Merge remote-tracking branch 'origin/main'

Author: MarkusNeusinger

.github/ISSUE_TEMPLATE/report-impl-issue.yml

@@ -42,6 +42,7 @@ body:
       label: Category
       description: What type of issue is this?
       options:
+        - Please select...
         - Visual (design, clarity, hard to read)
         - Data (unrealistic values, inappropriate context)
         - Functional (doesn't work as expected)

.github/ISSUE_TEMPLATE/report-spec-issue.yml

@@ -26,6 +26,7 @@ body:
       label: Category
       description: What type of issue is this?
       options:
+        - Please select...
         - Visual (design, clarity, hard to read)
         - Data (unrealistic values, inappropriate context)
         - Functional (doesn't work as expected)

plots/linked-views-selection/implementations/plotnine.py

@@ -0,0 +1,247 @@
+""" pyplots.ai
+linked-views-selection: Multiple Linked Views with Selection Sync
+Library: plotnine 0.15.2 | Python 3.13.11
+Quality: 91/100 | Created: 2026-01-08
+"""
+
+import os
+import tempfile
+
+import numpy as np
+import pandas as pd
+from PIL import Image, ImageDraw, ImageFont
+from plotnine import (
+    aes,
+    element_blank,
+    element_line,
+    element_rect,
+    element_text,
+    geom_bar,
+    geom_histogram,
+    geom_point,
+    geom_vline,
+    ggplot,
+    labs,
+    position_dodge,
+    scale_alpha_identity,
+    scale_color_manual,
+    scale_fill_manual,
+    theme,
+    theme_minimal,
+)
+
+
+# Data - Multivariate dataset with 3 clusters
+np.random.seed(42)
+n_per_cluster = 50
+
+categories = np.repeat(["Cluster A", "Cluster B", "Cluster C"], n_per_cluster)
+
+x = np.concatenate(
+    [
+        np.random.normal(2.5, 0.6, n_per_cluster),
+        np.random.normal(5.5, 0.7, n_per_cluster),
+        np.random.normal(4.0, 0.8, n_per_cluster),
+    ]
+)
+
+y = np.concatenate(
+    [
+        np.random.normal(3.0, 0.5, n_per_cluster),
+        np.random.normal(5.5, 0.6, n_per_cluster),
+        np.random.normal(2.0, 0.7, n_per_cluster),
+    ]
+)
+
+value = np.concatenate(
+    [
+        np.random.normal(30, 6, n_per_cluster),
+        np.random.normal(55, 8, n_per_cluster),
+        np.random.normal(42, 7, n_per_cluster),
+    ]
+)
+
+# Selection: x > 4.5 (selects Cluster B and part of Cluster C)
+selection_threshold = 4.5
+selected = x > selection_threshold
+
+# Create main dataframe
+df = pd.DataFrame(
+    {
+        "x": x,
+        "y": y,
+        "category": categories,
+        "value": value,
+        "selected": selected,
+        "Selection": np.where(selected, "Selected", "Unselected"),
+        "point_alpha": np.where(selected, 0.9, 0.35),
+    }
+)
+
+n_selected = int(selected.sum())
+n_total = len(df)
+
+# Colors - Python Blue/Yellow colorblind-safe
+color_selected = "#306998"
+color_unselected = "#AAAAAA"
+color_threshold = "#FFD43B"
+colors = {"Selected": color_selected, "Unselected": color_unselected}
+
+# Common theme for all plots with larger sizes for visibility
+base_theme = theme_minimal() + theme(
+    figure_size=(8, 7),
+    text=element_text(size=16),
+    axis_title=element_text(size=22),
+    axis_text=element_text(size=16),
+    plot_title=element_text(size=24, weight="bold"),
+    legend_position="none",
+    panel_grid_major=element_line(color="#E0E0E0", size=0.4),
+    panel_grid_minor=element_blank(),
+    plot_background=element_rect(fill="white", color="white"),
+    panel_background=element_rect(fill="white"),
+)
+
+# View 1: Scatter Plot (x vs y) - Selection source
+p1 = (
+    ggplot(df, aes("x", "y", color="Selection", alpha="point_alpha"))
+    + geom_point(size=5)
+    + geom_vline(xintercept=selection_threshold, linetype="dashed", color=color_threshold, size=2)
+    + scale_color_manual(values=colors)
+    + scale_alpha_identity()
+    + labs(title="Scatter Plot (X vs Y)", x="X Value", y="Y Value")
+    + base_theme
+)
+
+# View 2: Histogram of Value distribution
+p2 = (
+    ggplot(df, aes("value", fill="Selection"))
+    + geom_histogram(bins=15, position="identity", color="white", size=0.4, alpha=0.75)
+    + scale_fill_manual(values=colors)
+    + labs(title="Value Distribution", x="Value", y="Count")
+    + base_theme
+)
+
+# View 3: Bar chart by category showing selection breakdown
+bar_data = df.groupby(["category", "Selection"]).size().reset_index(name="count")
+
+p3 = (
+    ggplot(bar_data, aes("category", "count", fill="Selection"))
+    + geom_bar(stat="identity", position=position_dodge(width=0.8), width=0.7, color="white", size=0.6)
