feat(seaborn): implement learning-curve-basic (#2285)

## Implementation: `learning-curve-basic` - seaborn

Implements the **seaborn** version of `learning-curve-basic`.

**File:** `plots/learning-curve-basic/implementations/seaborn.py`

---
:robot: *[impl-generate
workflow](https://github.com/MarkusNeusinger/pyplots/actions/runs/20526600960)*

---------

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/learning-curve-basic/implementations/seaborn.py

@@ -0,0 +1,85 @@
+""" pyplots.ai
+learning-curve-basic: Model Learning Curve
+Library: seaborn 0.13.2 | Python 3.13.11
+Quality: 92/100 | Created: 2025-12-26
+"""
+
+import matplotlib.pyplot as plt
+import numpy as np
+import seaborn as sns
+
+
+# Data - Simulating a learning curve with typical patterns
+np.random.seed(42)
+
+# Training set sizes
+train_sizes = np.array([50, 100, 200, 400, 600, 800, 1000, 1200, 1500, 2000])
+n_sizes = len(train_sizes)
+n_folds = 5
+
+# Generate realistic learning curve pattern:
+# - Training score starts high and slightly decreases (model fits less perfectly with more data)
+# - Validation score starts low and increases (model generalizes better with more data)
+# - Gap narrows as training size increases
+
+# Training scores - high and slightly decreasing
+train_base = 0.98 - 0.03 * (train_sizes / train_sizes.max())
+train_scores = np.array([train_base + np.random.normal(0, 0.01, n_sizes) for _ in range(n_folds)])
+train_scores = np.clip(train_scores, 0.85, 1.0)
+
+# Validation scores - starts lower, increases with more data
+val_base = 0.65 + 0.25 * (1 - np.exp(-train_sizes / 500))
+validation_scores = np.array([val_base + np.random.normal(0, 0.02, n_sizes) for _ in range(n_folds)])
+validation_scores = np.clip(validation_scores, 0.55, 0.95)
+
+# Calculate means and standard deviations
+train_mean = train_scores.mean(axis=0)
+train_std = train_scores.std(axis=0)
+val_mean = validation_scores.mean(axis=0)
+val_std = validation_scores.std(axis=0)
+
+# Plot setup
+sns.set_context("talk", font_scale=1.1)
+sns.set_style("whitegrid")
+fig, ax = plt.subplots(figsize=(16, 9))
+
+# Define colors - Python Blue for training, Python Yellow for validation
+train_color = "#306998"
+val_color = "#FFD43B"
+
+# Plot training curve with confidence band
+ax.fill_between(train_sizes, train_mean - train_std, train_mean + train_std, alpha=0.2, color=train_color)
+sns.lineplot(
+    x=train_sizes,
+    y=train_mean,
+    ax=ax,
+    color=train_color,
+    linewidth=3,
+    marker="o",
+    markersize=10,
+    label="Training Score",
+)
+
+# Plot validation curve with confidence band
+ax.fill_between(train_sizes, val_mean - val_std, val_mean + val_std, alpha=0.2, color=val_color)
+sns.lineplot(
+    x=train_sizes, y=val_mean, ax=ax, color=val_color, linewidth=3, marker="s", markersize=10, label="Validation Score"
+)
+
+# Labels and styling
+ax.set_xlabel("Training Set Size", fontsize=20)
+ax.set_ylabel("Accuracy Score", fontsize=20)
+ax.set_title("learning-curve-basic · seaborn · pyplots.ai", fontsize=24)
+ax.tick_params(axis="both", labelsize=16)
+
+# Set y-axis limits for better visualization
+ax.set_ylim(0.5, 1.02)
+
+# Configure legend
+ax.legend(fontsize=16, loc="lower right", framealpha=0.9)
+
+# Subtle grid
+ax.grid(True, alpha=0.3, linestyle="--")
+
+plt.tight_layout()
+plt.savefig("plot.png", dpi=300, bbox_inches="tight")

plots/learning-curve-basic/metadata/seaborn.yaml

@@ -0,0 +1,29 @@
+library: seaborn
+specification_id: learning-curve-basic
+created: '2025-12-26T17:37:02Z'
+updated: '2025-12-26T17:41:18Z'
+generated_by: claude-opus-4-5-20251101
+workflow_run: 20526600960
+issue: 0
+python_version: 3.13.11
+library_version: 0.13.2
+preview_url: https://storage.googleapis.com/pyplots-images/plots/learning-curve-basic/seaborn/plot.png
+preview_thumb: https://storage.googleapis.com/pyplots-images/plots/learning-curve-basic/seaborn/plot_thumb.png
+preview_html: null
+quality_score: 92
+review:
+  strengths:
+  - Excellent visual clarity with well-chosen colors (Python blue/yellow) that are
+    both aesthetically pleasing and colorblind-safe
+  - Perfect implementation of learning curve pattern showing the classic bias-variance
+    tradeoff visualization
+  - Proper use of seaborn styling (set_context, set_style) creating a polished, professional
+    appearance
+  - Text sizing follows best practices exactly (24pt title, 20pt labels, 16pt ticks)
+  - Confidence bands are appropriately transparent (alpha=0.2) making overlapping
+    regions visible
+  weaknesses:
+  - Axis labels lack units (e.g., "Accuracy Score (0-1)" or "Training Set Size (samples)"
+    would be clearer)
+  - fill_between is used directly from matplotlib rather than a seaborn-native approach
+  - Data scenario is generic rather than depicting a specific real-world ML application