atlasify 0.4.1

Applies ATLAS style to matplotlib plots

atlasify

The Python package atlasify applies the ATLAS style to matplotlib plots. This includes

• Switching to Arial font (not Helvetica since it's not widely available),
• Adding ticks on all edges,
• Making ticks to inward,
• Adding the ATLAS badge with optional labels (e.g. Internal),
• Adding a description below the badge, and
• Moving the ATLAS badge outside the axes area.

Quickstart

The package will use Helvetica. The package ships with GPL-licensed Nimbus Sans L as a fallback.

The atlasify package can be installed using pip.

pip install atlasify
# or 
pip install https://gitlab.cern.ch/fsauerbu/atlasify/-/archive/master/atlasify-master.tar.gz

Usage

To apply the basic style, simply call the method without any arguments.

import matplotlib.pyplot as plt
import numpy as np
from atlasify import atlasify

x = np.linspace(-3, 3, 200)
y = np.exp(-x**2)

plt.plot(x, y)
atlasify()
plt.savefig("test_1.pdf")

Label

If the first argument is a string, e.g. Internal, it is added after the ATLAS badge.

plt.plot(x, y)
atlasify("Internal")
plt.savefig("test_2.pdf")

Subtext

The second argument can be used to add text on the second line. Multiple lines are rendered independently.

plt.plot(x, y)
atlasify("Internal", 
         "The Gaussian is defined by the\n"
         "function $f(x) = e^{-x^2}$.\n")
plt.savefig("test_3.pdf")

Enlarge

Usually there is not enought space for the additinal ATLAS badge. By default, the method enlarges the y-axis by a factor of 1.3. The factor can be changed with the enlarge keyword argument.

plt.plot(x, y)
atlasify("Internal", enlarge=1.5)
plt.savefig("test_4.pdf")

Changing ATLAS

Plots for the ATLAS upgrade are not tagged wth ATLAS itself. The text of the badge can be modified via a module constant.

import atlasify as atl
atl.ATLAS = "ITk Strip"

plt.plot(x, y)
atlasify("Test beam")
plt.savefig("test_9.pdf")

Resolution, Font and figure size

The font sizes are defined in module constants and can be changed on demand. Please note that the apparent size of the badge does not change when the resolution is changed. However, the badge appears to be larger when the figure size is made smaller.

In the two following plots with different resolution, the badges take the same fraction of the canvas.

plt.plot(x, y)
atlasify("Internal")
plt.savefig("test_5.png", dpi=72)
plt.savefig("test_6.png", dpi=300)

When a smaller figure size is choose, the badge takes a larger fraction of the canvas.

plt.figure(figsize=(4,3))
plt.plot(x, y)
atlasify("Internal")
plt.savefig("test_7.pdf")

plt.figure(figsize=(4, 4))
heatmap = np.random.normal(size=(4, 4))

plt.imshow(heatmap)
atlasify("Internal", "Random heatmap, Outside badge", outside=True)
plt.tight_layout()
plt.savefig("test_8.pdf")

Axis labels

The module provides the convenience methods set_xlabel() and set_ylabel() to set the axis labels. This will make then automatically aligned to the right and top. For event more convenience you can call monkeypatch_axis_labels() to change the default align behavior.

from atlasify import monkeypatch_axis_labels
monkeypatch_axis_labels()

plt.figure(figsize=(4,3))
plt.plot(x, y)
plt.xlabel(r"$\phi$")
plt.ylabel(r"Depth (a. u.)")
atlasify("Internal")
plt.tight_layout()
plt.savefig("test_10.pdf")

Example

Real world example histogram showing two Gaussian blobs representing a Z boson background and a Higgs boson signal.

# Unbinned data
Z = np.random.normal(90, 10, size=10000)
H = np.random.normal(125, 10, size=1000)

# Manual binning, or reading from TH1F
bins = np.linspace(50, 200, 31)
Z_counts, _ = np.histogram(Z, bins=bins)
H_counts, _ = np.histogram(H, bins=bins)

plt.figure(figsize=(5,4))

# Drawing shapes
plt.hist(bins[:-1], bins=bins, weights=Z_counts,
                    label="$Z$ boson", histtype='stepfilled')
plt.hist(bins[:-1], bins=bins, weights=H_counts, bottom=Z_counts,
                    label="Higgs boson", histtype='stepfilled')

# Styling
plt.xlabel("Mass $m$ / GeV", ha='right', x=0.95)
plt.ylabel("Events / 5 GeV", ha='right', y=0.95)
plt.xlim((bins[0], bins[-1]))
atlasify("Internal", r"$\sqrt{s} = 13\,\mathrm{TeV}$")
plt.tight_layout()
plt.savefig("test_histo.pdf")