-31
+"""
+
+import tempfile
+import time
+import urllib.request
+from pathlib import Path
+
+import numpy as np
+from highcharts_core.chart import Chart
+from highcharts_core.options import HighchartsOptions
+from highcharts_core.options.series.scatter import ScatterSeries
+from selenium import webdriver
+from selenium.webdriver.chrome.options import Options
+
+
+# Data - Simulated SHAP values for a regression model
+np.random.seed(42)
+n_samples = 200
+n_features = 12
+
+feature_names = [
+    "House Size (sqft)",
+    "Bedrooms",
+    "Location Score",
+    "Age (years)",
+    "Bathrooms",
+    "Garage Spaces",
+    "Lot Size (acres)",
+    "School Rating",
+    "Crime Rate",
+    "Distance to City (mi)",
+    "Year Built",
+    "HOA Fee ($)",
+]
+
+# Generate feature values (normalized 0-1 for coloring)
+feature_values = np.random.rand(n_samples, n_features)
+
+# Generate SHAP values with varying importance per feature
+# Features at top have higher magnitude SHAP values
+importance_weights = np.linspace(1.5, 0.2, n_features)
+shap_values = np.zeros((n_samples, n_features))
+
+for i in range(n_features):
+    # Create correlation between feature value and SHAP value
+    base_effect = (feature_values[:, i] - 0.5) * importance_weights[i] * 2
+    noise = np.random.randn(n_samples) * importance_weights[i] * 0.3
+    shap_values[:, i] = base_effect + noise
+
+# Sort features by mean absolute SHAP value (most important first)
+mean_abs_shap = np.mean(np.abs(shap_values), axis=1)
+feature_importance = np.mean(np.abs(shap_values), axis=0)
+sorted_indices = np.argsort(feature_importance)[::-1]
+
+# Take top 10 features
+top_n = 10
+sorted_indices = sorted_indices[:top_n]
+
+# Prepare series data - one series per color bucket for gradient effect
+# Use blue (low) to red (high) color gradient
+n_color_bins = 10
+color_gradient = [
+    "#3B4CC0",  # Blue (low)
+    "#5A7DC7",
+    "#7AAAD0",
+    "#A0C4DE",
+    "#C5D5E8",
+    "#E8C5C5",
+    "#DEA0A0",
+    "#D07A7A",
+    "#C75A5A",
+    "#C03B3B",  # Red (high)
+]
+
+all_series = []
+
+for bin_idx in range(n_color_bins):
+    bin_low = bin_idx / n_color_bins
+    bin_high = (bin_idx + 1) / n_color_bins
+
+    series_data = []
+
+    for feat_idx, sorted_feat_idx in enumerate(sorted_indices):
+        y_pos = top_n - 1 - feat_idx  # Invert so most important is at top
+
+        for sample_idx in range(n_samples):
+            feat_val = feature_values[sample_idx, sorted_feat_idx]
+
+            if bin_low <= feat_val < bin_high or (bin_idx == n_color_bins - 1 and feat_val == 1.0):
+                shap_val = shap_values[sample_idx, sorted_feat_idx]
+                # Add jitter to reduce overlap
+                jitter = np.random.uniform(-0.3, 0.3)
+                series_data.append({"x": round(shap_val, 4), "y": y_pos + jitter})
+
+    if series_data:
+        series = ScatterSeries()
+        series.data = series_data
+        series.name = f"Feature Value: {bin_low:.1f}-{bin_high:.1f}"
+        series.color = color_gradient[bin_idx]
+        series.marker = {"radius": 8, "symbol": "circle"}
+        series.show_in_legend = bin_idx in [0, 4, 9]  # Show only low, mid, high
+        all_series.append(series)
+
+# Create chart
+chart = Chart(container="container")
+chart.options = HighchartsOptions()
+
+# Chart configuration
+chart.options.chart = {
+    "type": "scatter",
+    "width": 4800,
+    "height": 2700,
+    "backgroundColor": "#ffffff",
+    "marginLeft": 350,
+    "marginBottom": 150,
+}
+
+# Title
+chart.options.title = {
+    "text": "shap-summary \u00b7 highcharts \u00b7 pyplots.ai",
+    "style": {"fontSize": "48px", "fontWeight": "bold"},
+}
+
+# Subtitle
+chart.options.subtitle = {"text": "Feature Importance and Impact on Model Predictions", "style": {"fontSize": "28px"}}
+
+# X-axis (SHAP value)
+chart.options.x_axis = {
+    "title": {"text": "SHAP Value (Impact on Prediction)", "style": {"fontSize": "36px"}, "margin": 20},
+    "labels": {"style": {"fontSize": "28px"}},
+    "gridLineWidth": 1,
+    "gridLineColor": "#e0e0e0",
+    "plotLines": [{"value": 0, "color": "#333333", "width": 4, "zIndex": 5}],
