Merge remote-tracking branch 'origin/main'

Author: MarkusNeusinger

plots/andrews-curves/implementations/letsplot.py

@@ -0,0 +1,97 @@
+""" pyplots.ai
+andrews-curves: Andrews Curves for Multivariate Data
+Library: letsplot 4.8.2 | Python 3.13.11
+Quality: 91/100 | Created: 2025-12-31
+"""
+
+import numpy as np
+import pandas as pd
+from lets_plot import (
+    LetsPlot,
+    aes,
+    element_text,
+    geom_line,
+    ggplot,
+    ggsize,
+    labs,
+    scale_color_manual,
+    scale_x_continuous,
+    theme,
+    theme_minimal,
+)
+from lets_plot.export import ggsave
+from sklearn.datasets import load_iris
+from sklearn.preprocessing import StandardScaler
+
+
+LetsPlot.setup_html()
+
+# Load and prepare data
+iris = load_iris()
+X = iris.data
+y = iris.target
+feature_names = iris.feature_names
+target_names = iris.target_names
+
+# Normalize variables to similar scales
+scaler = StandardScaler()
+X_scaled = scaler.fit_transform(X)
+
+# Create DataFrame with normalized features and species
+df_features = pd.DataFrame(X_scaled, columns=feature_names)
+df_features["species"] = [target_names[i] for i in y]
+
+# Andrews curves transformation
+# f(t) = x1/sqrt(2) + x2*sin(t) + x3*cos(t) + x4*sin(2t) + x5*cos(2t) + ...
+t_values = np.linspace(-np.pi, np.pi, 200)
+
+curves_data = []
+for idx, row in df_features.iterrows():
+    values = row[feature_names].values
+    species = row["species"]
+
+    for t in t_values:
+        # Fourier expansion
+        y_val = values[0] / np.sqrt(2)
+        for i in range(1, len(values)):
+            if i % 2 == 1:
+                y_val += values[i] * np.sin((i // 2 + 1) * t)
+            else:
+                y_val += values[i] * np.cos((i // 2) * t)
+
+        curves_data.append({"t": t, "y": y_val, "observation": idx, "species": species})
+
+df_curves = pd.DataFrame(curves_data)
+
+# Define colors for species - Python blue, yellow, and a third color
+species_colors = {"setosa": "#306998", "versicolor": "#FFD43B", "virginica": "#DC2626"}
+
+# Create plot
+plot = (
+    ggplot(df_curves, aes(x="t", y="y", group="observation", color="species"))
+    + geom_line(alpha=0.4, size=0.8)
+    + scale_color_manual(values=list(species_colors.values()))
+    + scale_x_continuous(breaks=[-np.pi, -np.pi / 2, 0, np.pi / 2, np.pi], labels=["-π", "-π/2", "0", "π/2", "π"])
+    + labs(
+        x="Parameter t (radians)",
+        y="Fourier Function Value",
+        title="andrews-curves · letsplot · pyplots.ai",
+        color="Species",
+    )
+    + theme_minimal()
+    + theme(
+        axis_title=element_text(size=20),
+        axis_text=element_text(size=16),
+        plot_title=element_text(size=24),
+        legend_title=element_text(size=18),
+        legend_text=element_text(size=16),
+        legend_position="right",
+    )
+    + ggsize(1600, 900)
+)
+
+# Save PNG (scale 3x to get 4800 × 2700 px)
+ggsave(plot, "plot.png", path=".", scale=3)
+
+# Save HTML for interactivity
+ggsave(plot, "plot.html", path=".")

