|
| 1 | +""" pyplots.ai |
| 2 | +scatter-regression-lowess: Scatter Plot with LOWESS Regression |
| 3 | +Library: highcharts unknown | Python 3.13.11 |
| 4 | +Quality: 91/100 | Created: 2025-12-30 |
| 5 | +""" |
| 6 | + |
| 7 | +import tempfile |
| 8 | +import time |
| 9 | +import urllib.request |
| 10 | +from pathlib import Path |
| 11 | + |
| 12 | +import numpy as np |
| 13 | +from highcharts_core.chart import Chart |
| 14 | +from highcharts_core.options import HighchartsOptions |
| 15 | +from highcharts_core.options.series.scatter import ScatterSeries |
| 16 | +from highcharts_core.options.series.spline import SplineSeries |
| 17 | +from selenium import webdriver |
| 18 | +from selenium.webdriver.chrome.options import Options |
| 19 | + |
| 20 | + |
| 21 | +def lowess(x, y, frac=0.3): |
| 22 | + """Simple LOWESS implementation using tricube weighting.""" |
| 23 | + n = len(x) |
| 24 | + k = int(np.ceil(frac * n)) # Number of neighbors to use |
| 25 | + y_smooth = np.zeros(n) |
| 26 | + sorted_idx = np.argsort(x) |
| 27 | + x_sorted = x[sorted_idx] |
| 28 | + y_sorted = y[sorted_idx] |
| 29 | + |
| 30 | + for i in range(n): |
| 31 | + # Calculate distances to all points |
| 32 | + distances = np.abs(x_sorted - x_sorted[i]) |
| 33 | + # Find k nearest neighbors |
| 34 | + nearest_idx = np.argsort(distances)[:k] |
| 35 | + # Maximum distance among neighbors |
| 36 | + max_dist = distances[nearest_idx[-1]] |
| 37 | + if max_dist == 0: |
| 38 | + max_dist = 1.0 |
| 39 | + # Tricube weights |
| 40 | + u = distances[nearest_idx] / max_dist |
| 41 | + weights = (1 - u**3) ** 3 |
| 42 | + # Weighted linear regression |
| 43 | + x_local = x_sorted[nearest_idx] |
| 44 | + y_local = y_sorted[nearest_idx] |
| 45 | + # Weighted least squares |
| 46 | + w_sum = np.sum(weights) |
| 47 | + wx_sum = np.sum(weights * x_local) |
| 48 | + wy_sum = np.sum(weights * y_local) |
| 49 | + wxx_sum = np.sum(weights * x_local * x_local) |
| 50 | + wxy_sum = np.sum(weights * x_local * y_local) |
| 51 | + # Solve for slope and intercept |
| 52 | + denom = w_sum * wxx_sum - wx_sum**2 |
| 53 | + if abs(denom) < 1e-10: |
| 54 | + y_smooth[i] = wy_sum / w_sum if w_sum > 0 else y_sorted[i] |
| 55 | + else: |
| 56 | + slope = (w_sum * wxy_sum - wx_sum * wy_sum) / denom |
| 57 | + intercept = (wy_sum - slope * wx_sum) / w_sum |
| 58 | + y_smooth[i] = slope * x_sorted[i] + intercept |
| 59 | + |
| 60 | + # Return in original order |
| 61 | + result = np.zeros(n) |
| 62 | + result[sorted_idx] = y_smooth |
| 63 | + return x, result |
| 64 | + |
| 65 | + |
| 66 | +# Data - Generate complex non-linear relationship |
| 67 | +np.random.seed(42) |
| 68 | +n_points = 200 |
| 69 | +x = np.linspace(0, 10, n_points) |
| 70 | +# Create a complex non-linear pattern: sine wave + quadratic + noise |
| 71 | +y = 3 * np.sin(x * 1.2) + 0.3 * x**2 - 0.5 * x + np.random.normal(0, 1.5, n_points) |
| 72 | + |
| 73 | +# Compute LOWESS regression curve |
| 74 | +x_lowess, y_lowess = lowess(x, y, frac=0.3) |
| 75 | + |
| 76 | +# Create chart with container |
| 77 | +chart = Chart(container="container") |
| 78 | +chart.options = HighchartsOptions() |
| 79 | + |
| 80 | +# Chart configuration |
| 81 | +chart.options.chart = { |
| 82 | + "type": "scatter", |
| 83 | + "width": 4800, |
| 84 | + "height": 2700, |
| 85 | + "backgroundColor": "#ffffff", |
| 86 | + "marginBottom": 250, |
| 87 | + "marginLeft": 150, |
| 88 | + "spacingTop": 60, |
| 89 | + "spacingRight": 100, |
| 90 | +} |
| 91 | + |
| 92 | +# Title |
| 93 | +chart.options.title = { |
| 94 | + "text": "scatter-regression-lowess · highcharts · pyplots.ai", |
| 95 | + "style": {"fontSize": "48px", "fontWeight": "bold"}, |
| 96 | +} |
| 97 | + |
| 98 | +# Subtitle with context |
| 99 | +chart.options.subtitle = {"text": "Non-linear Trend with LOWESS Smoothing (frac=0.3)", "style": {"fontSize": "32px"}} |
| 100 | + |