+}
+
+# Y-axis (features)
+y_categories = [feature_names[i] for i in sorted_indices][::-1]  # Reverse for top-to-bottom
+chart.options.y_axis = {
+    "title": {"text": "", "style": {"fontSize": "28px"}},
+    "categories": y_categories,
+    "labels": {"style": {"fontSize": "30px"}},
+    "gridLineWidth": 0,
+    "reversed": False,
+}
+
+# Legend configuration for color scale
+chart.options.legend = {
+    "enabled": True,
+    "align": "right",
+    "verticalAlign": "middle",
+    "layout": "vertical",
+    "title": {"text": "Feature Value", "style": {"fontSize": "32px", "fontWeight": "bold"}},
+    "itemStyle": {"fontSize": "26px"},
+    "symbolRadius": 8,
+    "symbolHeight": 20,
+    "symbolWidth": 20,
+    "itemMarginBottom": 10,
+}
+
+# Plot options
+chart.options.plot_options = {
+    "scatter": {
+        "marker": {"radius": 8, "states": {"hover": {"enabled": True, "lineColor": "#333333"}}},
+        "jitter": {"x": 0, "y": 0},
+    },
+    "series": {"animation": False},
+}
+
+# Tooltip
+chart.options.tooltip = {
+    "headerFormat": "<b>{series.name}</b><br>",
+    "pointFormat": "SHAP Value: {point.x:.3f}",
+    "style": {"fontSize": "18px"},
+}
+
+# Add all series
+for s in all_series:
+    chart.add_series(s)
+
+# Download Highcharts JS
+highcharts_url = "https://code.highcharts.com/highcharts.js"
+with urllib.request.urlopen(highcharts_url, timeout=30) as response:
+    highcharts_js = response.read().decode("utf-8")
+
+# Generate HTML with inline scripts
+html_str = chart.to_js_literal()
+html_content = f"""<!DOCTYPE html>
+<html>
+<head>
+    <meta charset="utf-8">
+    <script>{highcharts_js}</script>
+</head>
+<body style="margin:0;">
+    <div id="container" style="width: 4800px; height: 2700px;"></div>
+    <script>{html_str}</script>
+</body>
+</html>"""
+
+# Write temp HTML and take screenshot
+with tempfile.NamedTemporaryFile(mode="w", suffix=".html", delete=False, encoding="utf-8") as f:
+    f.write(html_content)
+    temp_path = f.name
+
+chrome_options = Options()
+chrome_options.add_argument("--headless")
+chrome_options.add_argument("--no-sandbox")
+chrome_options.add_argument("--disable-dev-shm-usage")
+chrome_options.add_argument("--disable-gpu")
+chrome_options.add_argument("--window-size=4800,2700")
+
+driver = webdriver.Chrome(options=chrome_options)
+driver.get(f"file://{temp_path}")
+time.sleep(5)
+driver.save_screenshot("plot.png")
+driver.quit()
+
+Path(temp_path).unlink()
+
+# Also save HTML for interactive version
+with open("plot.html", "w", encoding="utf-8") as f:
+    f.write(html_content)

plots/shap-summary/metadata/highcharts.yaml

@@ -0,0 +1,30 @@
+library: highcharts
+specification_id: shap-summary
+created: '2025-12-31T06:10:37Z'
+updated: '2025-12-31T09:04:07Z'
+generated_by: claude-opus-4-5-20251101
+workflow_run: 20612964334
+issue: 2923
+python_version: 3.13.11
+library_version: unknown
+preview_url: https://storage.googleapis.com/pyplots-images/plots/shap-summary/highcharts/plot.png
+preview_thumb: https://storage.googleapis.com/pyplots-images/plots/shap-summary/highcharts/plot_thumb.png
+preview_html: https://storage.googleapis.com/pyplots-images/plots/shap-summary/highcharts/plot.html
+quality_score: 91
+review:
+  strengths:
+  - Excellent implementation of SHAP summary plot using scatter series with color-coded
+    bins
+  - Well-chosen realistic context (house price prediction) with meaningful feature
+    names
+  - Proper sorting of features by mean absolute SHAP value (most important at top)
+  - Clear visual separation using vertical line at x=0
+  - Good use of jittering to reduce point overlap
+  - Correct title format and professional subtitle
+  - Properly saves both PNG and interactive HTML versions
+  weaknesses:
+  - Blue-red color scheme, while conventional for SHAP, is not optimal for colorblind
+    accessibility (consider viridis or plasma alternatives in future)
+  - All features show similar linear correlation patterns - more diverse SHAP distributions
+    (e.g., some features with non-linear effects, clustered values) would better demonstrate
+    the plot type capabilities