paspale 0.1.0

A unified plotting framework for academic papers

PasPale

PasPale provides a consistent, configurable interface for creating publication-ready figures with minimal boilerplate code.

Features

• 🎨 Unified Style - Consistent colors, fonts, and layouts across all plots
• ⚙️ Configurable - Dataclass-based configuration with sensible defaults
• 📊 Multiple Plot Types - Bar charts, stacked bars, dual-axis, KDE, and more
• 📄 Publication Ready - PDF output with proper fonts and sizing
• 🔧 Extensible - Easy to create custom plotters by inheritance

Installation

git clone https://github.com/undefined-c0der/paspale.git
cd paspale
pip install -e .

Quick Start

1. Simple Bar Chart

from paspale import paspale

paspale(
    data="data/results.csv",
    output="figures/comparison.pdf",
    ylabel="Speedup",
    legend=["Method A", "Method B", "Method C"],
)

2. Using Configuration

from paspale import BarPlotter, PlotConfig

config = PlotConfig(
    output="figures/performance.pdf",
    ylabel="Speedup",
    ylim=6.0,
    legend=["Baseline", "Optimized", "Ours"],
    colors="primary",
)

plotter = BarPlotter(config)
plotter.load_csv("data/results.csv")
plotter.plot()

3. KDE Distribution Plot

from paspale import KDEPlotter, KDEConfig

config = KDEConfig(
    output="figures/distribution.pdf",
    xlabel="Value",
    ylabel="Density",
)

# From CSV (each column is a distribution)
plotter = KDEPlotter(config)
plotter.load_csv("data/distributions.csv")
plotter.plot()

# Or from arrays
import numpy as np
plotter = KDEPlotter(config)
plotter.load_data([arr1, arr2], labels=["A", "B"])
plotter.plot()

4. Custom Plotter

from paspale import BarPlotter, PlotConfig

class MyPlotter(BarPlotter):
    def configure(self) -> PlotConfig:
        return PlotConfig(
            output="figures/my_plot.pdf",
            ylabel="Performance",
            ylim=10.0,
            legend=["A", "B", "C"],
        )
    
    def customize(self):
        self.ax.axhline(y=1.0, color='red', linestyle='--')

plotter = MyPlotter()
plotter.load_csv("data/my_data.csv")
plotter.plot()

Unified Interface

All plotters follow the same pattern:

plotter = SomePlotter(config)
plotter.load_csv("data.csv")   # or load_data(df)
plotter.plot()

Plotter	Config	Description
BarPlotter	PlotConfig	Grouped bar charts
StackedBarPlotter	PlotConfig	Stacked bar charts
DualAxisPlotter	PlotConfig	Dual Y-axis plots
LinePlotter	PlotConfig	Line charts
KDEPlotter	KDEConfig	Kernel density estimation

Configuration

PlotConfig Options

Option	Type	Default	Description
output	str	Required	Output file path
size	tuple	(14, 4.5)	Figure size (width, height)
ylabel	str	""	Y-axis label
xlabel	str	""	X-axis label
ylim	float	None	Y-axis upper limit
ylim_min	float	0.0	Y-axis lower limit
legend	list	[]	Legend labels
colors	str/list	"primary"	Color scheme or list
log_scale	bool	False	Use log scale for Y-axis

KDEConfig Options

Option	Type	Default	Description
output	str	""	Output file path
size	tuple	(8, 5)	Figure size
xlabel	str	""	X-axis label
ylabel	str	"Density"	Y-axis label
xlim	tuple	None	X-axis limits
fill	bool	True	Fill under curve
alpha	float	0.3	Fill transparency
use_percentage	bool	False	Show Y as percentage

Color Schemes

from paspale import colors

colors.PRIMARY    # 6 pastel colors
colors.EXTENDED   # 11 colors
colors.GRADIENT   # Blue gradient for ablations
colors.BREAKDOWN  # Component colors
colors.LINES      # Line plot colors

Examples

See examples/ directory:

• examples/bar.py - Basic bar chart
• examples/ablation.py - Ablation study with gradient colors
• examples/dual_axis.py - Bandwidth with miss rate overlay
• examples/kde.py - Distribution visualization

License

Paspale is licensed under the MIT License - see the LICENSE file for details.