| 101 | +# X-axis configuration |
| 102 | +chart.options.x_axis = { |
| 103 | + "title": {"text": "X Value", "style": {"fontSize": "36px"}}, |
| 104 | + "labels": {"style": {"fontSize": "28px"}}, |
| 105 | + "gridLineWidth": 1, |
| 106 | + "gridLineColor": "rgba(0, 0, 0, 0.1)", |
| 107 | +} |
| 108 | + |
| 109 | +# Y-axis configuration |
| 110 | +chart.options.y_axis = { |
| 111 | + "title": {"text": "Y Value", "style": {"fontSize": "36px"}}, |
| 112 | + "labels": {"style": {"fontSize": "28px"}}, |
| 113 | + "gridLineWidth": 1, |
| 114 | + "gridLineColor": "rgba(0, 0, 0, 0.1)", |
| 115 | +} |
| 116 | + |
| 117 | +# Legend |
| 118 | +chart.options.legend = { |
| 119 | + "enabled": True, |
| 120 | + "itemStyle": {"fontSize": "28px"}, |
| 121 | + "align": "right", |
| 122 | + "verticalAlign": "top", |
| 123 | + "layout": "vertical", |
| 124 | + "x": -50, |
| 125 | + "y": 100, |
| 126 | +} |
| 127 | + |
| 128 | +# Plot options |
| 129 | +chart.options.plot_options = { |
| 130 | + "scatter": { |
| 131 | + "marker": {"radius": 10, "fillColor": "rgba(48, 105, 152, 0.6)", "lineWidth": 1, "lineColor": "#306998"} |
| 132 | + }, |
| 133 | + "spline": {"marker": {"enabled": False}, "lineWidth": 5}, |
| 134 | +} |
| 135 | + |
| 136 | +# Add scatter series (data points) |
| 137 | +scatter_series = ScatterSeries() |
| 138 | +scatter_series.name = "Data Points" |
| 139 | +scatter_series.data = [[float(xi), float(yi)] for xi, yi in zip(x, y, strict=True)] |
| 140 | +scatter_series.color = "#306998" |
| 141 | +chart.add_series(scatter_series) |
| 142 | + |
| 143 | +# Add LOWESS curve as spline series |
| 144 | +lowess_series = SplineSeries() |
| 145 | +lowess_series.name = "LOWESS Curve" |
| 146 | +lowess_series.data = [[float(xi), float(yi)] for xi, yi in zip(x_lowess, y_lowess, strict=True)] |
| 147 | +lowess_series.color = "#FFD43B" |
| 148 | +lowess_series.marker = {"enabled": False} |
| 149 | +lowess_series.line_width = 6 |
| 150 | +chart.add_series(lowess_series) |
| 151 | + |
| 152 | +# Download Highcharts JS for inline embedding |
| 153 | +highcharts_url = "https://code.highcharts.com/highcharts.js" |
| 154 | +with urllib.request.urlopen(highcharts_url, timeout=30) as response: |
| 155 | + highcharts_js = response.read().decode("utf-8") |
| 156 | + |
| 157 | +# Generate HTML with inline scripts |
| 158 | +html_str = chart.to_js_literal() |
| 159 | +html_content = f"""<!DOCTYPE html> |
| 160 | +<html> |
| 161 | +<head> |
| 162 | + <meta charset="utf-8"> |
| 163 | + <script>{highcharts_js}</script> |
| 164 | +</head> |
| 165 | +<body style="margin:0;"> |
| 166 | + <div id="container" style="width: 4800px; height: 2700px;"></div> |
| 167 | + <script>{html_str}</script> |
| 168 | +</body> |
| 169 | +</html>""" |
| 170 | + |
| 171 | +# Write temp HTML and take screenshot |
| 172 | +with tempfile.NamedTemporaryFile(mode="w", suffix=".html", delete=False, encoding="utf-8") as f: |
| 173 | + f.write(html_content) |
| 174 | + temp_path = f.name |
| 175 | + |
| 176 | +# Also save HTML for interactive version |
| 177 | +with open("plot.html", "w", encoding="utf-8") as f: |
| 178 | + f.write(html_content) |
| 179 | + |
| 180 | +# Configure Chrome for headless screenshot |
| 181 | +chrome_options = Options() |
| 182 | +chrome_options.add_argument("--headless") |
| 183 | +chrome_options.add_argument("--no-sandbox") |
| 184 | +chrome_options.add_argument("--disable-dev-shm-usage") |
| 185 | +chrome_options.add_argument("--disable-gpu") |
| 186 | +chrome_options.add_argument("--window-size=4800,2700") |
| 187 | + |
| 188 | +driver = webdriver.Chrome(options=chrome_options) |
| 189 | +driver.get(f"file://{temp_path}") |
| 190 | +time.sleep(5) # Wait for chart to render |
| 191 | +driver.save_screenshot("plot.png") |
| 192 | +driver.quit() |
| 193 | + |
| 194 | +# Clean up temp file |
| 195 | +Path(temp_path).unlink() |
0 commit comments