{"spec_id":"hexbin-basic","library":"pygal","language":"python","code":"\"\"\" anyplot.ai\nhexbin-basic: Basic Hexbin Plot\nLibrary: pygal 3.1.0 | Python 3.13.13\nQuality: 85/100 | Created: 2026-05-29\n\"\"\"\n\nimport importlib.util\nimport math\nimport os\nimport re\nimport sys\n\nimport cairosvg\nimport numpy as np\n\n\n# Ensure we import the installed pygal package, not this file\npygal_spec = importlib.util.find_spec(\"pygal\")\nif pygal_spec and pygal_spec.origin != __file__:\n import pygal\n from pygal.style import Style\nelse:\n cwd = os.getcwd()\n sys.path = [p for p in sys.path if os.path.abspath(p) != cwd]\n try:\n import pygal\n from pygal.style import Style\n finally:\n sys.path.insert(0, cwd)\n\nTHEME = os.getenv(\"ANYPLOT_THEME\", \"light\")\n\n# Theme-adaptive chrome tokens (Imprint style guide)\nPAGE_BG = \"#FAF8F1\" if THEME == \"light\" else \"#1A1A17\"\nINK = \"#1A1A17\" if THEME == \"light\" else \"#F0EFE8\"\nINK_MUTED = \"#6B6A63\" if THEME == \"light\" else \"#A8A79F\"\n\n\n# Imprint sequential colormap: brand green (#009E73) → blue (#4467A3), sparse → dense\nN_LEVELS = 6\ndensity_colors = tuple(\n f\"#{round(0x00 + (0x44 - 0x00) * i / (N_LEVELS - 1)):02X}\"\n f\"{round(0x9E + (0x67 - 0x9E) * i / (N_LEVELS - 1)):02X}\"\n f\"{round(0x73 + (0xA3 - 0x73) * i / (N_LEVELS - 1)):02X}\"\n for i in range(N_LEVELS)\n)\n\n# Data: urban air quality sensor network with three pollution hotspots\nnp.random.seed(42)\ncore_x = np.random.randn(2500) * 0.6 + 1.5\ncore_y = np.random.randn(2500) * 0.6 + 3.0\nindustrial_x = np.random.randn(1200) * 1.1 - 2.5\nindustrial_y = np.random.randn(1200) * 0.9 - 1.0\nhighway_x = np.random.randn(600) * 0.4 + 5.0\nhighway_y = np.random.randn(600) * 1.8 + 1.0\nbg_x = np.random.uniform(-4.0, 6.5, 150)\nbg_y = np.random.uniform(-3.0, 5.5, 150)\n\nsensor_x = np.concatenate([core_x, industrial_x, highway_x, bg_x])\nsensor_y = np.concatenate([core_y, industrial_y, highway_y, bg_y])\n\n# Hexagonal binning: assign points to tessellating hex cells, count per cell\nGRIDSIZE = 20\nPAD = 0.2\nx_min, x_max = sensor_x.min() - PAD, sensor_x.max() + PAD\ny_min = sensor_y.min() - PAD\n\nhex_width = (x_max - x_min) / GRIDSIZE # pointy-top: x-spacing = sqrt(3)*R\nhex_height = hex_width * 2 / np.sqrt(3) # y-spacing = 2R, circumradius R = hex_width/sqrt(3)\n\nrows = ((sensor_y - y_min) / hex_height).astype(int)\nodd_shift = np.where(rows % 2 == 1, 0.5, 0.0)\ncols = ((sensor_x - x_min) / hex_width - odd_shift).astype(int)\n\ncell_ids = cols * 10000 + rows\nunique_cells, counts = np.unique(cell_ids, return_counts=True)\ncell_cols = unique_cells // 10000\ncell_rows = unique_cells % 10000\ncx = x_min + (cell_cols + np.where(cell_rows % 2 == 1, 0.5, 0.0)) * hex_width + hex_width / 2\ncy = y_min + cell_rows * hex_height + hex_height / 2\n\n# Log-spaced density thresholds → 6 colour levels\nc_min, c_max = float(counts.min()), float(counts.max())\nedges = np.logspace(np.log10(c_min), np.log10(c_max + 1), N_LEVELS + 1)\nedges[0] = c_min\nedges[-1] = c_max + 1\n\nlevel_names = [\"Sparse\", \"Low\", \"Medium\", \"Moderate\", \"Dense\", \"Hotspot\"]\nlabels = [f\"{level_names[i]} ({int(edges[i])}–{int(edges[i + 1])})\" for i in range(N_LEVELS)]\n\nseries_data: list[list] = [[] for _ in range(N_LEVELS)]\nfor x, y, cnt in zip(cx, cy, counts, strict=True):\n level = min(int(np.searchsorted(edges[1:], cnt)), N_LEVELS - 1)\n series_data[level].append({\"value\": (round(float(x), 2), round(float(y), 2)), \"label\": f\"{int(cnt)} readings\"})\n\n# Imprint sequential palette, theme-adaptive chrome\ncustom_style = Style(\n background=PAGE_BG,\n plot_background=PAGE_BG,\n foreground=INK,\n foreground_strong=INK,\n foreground_subtle=INK_MUTED,\n colors=density_colors,\n opacity=0.95,\n opacity_hover=1.0,\n title_font_size=66,\n label_font_size=56,\n