plots/andrews-curves/implementations/plotly.py

@@ -0,0 +1,108 @@
+""" pyplots.ai
+andrews-curves: Andrews Curves for Multivariate Data
+Library: plotly 6.5.0 | Python 3.13.11
+Quality: 91/100 | Created: 2025-12-31
+"""
+
+import numpy as np
+import plotly.graph_objects as go
+from sklearn.datasets import load_iris
+
+
+# Data - Load Iris dataset
+iris = load_iris()
+X = iris.data  # 150 samples, 4 features
+y = iris.target
+species_names = ["Setosa", "Versicolor", "Virginica"]
+colors = ["#306998", "#FFD43B", "#E74C3C"]  # Python Blue, Python Yellow, Red
+
+# Normalize data to prevent dominant variables
+X_normalized = (X - X.mean(axis=0)) / X.std(axis=0)
+
+# Andrews curve transformation
+# f(t) = x1/sqrt(2) + x2*sin(t) + x3*cos(t) + x4*sin(2t) + x5*cos(2t) + ...
+t = np.linspace(-np.pi, np.pi, 200)
+
+
+def andrews_curve(x, t_vals):
+    """Transform a single observation to Andrews curve values."""
+    n_features = len(x)
+    curve = np.ones_like(t_vals) * x[0] / np.sqrt(2)
+    for i in range(1, n_features):
+        freq = (i + 1) // 2
+        if i % 2 == 1:
+            curve += x[i] * np.sin(freq * t_vals)
+        else:
+            curve += x[i] * np.cos(freq * t_vals)
+    return curve
+
+
+# Create figure
+fig = go.Figure()
+
+# Plot Andrews curves for each sample, colored by species
+for species_idx in range(3):
+    species_mask = y == species_idx
+    X_species = X_normalized[species_mask]
+
+    for i, x in enumerate(X_species):
+        curve_y = andrews_curve(x, t)
+        fig.add_trace(
+            go.Scatter(
+                x=t,
+                y=curve_y,
+                mode="lines",
+                line=dict(color=colors[species_idx], width=2),
+                opacity=0.4,
+                name=species_names[species_idx],
+                legendgroup=species_names[species_idx],
+                showlegend=(i == 0),  # Only show legend for first curve of each species
+                hovertemplate=f"{species_names[species_idx]}<br>t: %{{x:.2f}}<br>f(t): %{{y:.2f}}<extra></extra>",
+            )
+        )
+
+# Update layout for 4800x2700 canvas
+fig.update_layout(
+    title=dict(text="Iris Dataset · andrews-curves · plotly · pyplots.ai", font=dict(size=32), x=0.5, xanchor="center"),
+    xaxis=dict(
+        title=dict(text="Parameter t (radians)", font=dict(size=24)),
+        tickfont=dict(size=18),
+        gridcolor="rgba(128, 128, 128, 0.3)",
+        gridwidth=1,
+        zeroline=True,
+        zerolinecolor="rgba(128, 128, 128, 0.5)",
+        zerolinewidth=1,
+        range=[-np.pi, np.pi],
+        tickvals=[-np.pi, -np.pi / 2, 0, np.pi / 2, np.pi],
+        ticktext=["-π", "-π/2", "0", "π/2", "π"],
+    ),
+    yaxis=dict(
+        title=dict(text="f(t) (normalized units)", font=dict(size=24)),
+        tickfont=dict(size=18),
+        gridcolor="rgba(128, 128, 128, 0.3)",
+        gridwidth=1,
+        zeroline=True,
+        zerolinecolor="rgba(128, 128, 128, 0.5)",
+        zerolinewidth=1,
+    ),
+    template="plotly_white",
+    legend=dict(
+        font=dict(size=20),
+        x=0.98,
+        y=0.98,
+        xanchor="right",
+        yanchor="top",
+        bgcolor="rgba(255, 255, 255, 0.8)",
+        bordercolor="rgba(128, 128, 128, 0.3)",
+        borderwidth=1,
+    ),
+    margin=dict(l=100, r=80, t=100, b=80),
+    plot_bgcolor="white",
+    paper_bgcolor="white",
+)
+
+# Save as PNG (4800 x 2700 px)
+fig.write_image("plot.png", width=1600, height=900, scale=3)
+
+# Save interactive HTML version
+fig.write_html("plot.html", include_plotlyjs=True, full_html=True)

plots/andrews-curves/metadata/letsplot.yaml

@@ -0,0 +1,29 @@
+library: letsplot
+specification_id: andrews-curves
+created: '2025-12-31T21:36:48Z'
+updated: '2025-12-31T21:39:19Z'
+generated_by: claude-opus-4-5-20251101
+workflow_run: 20627522568
+issue: 2859
+python_version: 3.13.11
+library_version: 4.8.2
+preview_url: https://storage.googleapis.com/pyplots-images/plots/andrews-curves/letsplot/plot.png
+preview_thumb: https://storage.googleapis.com/pyplots-images/plots/andrews-curves/letsplot/plot_thumb.png
+preview_html: https://storage.googleapis.com/pyplots-images/plots/andrews-curves/letsplot/plot.html
+quality_score: 91
+review:
+  strengths:
+  - Excellent implementation of Andrews curves Fourier transformation algorithm
+  - Proper normalization of data using StandardScaler before transformation
+  - Good use of transparency (alpha=0.4) for overlapping curves as recommended in
+    spec
+  - Clear cluster separation visible between species
+  - Proper use of π symbols in x-axis tick labels
+  - Correct title format following pyplots.ai convention
+  - Clean, readable code structure
+  - Both PNG and HTML outputs generated for interactivity
+  weaknesses:
+  - No visible grid lines to aid visual reference (theme_minimal may have removed
+    them, but subtle grid would help)
+  - Missing explicit random seed (though data is deterministic from sklearn, good
+    practice to include)

plots/andrews-curves/metadata/plotly.yaml

@@ -0,0 +1,28 @@
+library: plotly
+specification_id: andrews-curves
+created: '2025-12-31T21:34:18Z'
+updated: '2025-12-31T21:36:51Z'
+generated_by: claude-opus-4-5-20251101
+workflow_run: 20627521331
+issue: 2859
+python_version: 3.13.11
+library_version: 6.5.0
+preview_url: https://storage.googleapis.com/pyplots-images/plots/andrews-curves/plotly/plot.png
+preview_thumb: https://storage.googleapis.com/pyplots-images/plots/andrews-curves/plotly/plot_thumb.png
+preview_html: https://storage.googleapis.com/pyplots-images/plots/andrews-curves/plotly/plot.html
+quality_score: 91
+review:
+  strengths:
+  - Excellent implementation of the Fourier transformation formula for Andrews curves
+  - Perfect use of transparency (alpha=0.4) to reveal curve density and overlap patterns
+  - Clear visual separation between Iris species demonstrates the technique utility
+  - 'Professional use of Plotly features: legendgroup for grouped legend entries,
+    custom hover templates'
+  - Proper data normalization prevents any single variable from dominating the curves
+  - Clean π notation on x-axis tick labels enhances readability
+  - Dual output (PNG + HTML) maximizes value for both static and interactive use cases
+  weaknesses:
+  - Code contains a helper function andrews_curve() violating the KISS no functions
+    rule; should inline the transformation logic
+  - Yellow color for Versicolor may be less distinguishable for colorblind users;
+    consider using a more accessible palette