feat(plotnine): implement subplot-mosaic (#3066)

## Implementation: `subplot-mosaic` - plotnine

Implements the **plotnine** version of `subplot-mosaic`.

**File:** `plots/subplot-mosaic/implementations/plotnine.py`

**Parent Issue:** #3002

---
:robot: *[impl-generate
workflow](https://github.com/MarkusNeusinger/pyplots/actions/runs/20617510969)*

---------

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/subplot-mosaic/implementations/plotnine.py

@@ -0,0 +1,165 @@
+""" pyplots.ai
+subplot-mosaic: Mosaic Subplot Layout with Varying Sizes
+Library: plotnine 0.15.2 | Python 3.13.11
+Quality: 90/100 | Created: 2025-12-31
+"""
+
+import matplotlib.pyplot as plt
+import numpy as np
+import pandas as pd
+from plotnine import (
+    aes,
+    element_line,
+    element_rect,
+    element_text,
+    geom_bar,
+    geom_line,
+    geom_point,
+    geom_text,
+    geom_tile,
+    ggplot,
+    labs,
+    scale_color_manual,
+    scale_fill_gradient2,
+    scale_fill_manual,
+    theme,
+    theme_minimal,
+)
+
+
+# Data - Product performance dashboard
+np.random.seed(42)
+
+# Daily metrics for overview chart (Panel A - large top panel)
+n_days = 60
+days = np.arange(n_days)
+products = ["Alpha", "Beta", "Gamma"]
+colors = ["#306998", "#FFD43B", "#5DADE2"]
+
+df_overview = pd.concat(
+    [
+        pd.DataFrame({"day": days, "sales": 100 + i * 15 + np.cumsum(np.random.randn(n_days) * 3), "product": name})
+        for i, name in enumerate(products)
+    ],
+    ignore_index=True,
+)
+
+# Category performance (Panel B - medium right panel)
+categories = ["Q1", "Q2", "Q3", "Q4"]
+revenues = [48, 35, 42, 55]
+df_category = pd.DataFrame({"quarter": categories, "revenue": revenues})
+df_category["quarter"] = pd.Categorical(df_category["quarter"], categories=categories, ordered=True)
+
+# Distribution data (Panel C - bottom left)
+df_scatter = pd.DataFrame({"units": np.random.uniform(50, 400, 80), "margin": 15 + np.random.randn(80) * 8})
+df_scatter["margin"] = df_scatter["margin"] + 0.03 * df_scatter["units"]
+
+# Heatmap data (Panel D - bottom middle)
+regions = ["North", "South", "East", "West"]
+metrics_list = ["Sales", "Profit", "Growth"]
+heatmap_vals = np.random.rand(len(metrics_list), len(regions)) * 100
+df_heat = pd.DataFrame(
+    [
+        {"region": regions[j], "metric": metrics_list[i], "value": heatmap_vals[i, j]}
+        for i in range(len(metrics_list))
+        for j in range(len(regions))
+    ]
+)
+df_heat["region"] = pd.Categorical(df_heat["region"], categories=regions, ordered=True)
+df_heat["metric"] = pd.Categorical(df_heat["metric"], categories=metrics_list[::-1], ordered=True)
+
+# Small metric panel (Panel E - bottom right)
+months = ["Jan", "Feb", "Mar", "Apr", "May", "Jun"]
+monthly_perf = [82, 78, 91, 88, 95, 92]
+df_monthly = pd.DataFrame({"month": months, "score": monthly_perf})
+df_monthly["month"] = pd.Categorical(df_monthly["month"], categories=months, ordered=True)
+
+# Create all plotnine plots with proper theming for 4800x2700 output
+base_theme = theme_minimal() + theme(
+    plot_title=element_text(size=22, face="bold", ha="center"),
+    axis_title=element_text(size=18),
+    axis_text=element_text(size=15),
+    legend_text=element_text(size=14),
+    legend_title=element_text(size=16, face="bold"),
+    panel_grid_major=element_line(color="#cccccc", alpha=0.3),
+    panel_grid_minor=element_line(alpha=0),
+    panel_background=element_rect(fill="white"),
+)
+
+# Panel A: Sales Overview - plotnine line + point plot (LARGE - 2/3 width)
+p_overview = (
+    ggplot(df_overview, aes(x="day", y="sales", color="product"))
+    + geom_line(size=1.5, alpha=0.9)
+    + geom_point(size=2.5, alpha=0.6)
+    + scale_color_manual(values=colors)
+    + labs(x="Day", y="Sales (Units)", color="Product")
+    + base_theme
+    + theme(legend_position="right", figure_size=(12, 5.5))
+)
+
+# Panel B: Quarterly Revenue - plotnine bar plot (SMALL - 1/3 width)
+p_category = (
+    ggplot(df_category, aes(x="quarter", y="revenue", fill="quarter"))
+    + geom_bar(stat="identity", width=0.7, show_legend=False)
+    + scale_fill_manual(values=["#306998", "#4A8BBF", "#6BA3D6", "#FFD43B"])
+    + labs(x="Quarter", y="Revenue (k$)")
+    + base_theme
+    + theme(figure_size=(4, 5.5))
+)
+
+# Panel C: Units vs Margin - plotnine scatter plot
+p_scatter = (
+    ggplot(df_scatter, aes(x="units", y="margin"))
+    + geom_point(size=3.5, color="#306998", alpha=0.7)
+    + labs(x="Units Sold", y="Margin (%)")
+    + base_theme
+    + theme(figure_size=(5, 3.5))
+)
+
+# Panel D: Regional Performance - plotnine heatmap with geom_tile
+p_heatmap = (
+    ggplot(df_heat, aes(x="region", y="metric", fill="value"))
+    + geom_tile(color="white", size=1)
+    + geom_text(aes(label="value"), format_string="{:.0f}", size=10, color="white")
+    + scale_fill_gradient2(low="#FFD43B", mid="#5DADE2", high="#306998", midpoint=50)
+    + labs(x="Region", y="", fill="Score")
+    + base_theme
+    + theme(legend_position="bottom", legend_direction="horizontal", figure_size=(5.5, 3.5))
+)
+
+# Panel E: Monthly Score - plotnine bar plot
+p_monthly = (
+    ggplot(df_monthly, aes(x="month", y="score"))
+    + geom_bar(stat="identity", fill="#306998", width=0.6)
+    + labs(x="Month", y="Score")
+    + base_theme
+    + theme(figure_size=(5, 3.5))
+)
+
+# Use plotnine's composition to create a mosaic-like layout
+# The figure_size differences create visual hierarchy:
+# Panel A (12x5.5) vs Panel B (4x5.5) = ~3:1 width ratio in top row
+# This approximates the mosaic pattern where A spans more than B
+
+# Create the composition
+# Top section: overview (large ~3/4) | category (small ~1/4)
+# Bottom section: scatter | heatmap | monthly (equal thirds)
+top_row = p_overview | p_category
+bottom_row = p_scatter | p_heatmap | p_monthly
+full_layout = top_row / bottom_row
+
+# Draw the composed layout
+fig = full_layout.draw()
+
+# Adjust figure size for 4800x2700 target
+fig.set_size_inches(16, 9)
+
+# Adjust spacing - tighter wspace to keep heatmap legend near panel
+fig.subplots_adjust(top=0.88, bottom=0.10, left=0.05, right=0.95, hspace=0.30, wspace=0.18)
+
+# Add main figure title using suptitle
+fig.suptitle("subplot-mosaic · plotnine · pyplots.ai", fontsize=32, fontweight="bold", y=0.97)
+
+# Save
+fig.savefig("plot.png", dpi=300, bbox_inches="tight", facecolor="white")
+plt.close(fig)