major_label_font_size=44,\n legend_font_size=44,\n value_font_size=36,\n title_font_family=\"sans-serif\",\n label_font_family=\"sans-serif\",\n major_label_font_family=\"sans-serif\",\n legend_font_family=\"sans-serif\",\n)\n\n# Axis viewport: 2nd/98th percentile + padding to trim outlier tails\nx_lo = float(np.percentile(sensor_x, 2)) - 0.4\nx_hi = float(np.percentile(sensor_x, 98)) + 0.4\ny_lo = float(np.percentile(sensor_y, 2)) - 0.4\ny_hi = float(np.percentile(sensor_y, 98)) + 0.4\n\n# Uniform circumradius for tessellating hexagons (SVG user units = pixels at 3200px width)\n# Estimate: plot area ≈ 2900px wide, x_range ≈ 10.8 km, gridsize=20 →\n# hex_width_px = 2900/20 = 145, circumradius = 145/sqrt(3) ≈ 84\n# Use 56 for crisp gaps between hexes (standard hexbin look) and title headroom\nHEX_R = 56\n\nchart = pygal.XY(\n width=3200,\n height=1800,\n style=custom_style,\n title=\"hexbin-basic · python · pygal · anyplot.ai\",\n x_title=\"Sensor Grid X (km)\",\n y_title=\"Sensor Grid Y (km)\",\n show_legend=True,\n legend_at_bottom=True,\n legend_at_bottom_columns=3,\n legend_box_size=24,\n stroke=False,\n dots_size=HEX_R,\n show_x_guides=False,\n show_y_guides=False,\n xrange=(x_lo, x_hi),\n range=(y_lo, y_hi),\n truncate_legend=-1,\n print_values=False,\n human_readable=True,\n x_labels_major_count=6,\n y_labels_major_count=6,\n show_minor_x_labels=False,\n show_minor_y_labels=False,\n value_formatter=lambda v: f\"{v:.1f} km\",\n)\n\nfor i in range(N_LEVELS):\n if series_data[i]:\n chart.add(labels[i], series_data[i], dots_size=HEX_R)\n\n# SVG post-processing: replace circle markers with hexagonal polygons\nsvg_raw = chart.render()\nsvg_text = svg_raw.decode(\"utf-8\") if isinstance(svg_raw, bytes) else svg_raw\n\n\ndef _circle_to_hex(match: re.Match) -> str:\n tag = match.group(0)\n cx_m = re.search(r'cx=\"([\\d.e+-]+)\"', tag)\n cy_m = re.search(r'cy=\"([\\d.e+-]+)\"', tag)\n r_m = re.search(r\"\\br=\\\"([\\d.e+-]+)\\\"\", tag)\n if not (cx_m and cy_m and r_m):\n return tag\n r_v = float(r_m.group(1))\n if r_v < 1.0: # skip tiny decorative circles (legend dots, etc.)\n return tag\n xc, yc = float(cx_m.group(1)), float(cy_m.group(1))\n pts = \" \".join(\n f\"{xc + r_v * math.cos(math.radians(a)):.2f},{yc + r_v * math.sin(math.radians(a)):.2f}\"\n for a in range(0, 360, 60)\n )\n result = re.sub(r'\\bcx=\"[\\d.e+-]+\"', \"\", tag)\n result = re.sub(r'\\bcy=\"[\\d.e+-]+\"', \"\", result)\n result = re.sub(r\"\\br=\\\"[\\d.e+-]+\\\"\", f'points=\"{pts}\"', result, count=1)\n return result.replace(\"]*/>\", _circle_to_hex, svg_text)\n\n# Inject a clip-path for the data area so hexagons don't bleed into title/margins.\n# Extract the plot group transform to find the data area bounds in SVG coordinates.\nplot_transform_m = re.search(r''\n )\n # Insert clipPath into \n svg_hex = svg_hex.replace(\"\", clip_def + \"\", 1)\n # Apply clip to each series group (leaves axis/label groups unclipped)\n svg_hex = re.sub(\n r'()', lambda m: m.group(0).replace(\">\", f' clip-path=\"url(#{clip_id})\">', 1), svg_hex\n )\n # Subtle plot frame — inner background rect from plot group (width/height relative to transform)\n inner_m = re.search(r']+class=\"plot_background\"[^>]*/>', svg_hex) or re.search(\n r']+class=\"graph\"[^>]*/>', svg_hex\n )\n if inner_m:\n inner_tag = inner_m.group(0)\n iw_m = re.search(r'width=\"([\\d.]+)\"', inner_tag)\n ih_m = re.search(r'height=\"([\\d.]+)\"', inner_tag)\n if iw_m and ih_m:\n iw, ih = float(iw_m.group(1)), float(ih_m.group(1))\n frame_svg = (\n f''\n )\n svg_hex = svg_hex.replace(\"\", frame_svg + \"\\n\")\n\n# Save interactive HTML and static PNG\nwith open(f\"plot-{THEME}.html\", \"w\", encoding=\"utf-8\") as f:\n f.write(svg_hex)\n\ncairosvg.svg2png(bytestring=svg_hex.encode(\"utf-8\"), write_to=f\"plot-{THEME}.png\")\n"}