+    + scale_fill_manual(values=colors)
+    + labs(title="Category Breakdown", x="Category", y="Count")
+    + base_theme
+    + theme(axis_text_x=element_text(size=16))
+)
+
+# Save individual plots to temp files
+temp_dir = tempfile.mkdtemp()
+p1.save(os.path.join(temp_dir, "p1.png"), dpi=200, verbose=False)
+p2.save(os.path.join(temp_dir, "p2.png"), dpi=200, verbose=False)
+p3.save(os.path.join(temp_dir, "p3.png"), dpi=200, verbose=False)
+
+# Load images
+img1 = Image.open(os.path.join(temp_dir, "p1.png"))
+img2 = Image.open(os.path.join(temp_dir, "p2.png"))
+img3 = Image.open(os.path.join(temp_dir, "p3.png"))
+
+# Target size: 4800 x 2700 (16:9)
+final_width = 4800
+final_height = 2700
+
+# Create final canvas
+final = Image.new("RGB", (final_width, final_height), "white")
+
+# Layout: top row has 2 plots side by side, bottom row has 1 centered plot
+# Calculate panel sizes to fill canvas well
+top_panel_width = 2200
+top_panel_height = 1250
+bottom_panel_width = 2800
+bottom_panel_height = 1100
+
+# Resize panels maintaining aspect ratio by using LANCZOS
+img1_resized = img1.resize((top_panel_width, top_panel_height), Image.Resampling.LANCZOS)
+img2_resized = img2.resize((top_panel_width, top_panel_height), Image.Resampling.LANCZOS)
+img3_resized = img3.resize((bottom_panel_width, bottom_panel_height), Image.Resampling.LANCZOS)
+
+# Position panels
+title_space = 180
+panel_gap = 50
+
+# Top row: two panels side by side
+left_margin = (final_width - 2 * top_panel_width - panel_gap) // 2
+final.paste(img1_resized, (left_margin, title_space))
+final.paste(img2_resized, (left_margin + top_panel_width + panel_gap, title_space))
+
+# Bottom row: one centered panel
+bottom_left = (final_width - bottom_panel_width) // 2
+final.paste(img3_resized, (bottom_left, title_space + top_panel_height + panel_gap))
+
+# Add title and subtitle using PIL
+draw = ImageDraw.Draw(final)
+
+# Load fonts (with fallback)
+try:
+    title_font = ImageFont.truetype("/usr/share/fonts/truetype/dejavu/DejaVuSans-Bold.ttf", 64)
+    subtitle_font = ImageFont.truetype("/usr/share/fonts/truetype/dejavu/DejaVuSans.ttf", 40)
+    legend_font = ImageFont.truetype("/usr/share/fonts/truetype/dejavu/DejaVuSans.ttf", 36)
+except OSError:
+    title_font = ImageFont.load_default()
+    subtitle_font = title_font
+    legend_font = title_font
+
+# Draw main title
+title = "linked-views-selection · plotnine · pyplots.ai"
+title_bbox = draw.textbbox((0, 0), title, font=title_font)
+title_width = title_bbox[2] - title_bbox[0]
+draw.text(((final_width - title_width) // 2, 30), title, fill="#333333", font=title_font)
+
+# Draw subtitle with selection info
+subtitle = f"Selection: x > {selection_threshold} highlights {n_selected}/{n_total} points across all views"
+subtitle_bbox = draw.textbbox((0, 0), subtitle, font=subtitle_font)
+subtitle_width = subtitle_bbox[2] - subtitle_bbox[0]
+draw.text(((final_width - subtitle_width) // 2, 105), subtitle, fill="#666666", font=subtitle_font)
+
+# Add legend at bottom center
+legend_y = final_height - 70
+legend_box_size = 35
+
+# Calculate centered legend position
+legend_text_selected = "Selected"
+legend_text_unselected = "Unselected"
+selected_bbox = draw.textbbox((0, 0), legend_text_selected, font=legend_font)
+unselected_bbox = draw.textbbox((0, 0), legend_text_unselected, font=legend_font)
+total_legend_width = (
+    legend_box_size
+    + 10
+    + (selected_bbox[2] - selected_bbox[0])
+    + 80
+    + legend_box_size
+    + 10
+    + (unselected_bbox[2] - unselected_bbox[0])
+)
+legend_x = (final_width - total_legend_width) // 2
+
+# Selected legend item
+draw.rectangle([legend_x, legend_y, legend_x + legend_box_size, legend_y + legend_box_size], fill=color_selected)
+draw.text((legend_x + legend_box_size + 10, legend_y - 2), legend_text_selected, fill="#333333", font=legend_font)
+
+# Unselected legend item
+unselected_x = legend_x + legend_box_size + 10 + (selected_bbox[2] - selected_bbox[0]) + 80
+draw.rectangle(
+    [unselected_x, legend_y, unselected_x + legend_box_size, legend_y + legend_box_size], fill=color_unselected
+)
+draw.text((unselected_x + legend_box_size + 10, legend_y - 2), legend_text_unselected, fill="#333333", font=legend_font)
+
+# Save final composite image
+final.save("plot.png")
+
+# Cleanup temp files
+for fname in ["p1.png", "p2.png", "p3.png"]:
+    try:
+        os.remove(os.path.join(temp_dir, fname))
+    except OSError:
+        pass
+try:
+    os.rmdir(temp_dir)
+except OSError:
+    pass