plots/subplot-mosaic/metadata/plotnine.yaml

@@ -0,0 +1,30 @@
+library: plotnine
+specification_id: subplot-mosaic
+created: '2025-12-31T11:20:28Z'
+updated: '2025-12-31T11:30:45Z'
+generated_by: claude-opus-4-5-20251101
+workflow_run: 20617510969
+issue: 3002
+python_version: 3.13.11
+library_version: 0.15.2
+preview_url: https://storage.googleapis.com/pyplots-images/plots/subplot-mosaic/plotnine/plot.png
+preview_thumb: https://storage.googleapis.com/pyplots-images/plots/subplot-mosaic/plotnine/plot_thumb.png
+preview_html: null
+quality_score: 90
+review:
+  strengths:
+  - Excellent use of plotnine composition operators (| and /) to create mosaic-like
+    layout
+  - Clean, consistent color scheme using blue/yellow palette throughout all panels
+  - Good visual hierarchy with larger overview panel and smaller detail panels
+  - Well-structured code with clear separation of data generation and plotting
+  - Appropriate use of diverse geoms (geom_line, geom_point, geom_bar, geom_tile,
+    geom_text)
+  - Professional theming with customized text sizes for high-resolution output
+  weaknesses:
+  - Heatmap colorbar legend is placed at the very bottom of the figure, visually separated
+    from its panel
+  - Title uses suptitle which works but could be more integrated with plotnine native
+    theme system
+  - Could demonstrate the mosaic empty cell feature mentioned in the spec using placeholder
+    characters